Eli Bendersky's website Uh oh... here it is, THE review. Yep, ladies and gentlemen, it's the review I'm most excited to write, because it is a review of the best book I've read in my life. Twice. And will read again, and again, and again. However, don't expect much from the review, as I'm confident I won't be able to express all my emotions and impressions from this book in writing. If you have a few hours to spare, I'll delightfully discuss it.
I first read GEB when I bought it a few years ago. Back then, it was hard for me to understand. There were a lot of things I didn't fully grasp after finishing it. Did it get better this time ? Nope. Though I understood much more, I still feel there's a whole tanker of ideas and beauty I missed.
And how can one understand such a book in one read ? On its ~750 pages, the author covers math, logic, reasoning, music, art, computers, programming, AI, biology, zen, and much, much more. Before I read this book, I never really understood how much an art is book-writing. It seems that so much thought and caring went into writing each chapter, so you just sit befuddled, trying to grasp the meaning of each word.
In the 20 years this book is on the shelves, opinions vary on what it's about. I have the anniversary edition, where Hofstadter confronts this question in a long preface. I feel, however, that each person can find something different in this book. People with different mind-settings and experiences may conclude that this book is about different topics. So in my review, I'll not try to guess what the author meant this book to be, I'll focus on what this book is about for me, from my point of view.
What is it, then ? Well, I feel that the single most important theme of this book, it's ultimate goal, is to show that "meaningful" consciousness can emerge from "meaningless" formal systems, with very simple rules. That complexity can rise from trivial governing laws of small-scale systems.
Hofstadter artfully approaches this topic from multiple angles simultaneously. First, formal systems are introduced. These systems are "typographical" - i.e. their rules and forms can be expressed on paper. This is essential to make the author's point, as he developes these systems on the pages of his book.
MIU, PQ, TNT, Typogenetics. All formal systems with underlying simple rules, used by the author to represent his points. Computing and computability (partial and general recursive) is introduced through B/F/Gloop.
Hofstadter speaks about AI a lot, which is exciting for people interested in the topic. He presents various AI programs, discussing their structure and what they teach us about "simulating minds". In his discussion of Church's thesis, in various versions, he shows what can be computed by machines vs. what can be computed by human beings.
Another appreciation I must express for the author is his control of the English language. The amount of word-plays and tricks in this book is astounding. Whole passages are written to be similar to fugues and canons in music. The style, especially of the dialogues, is something I've never seen or experienced before.
I thoroughly enjoyed reading this book. And unfortunately, I feel that I just can't express it all in this review. If you haven't read this yet, do yourself a favor. If you read only one book about consciousness and AI in your life, this is it. If you've read this book, I warmly recommend tho others: "Chaos" by Gleick - to get more insight into the emergence of complexity from trivial rules, and "The selfish gene" by Dawkins, to discuss this emergence in relation to our origin and evolution.