Ask HN: Experiences with 'Gödel, Escher, Bach' (GEB)? | Hacker News Reader

Ask HN: Experiences with 'Gödel, Escher, Bach' (GEB)?

29 pointsmbm38 comments

I've heard so many good things about Hofstadter's famous book and I'm about to begin it myself. Anyone care to share their experiences with GEB/whether reading it was as transformative as I've heard it is? Also, any suggestions for how to approach the book itself?

View on HN

I finished it, but found it pretentious, boring as hell and thin on content. In my opinion (which is not shared widely ;-)) it would have made a good 15 page pamphlet. The rest is turgid verbiage(imo). I can understand someone for whom recursion(say) is some kind of mind blowing concept getting some value out of the book, but I've never understood why programmers find it fascinating. It has a few ultra basic concepts, stretched across multiple pages with lots of padding, then endlessly repeated in minor variations with lots of faux insight.

In my opinion (which is not shared widely ;-))

Share. Hofstadter always struck me as the ultimate dilettante.

Dilettante - from the Italian "to delight". I think that's fair.

Hofstadter doesn't make any claims to being an expert in any of the areas he discusses (at least not in 1979). I think of GEB as an extremely idiosyncratic but well written piece of popular science. He provides plenty of further reading for those who want to delve into the various topics he touches upon.

Out of genuine curiosity, what were expecting to get out of it before you started reading it? Also, if you found the proof of Gödel's incompleteness and RNA transcription "ultra basic concepts" my hat is off to you.

I'm a programmer with a maths background[1] and I've read the book twice–once when I was 19 and again when I was 26. Both times I really enjoyed the digressions, language games, dialogues, and playfulness of the book. I got the sense Hofstadter was deeply curious and enthusiastic about the topics he explores and did a wonderful job at sharing that enthusiasm.

In my opinion it is a great example of writing for the sheer joy of exploring ideas. I can see how approaching such a book with a "just the facts please" attitude might end badly.

[1]: I've been told I'm a strange math/programmer though. I enjoy a wide range of literature too: Cervantes, Nabokov, Wallace, and Coetzee being some of my favourite authors.

"Out of genuine curiosity, what were expecting to get out of it before you started reading it? "

I wasn't expecting anything but a good read.Which is exactly what I didn't get.

The writing style is (imo) very turgid on the other end of the spectrum from (say) pg's crisp and incisive style. Feynman is another example of a writer who exposes his interest in a wide variety of stuff(Brazilian drumming for eg) but he isn't yawn inducing. I hate writing that tests readers' patience by prolonged aimless meandering. Fwiw I didn't like _why's book on Ruby either.So there does seem to be a pattern there somewhere ;-)

I have no problems if other people like such books. Different folks different strokes and all that. I was just expressing my opinion.

"Also, if you found the proof of Gödel's incompleteness and RNA transcription "ultra basic concepts" my hat is off to you."

I am sorry but these were basic to me or I could find crisper explanations elsewhere. It has been a while since I read the book (I read this when I was 19 or so iirc) but I would guess that a few hundred pages could be lopped off with "computer programs, music and some kinds of drawing exhibit recursive structure" and giving a few examples each.

I don't remember the RNA transcription bit, but I would be very surprised if it didn't meander all over the place without saying anything much for a few dozen pages or so.

There is plenty of popular science and non fiction written crisply, interestingly and in an uncondescending fashion which explores ideas without boring the reader silly.That said I completely acknowledge this is a matter of taste and that one man's food is another's poison.

The question to which that post was a reply did ask for opinions of HNers on the book. And my opinion is that GEB is overhyped and overblown and I said as much. YMMV.And that is as it should be.

Thanks for the answers. I wasn't disputing your opinion, I just wanted to know your experience differed so much from mine. De gustibus non est disputandum.

It seems the main point of difference is that you found the meandering digressions annoying whereas I really enjoyed them. For me, understanding the topics (computation, recursion, use/mention, self-reference) wasn't the main reason for reading the book. The blurring of form and function in the writing and general playfulness were just as important and anything but boring.

PG, Feynman, and _why are useful points of reference. I like all three but recognise how different the latter one is to the former two. A similar distinction would be between Coetzee (precision) vs. Wallace (verbosity). I strenuously suggest you avoid Wallace even though I think both authors are amazing. :) Having enjoyed Wallace's non-fiction book on infinity, "Everything and More", I recommended it to a work mate (also a programmer) and he hated it. Upon reflection, I think he has similar tastes to you and would probably not like GEB either.

Out of interest, what style of fiction (if any) do you enjoy?

hey I didn't think you were disputing either. Just trying to say that I was tossing off an opinion without thinking too much about it because that was what the OP asked for.

"It seems the main point of difference is that you found the meandering digressions annoying whereas I really enjoyed them. The blurring of form and function in the writing and general playfulness were just as important and anything but boring."

I agree. Our tastes differ.

As for fiction, I generally like fiction which has a strong plot and/or characterization. Taking fantasy as a random genre, I like the first few books of George Martin's "Ice and Fire" fantasy series (the last two books meander too much imo :-P) and I can't stand Robert Jordan. I hate cardboard cut out characters and too predictable plots.

But, I digress too much from the OP's topic here so will stop now.

"Feynman is another example of a writer who exposes his interest in a wide variety of stuff(Brazilian drumming for eg) but he isn't yawn inducing."

Do keep in mind that Feynman's biographical books (Surely You're Joking, etc.) were not written by Feynman in the classical sense. They were drawn from taped conversations and written/edited by Ralph Leighton.

Many people often speak and write with different tones and levels of formality, and this can cause a message to come across in a different way.

Perhaps Hofstadter would have done better to have a number of conversations with, say, Thomas Pynchon, and then TRP could have written a very twisted version of the same book.

I think the nerdiverse would implode upon itself as the multiplicative power of people claiming to have read books they haven't would create a third book, vaguely resembling a comic book version of Ulysses.

Just to note - thats twice now you've noted the book is 'turgid' (in your opinion) - it seems to have stuck with you from the reading of the book: http://en.wikipedia.org/wiki/Egbert_B._Gebstadter ;)

I love it. But:

The book is NOT a good introduction to the incompleteness theorem. It is not a book about M.C. Escher. It could, barely, be said to be about papa Bach.

If you read it for math or science, you will be disappointed. It's a horrendous textbook.

Read it paying attention to his words, how and why he choses them, and GEB is a pretty marvelous thing. If the English language excites you, then read it, and read Le Ton Beau de Marot. GEB is a demonstration of ideas, not a discussion of them.

Hofstadter is not a computer scientist. He's a philosopher, a linguist, and a gedankenspiel-smith. Read him to laugh and play. Expect pretension because big words are fun.

Or skip it. I don't think it will change your life.

Frankly it made me feel like I'm not smart enough much of the time. But trying not to let that bother me, I plowed ahead anyway. I found quite a few things actually quite mind expanding. His exposition about infinite recursion was worth the price of admission alone. A small segment of that is here: http://amberbaldet.com/uploads/little-harmonic-labrynth.html

Another book which I did find truly transformative include "The Art of Game Design" by Jessie Schell. Also Anathem, by Neal Stephenson, is also a very worthwhile mind-broadening scientific read. A New Kind of Science by Wolfram is also quite a bit of fun.

But by far my most prized techie book is Genetic Programming, by the legendary John Koza (if you're not a Lisp person like I wasn't when I read this, then you really need to read it!). Braid and all of the books I mentioned have something in common: They are all fat, heavy, meaty, thought-provoking tomes.

Seconded, this book caused me to do more research to cover my ignorance than any text book I've ever picked up, but I absolutely enjoyed every minute of it. I also think Anathem is Stephenson's best book by far.

I've read it two times through. The first time, I thought it was one of the most brilliant books I had ever read (I read it the first time in the early 80's). I became especially fascinated with Godel's Proof and the thoughts of E.O. Wilson on whether an ant colony was in reality a single organism. I was very, very excited about the possibilities of artificial intelligence and especially natural language processing.

The second time I read it (somewhere in the 90's), I enjoyed it tremendously but I was surprised by how excited I had been before. The second time through, I caught a lot of additional insights but somehow, it felt less revolutionary.

I plan to read it again in a few years. It's been over ten years since the last time I read it. I have no idea how I'll react to it.

I read the last ~1/3 of it while recovering from food poisoning in a Buenos Aires hotel room. The simpsons was playing on TV, in spanish. Although the context was trippy, the content seemed pretty reasonable to me. It made a lot of sense from a theory of computation perspective; my response was mainly "oh, just like turing machines." Of course, I may have missed something.

My major was Cognitive Science, so I always had to read excerpts of this book as part of course requirement, but never finished the whole thing.

Now I read it by jumping between pages and hoping the individual sentences/paragraphs could spark my own thought process (this is how I read most of the time nowadays).

For a much more concise overview on the similar subject matters, checkout this book by Andy Clark http://www.amazon.com/Mindware-Introduction-Philosophy-Cogni...

You should read GEB like a playful, mathematical Alice in Wonderland full of puzzles, philosophy and sometimes self-indulgent wit. Once you go down Hofstadter's rabbit hole you'll get a mind-expanding romp through the meta-logical underpinnings of modern mathematics, information theory and computer science. At the end, whether you find it deep or simultaneously twee and pretentious is a matter of taste, but you'll definitely find yourself a lot more capable of thinking of three impossible things before breakfast.

The first time I read it straight through. I was in the middle of college, and transitioning from a dual bio/chem major to a math/cs major, and the book really hit a sweet spot for me with the nature of the universe.

If you're a programmer, then you have a big leg up on the core concepts introduced in the first half, but its still a tough read.

The biggest piece of advice: keep going, and don't be intimidated as the examples grow. There are examples that are drawn from many fields of study, and the dialogues in between chapters give a more intuitive sense of the concepts being introduced in the following chapter. Re-read those if you start finding the main discussions too dry.

There is a _lot_ going on in that book, and you will not get it all on the first read. You will have a feeling that you are missing something and you are. Don't worry about that. Some is revealed in the later parts of the book, and you might find yourself looping back to earlier passages.

Take your time with it, but get through to the end, you will not regret it, and you will find yourself picking through parts of it for a long time to come.

Also, check out Metamagical Themas, which is the opposite, a series of discrete chapters on different topics, but just as fascinating. It covers lisp, rubics cubes, the prisoners dilemma, gender roles in language and a whole host of other good things.

I highly recommend Gödel's Proof (Nagel and Newman) as a supplement. I got a lot out of Hofstadter's book, but honestly found much of the discussion of the incompleteness theorem more confusing than it needed to be. If you're mostly interested in the Gödel part, I actually recommend you read it instead of Hofstadter's book.

Gödel's Proof is very concise, and more clearly describes the core argument.

GEB is a great experience. It's not a deeply technical book, though it is both deep and technical. It's more specifically an adventure of thought.

I hear there are hidden messages in the book itself. I'm not motivated to track them down and solve them, but this is certainly an indication of the kind of tone Hofstadter was after: playful.

I absolutely love this book - I'm right in the middle, he's trying to answer the question, "Can minds be mapped onto each other?".

It's very engaging and fascinating, and I highly recommend it for all sentient beings. Admittedly I haven't finished the book just yet, but I hold the belief that I may have been better off not going to college, but instead reading GEB two or three times.

One reason I occasionally stumble while reading it is because it's such a vast flow of information. It's a bit hard to process all of his thoughts at once, as it's basically a well-organized dump of Hofstadter's brain. Reminds me of Pirsig's Zen & the Art of Motorcycle Maintenance in that way.

As for how to approach reading the book, I recommend reading it outdoors in the summertime, which is a bit tricky in the northern hemisphere these days. And don't forget to have a LOT of coffee on-hand!

enjoy! :)

I first read GEB 2 years ago and was similarly mind-blown. I found that it didn't answer any questions, but it did ask a lot. I would consider essential reading for anyone studying AI, and indeed for anyone with a curious mind.

I'm also in the middle of I Am a Strange Loop, and it feels very familiar. Hofstadter's wit and skill with words comes through strongly, and the technical depth of thought is equally as compelling.

With regards to approaching the book, I found it enjoyable and helpful to stop every few pages and think about some of the concepts that are posed. Don't consider it a book on AI, or computing, or Bach, or Escher, or Godel, or quirky narratives. Instead, just go with it and follow along with the story. It takes a while, but once you've finished, you'll want to read it again.

As someone who does the majority of their reading in bed, GEB took a long time to get through, and I definitely didn't pick up nearly as much of it as I should have.

That's not to say it's necessarily dry or boring, it's just, I imagine your brain has to be on to really appreciate it.

MIT OpenCourseWare has a couple of resources you may enjoy if you're looking at GEB:

A Highlights for High School course called Gödel, Escher, Bach: A Mental Space Odyssey. This course has a set of 16 lecture notes and 6 video lectures. http://ocw.mit.edu/high-school/courses/godel-escher-bach/

The undergraduate special course, SP.258 / ESG.SP258 Gödel, Escher, Bach. http://ocw.mit.edu/courses/special-programs/sp-258-goedel-es...

It remains one of my favorite books. I read it once my senior year in high school and I think that it influenced (perhaps unconsciously) my course selection at university (cognitive neuroscience/EECS simultaneous degrees). I re-read it again as a senior in college and was amazed by how little I really understood during my first read and how nearly all the classes I chose fit into the subjects covered in the book.

I thought it was awesome when I first read it in high school; the second time, after graduating university, I skipped all of the Achilles & the Tortoise sections and became quite frustrated at how many page turns it took to contain any given idea. Like Devilboy said, there's not reason it should have been 800pp.

I'm in the middle of it and reading it slowly. It hasn't quite lived up to the hype, but I have deeply enjoyed it. I think it's wonderfully quirky and while yes, it's probably too long, I think that's ok, because it's not a blog post and I am happy to make room for something massive and rambly on my bookshelf.

I made an attempt to read Hofstadter's more recent I'm a Strange Loop about a year ago and didn't make it all the way through.

After being told (a few weeks ago, actually) that GEB was a bit more accessible, relative at least to ISL, I started reading it, and have thus far been blown away.

Good fun but too long. I loved it when I first read it, and in college I ran a reading course where we read it over one semester. But when I learned more and later came back to it, I found it disappointing content-wise. I'm still glad I read it, though; it was pretty formative for me.

It got me hooked on programming, classical music and Escher at a young age. The best way to approach it is to realise the author's intention. He wants to explore the question: what is a self? If you keep that in mind, then his overall approach makes more sense.

One of the few books I've put down after less than 1/3rd. Really bad and over-hyped, IMHO.

i hate to be 'that guy' who says "me too", but i felt the same way when reading it 5 years ago. i had far higher expectations of a pulitzer prize winning book in non-fiction. another pulitzer prize winner that's almost as long but far more engaging is Jared Diamond's "Guns, Germs, and Steel"

GGS is an amazing read. never found a more elegant and simple explanantion of the history of mankind. in hindsight it looks perfectly rational and logical.

Yes, probably the single most important book I've read, and by far the best tutorial in symbolic logic that I know. Best approach: disappear for a month, take a holiday from the internet.

Based on what I'd heard, I expected a lot more from it. At times, Hofstadter somehow manages to turn what should be insanely interesting subject matter into something quite dry.

Great book. Like most people that start it, I didn't finish. It is a notoriously underfinished book. Give it a try anyways.

I've read it twice. That says it all, I guess.

I hated it. I felt like the author was just trying too hard. There's just not enough actual content in a book of 800 pages.

I am 200 pages in. For now, it has just radically changed my view of the concept of formal systems.

Reply on news.ycombinator.com