The "favorite books list" meme has been going around various circles I read, probably started by Dr. Principles being annoyed by CalPundit's being annoyed by Clinton's favorite book list, then getting kicked around several groups and finally ending up with me having enough time to think about mind and write it up with some sort of justification.
By the time I got around to thinking about it, the list question also included the meta-question, "How do you decide what your favorite books are?" For me, two common themes are whether or not the book is "just right," and whether or not the book has touched off a series of ideas in my head or in some other way substantially altered my point of view.
For the former, for fiction especially, that has to mean that the book has to work for me on all (or almost all) levels, very well-- the prose, the plotting, the characterization, the symbolism, etc. For non-fiction books, it helps to be well-written in its own right, to give a reasonably complete treatment of its subject at whatever level its aiming, and it helps to give a balanced perspective if there's some controversy in the field.
For the latter, well. It means what it means.
The very best books do it all at once-- readable, engaging, personable, and touch off an explosion of ideas in my head, giving me a new way to look at a problem, or an issue, or an entire field.
In the fiction category, these come to mind, in no particular order:
Deadhouse Gates by Steven Erikson. Really, the whole Malazan series thus far has been brilliant, but the second was a special volume in a special series. I love the series for its conjuration of a theme of simple human dignity out of a story that almost revels in showing the opposite. I love the second book becuase of the Chain of Dogs, and the telling of Coltaine's final fate. Brilliance.
The Neverness Cycle by David Zindell (Neverness, The Broken God, The Wild, and War in Heaven.) I tried to narrow this down to one, then to two, but just couldn't. The whole series succeeds like very very few others, in my experience, at creating a society (human based in this case) that is at once almost completely alien, but somehow still comprehensible. It's also just absolutely crackling with ideas and extrapolations and religious and mathematical scenes... and I just like the writing style.
Neverness is rough but still shines, like an uncut gem. The Broken God has one of the best protagonist-antagonist relationships I've ever seen. And any book-- in this case, The Wild-- that can make me appreciate the pacifist view, even for a split second, has something going for it.
Alas, while I don't think Neverness is a love it or hate it series, I do think it's a get it or don't get it series, and it probably doesn't appeal to everyone. But it's no accident that my blog is named Shih.
Tigana by Guy Gavriel Kay. Because I have pride of place, and ambition for myself and my place of pride.
The Cryptonomicon by Neal Stephenson. This is probably the best thing Stephenson has written (with the caveat that I haven't finished Quicksilver and the additional caveat that I've bogged down badly in the second section of it.) It's big, it's fast, it's larger than life, and Stephenson has a way of writing that has you laughing at the same time he's putting characters you care about through sheer torture. His books are often humorous, but not really funny if you think about them.
(Woo! Here we see Randy Waterhouse crowbar himself out of a doomed relationship with a woman who has nothing but contempt for him and eventually get sent to prison by shady forces trying to exploit his family knowledge! Ho ho! Here we see Goto Dengo get his ass kicked across the South Pacific, get his ass captured by headhunters and then shit on when he's rescued by his countrymen! And get this! We see Bobby Shaftoe abandon the woman he loves and the child he doesn't know about, get *his* ass kicked around the South Pacific, and deal with it all by developping a morphine addiction! What fun! The Big U is like that, too.)
It's also just plain well written. I don't care when people say the book is bloated and neeed editting. No, it didn't. Nor was the purpose of it just to enjoy the scenery. A lot of those extraneous scenes tended to loop back into the plot, one way or another.
A Deepness in the Sky by Vernor Vinge. I really like Vinge's fiction. I've read just about every piece of fiction he's written (possibly I've tracked all of them down by now, but I haven't finished Tatja Grimm's World) and very few of them are bad. As a rule, he just gets better and better as a writer-- read his short stories in chronological order, and you can actually see the skill accumulating.
Deepness is probably the most complex work he's written thus far, and the most satisfying. This is not surprising, since, with the exception of a efw much shorter pieces, it is his most recent as well.
The first five of The Amber Chronicales, by Roger Zelazny. Happily, the second five and the subsequent short stories did not ruin the whole thing for me. Zelazny was just a hell of a writer. I've been through those books so very often that I know each wart and blemish in them by heart, but the fun of it all still comes through.
All of an Instant by Richard Garfinkle, shares with Zindell a really wonderful portrayal of an absolutely alien situation which is nevertheless comprehensible. I couldn't possibly even begin to explain the actual book, except to exhort people to go read it.
Foucault's Pendulum by Umberto Eco. Heh. Heh heh. Any book that still has me chuckling at the scholarly in-jokes and general silliness of it all has something gong for it.
In non-fiction:
Groups: A Pathway to Geometry by R. P. Burn. Yes, yes, it's a book on a group theory approach to modern geometry. Ho hum. But if you like math, this and every other math book by Burn is wonderful, because it's pure Socratic method-- it will ask very bite-sized guided questions which a reasonable student can answer with some effort, and which magically lead up to complete proofs of some sophisticated mathematical ideas.
The First American by H. W. Brands. I like historical biographies. I like early American history. A biography of Ben Franklin is a must-read, therefore, and Brands is one of the most engaging biographers I've had the pleasure of reading. I cam away with not just a better understanding of Franklin (beyond the lightning and the kite thing) but of the whole background of the pre-revolutionary and revolutionary periods.
I am looking forward to his biographies of Wilson and Roosevelt, comma, Teddy, if I ever get around to them.
Models of Computation and Formal Languages by R. Gregory Taylor. This is one of computer science pedagogy's undiscovered gems, in my opinion. Every CS theorist must (and in my opinion, every CS non-theorist certainly should!) master certain basic concepts. Formal languages, and their mapping to various classes of automata, are in the must-know category. Taylor takes a weird, almost backward approach to it, but his explanations are some of the clearest I have ever read, yes, including even the legendary Papadimitriou, and the newly made classic, Sipser. Papadimitriou and Sipser cover more advanced material, though.
On that subject, Goedel, Escher, Bach: An Eternal Golden Braid, by Douglas Hofstadter is just a perfect (and perfectly readable) development of one of the seminal developments in 20th century mathematics: Goedel's Incompleteness Theorem. And a whole lot of other stuff. And it's fun. When I was a young man idly kicking through a used bookstore, years and years before I had any ambition of pursuing computer science at any level much less an eventual PhD, the owner pressed this book on me. He told me if I didn't like it, I could bring it back and he'd refund my money.
How could I argue?
But how did he know?!
Spacetime Physics by Taylor and Wheeler. I could teach relativity to a monkey with this book. It's that good.
The Hero With a Thousand Faces by Joseph Campbell. On the one hand, it is kinda heavy handed and can grate, after a while. On the other hand, although Campbell was a man with a hammer on a universal quest for nails, he was not a stupid or a crazy man, and there's a lot to what he has to say, about story structure, about religions, and about their intersection in mythic structures. This one goes on the list because after having read it (and other works of Campbell) I find I get a lot more out of other great works of fiction, and on different levels than before.
Nanosystems by Eric Drexler. This is pretty much the seminal work in nanotechnology, to this day. Written over ten years ago, it's still one of the very few books (as opposed to journal articles) on nanotechnology that is very much not a popular work. This is Drexler's PhD thesis from MIT, expanded, fleshed out, and solidified. It's a long hard slog through a lot of different branches of science, and it's really pretty damned impressive.
Even if the claims turn out not to be practicable, it's still a hell of an intellectual accomplishment. (And for the record, for anyone following the public debates, Smalley and Drexler are both embarassing themselves, at this point.)