My ten great books #8: Dictionary of the Khazars

The more books I read or movies I see, the more I’ve come to appreciate works of art that live up to their own promises. They don’t need to be vast or ambitious: I have great respect for straightforward genre pieces—the novels of John D. MacDonald, the movies of Michael Curtiz or Howard Hawks—that deliver on exactly what they say they will. This is doubly true of works that take big formal or conceptual risks. A movie like Memento is a pleasure because it sets itself a tremendous technical challenge and exploits it to its fullest extent. The same is true of a book like Pale Fire, which is irresistible in its conception and even better in execution. More often, you’ll see books that aim high on a structural level but can’t quite close the deal: I admire House of Leaves and Cloud Atlas, for instance, but both novels leave me with the sense that the authors, for all their obvious gifts, faltered near the end. And this isn’t their fault. For a novel to be both perfect and unique, you need more than talent: luck, ruthless patience, and the disposition of the reader all play their part. Which is all to say that Milorad Pavic’s Dictionary of the Khazars comes closer than any novel I know to laying out a series of increasingly improbable formal challenges and triumphing on every level, assuming that you’re prepared to read it on its own terms.

Dictionary of the Khazars, as its title implies, is a dictionary—or, more precisely, three dictionaries with some prefatory material and two appendices—in which the entries can be read in any order. (There’s also the small point that the book comes in two versions, male and female, that differ in a single crucial paragraph, although it’s not until you get to the final page that you understand why.) You can just read the entire book straight through, if you like, or you can read parallel entries in the three different sections, or you can follow the text from one cross-reference to the next. Characters mentioned briefly in one entry receive full treatment in another; you can read the end of one story before finding the beginning or middle; and throughout, there’s the teasing sense that you’re on the verge of uncovering the answer to a puzzle revolving around the fate of the Khazars, a tribe of Central Asian nomads that vanished shortly after their conversion to a neighboring religion, either Christianity, Islam, or Judaism. The fact that Pavic sets all these enigmas and expectations in motion and then actually resolves them is stunning enough: at first glance, the novel seems chaotic, but it’s really a perfect crystal, and it answers all the questions it raises. It’s even more miraculous that the journey is so beautiful, witty, and moving. It’s possible that I reacted to the last few pages so strongly because of the role that this book has played in my own life, as it followed me from one set of shelves to another for more than a decade, waiting patiently to be discovered. But I wouldn’t be writing this if I didn’t think it might hold the same magic for you, too.

Reading while writing

When Norman Mailer was working on The Naked and the Dead, still in his early twenties, he fell back on a trick that I suspect most novelists have utilized at one time or another. Here how he described it to his biographer Peter Manso:

I had four books on my desk all the time I was writing: Anna Karenina, Of Time and the River, U.S.A., and Studs Lonigan. And whenever I wanted to get in the mood to write I’d read one of them. The atmosphere of The Naked and the Dead, the overspirit, is Tolstoyan; the rococo comes out of Dos Passos; the fundamental, slogging style from Farrell; and the occasional overreach descriptions from Wolfe.

I haven’t looked into this in any systematic way, but I have a feeling that a lot of writers do much the same thing—they select a book by another author whom they admire, and when they start the day’s work, or feel their inspiration starting to flag, they read a few pages of it. If you’re like me, you try to move straight from the last sentence of your chosen model to your own writing, as if to carry over some of that lingering magic. And if you’re lucky, the push it provides will get you through another hour or so of work, at which point you do it again.

I’ve followed this routine ever since I started writing seriously, and it isn’t hard to figure out why it helps. One of the hardest things about writing is starting again after a break, and reading someone else’s pages has the same effect as the advice, often given to young writers, as retyping a paragraph of your work from the day before: like the running start before the long jump, it gives you just enough momentum to carry you past the hardest part. I’ve also developed a set of rather complicated rules about what I can and can’t allow myself to read while working. It needs to be something originally composed in English, since even the best translations lose something of the vitality of a novel in one’s native language. (Years ago, I saw one of Susan Sontag’s early novels described as being written in “translator’s prose,” and I’ve never forgotten it.) It has to be the work of a master stylist, but not so overwhelming or distinctive that the tics begin to overwhelm your own voice: I still vividly remember writing a few pages of a novel shortly after reading some Nabokov, and being humiliated when I went back to read the result the next day. I stay away from such writers for much the same reason that I avoid listening to music when I write these days. It’s all too easy to confuse the emotional effects produced by proximity to another work of art with the virtues of the writing itself. When you’re reading in parallel, you want a writer who bears you forward on the wave of his or her style without drowning you in it.

This also means that there are books that I can’t allow myself to read when I’m writing, out of fear that I’ll be contaminated by their influence, for better or for worse. Obviously, I avoid bad writers, but I also steer clear of great writers whom I’m afraid will infect my style. In practice, because I’m nearly always writing something, this means that I’ve avoided certain books for years. It took me a long time to read Cloud Atlas, for instance, because it seemed like exactly the kind of overwhelming stylistic experiment that could only have a damaging impact. Mailer makes a similar point:

I was very careful not to read things that would demoralize me. I knew that instinctively. There’s a navigator in us—I really do believe that—and I think this navigator knew I wanted to be a writer and had an absolute sense of what was good for me and what wasn’t. If somebody had said, “Go read Proust,” I’d say, “No, not now.”

Or as the great Sherlockian scholar Christopher Morley noted: “There is no harm in reading any number of unimportant books for pastime, but the significant books must be taken cautiously. You don’t want them to get in the way of what might perhaps be growing and brooding in yourself, taking its own time.”

And this search for books in English that have a great style, but not too much of it, has led to some curious patterns in my reading life. Usually, when I’m working on something and need a helping hand to get me over the rough patches, I go with Ian McEwan. I’m not sure that I’d describe him as my favorite living writer, but he’s arguably the one whose clean, lucid, observant prose comes closest to the ideal that I’d like to see in my own work. You can’t really go wrong with an imitation of McEwan, whereas there are other writers in the same vein, like Updike, who are more likely to lead you astray. With McEwan, at worst, you’ll end up with something boring, but it probably won’t be outright embarrassing. (It reminds me a little of what T.S. Eliot once said along similar lines: “If you follow Dante without talent, you will at worst be pedestrian and flat; if you follow Shakespeare or Pope without talent, you will make an utter fool of yourself.”) McEwan is the closest I’ve found to a foolproof choice, which is why I’m currently reading The Children Act, a few pages at a time, while I’m working up a new short story. James Salter and J.M. Coetzee are two other good options, and if I’m really stuck for inspiration, I’ll often fall back on an old favorite like Deliverance by James Dickey, or even Mailer himself, for early drafts when I’m pretty sure that I’ll have a chance to pare away any excesses of style. Every writer eventually develops his or her own personal list, and there aren’t any wrong answers. You just listen to your navigator. And maybe you don’t read Nabokov.

My ten great books #7: Dictionary of the Khazars

(Note: For the rest of the month, I’m counting down the ten works of fiction that have had the greatest influence on my life as an author and reader, in order of their first publication. For earlier entries in the series, please see here.) 

The more books I read or movies I see, the more I’ve come to appreciate works of art that live up to their own promises. These promises don’t need to be vast or ambitious: I have great respect for straightforward genre pieces—the novels of John D. MacDonald, the movies of Michael Curtiz or Howard Hawks—that gracefully deliver on exactly what they say they will. This is doubly true of works that take big formal or conceptual risks. A movie like Memento is a pleasure because it sets itself a tremendous technical challenge and exploits it to its fullest extent. The same is true of a book like Pale Fire, which is irresistible in its conception and even better in execution. More often, you’ll see books that aim high on a structural level but can’t quite close the deal: I admire House of Leaves and Cloud Atlas, for instance, but both novels leave me with a sense that the authors, for all their obvious gifts, faltered near the end. And this isn’t their fault. For a novel to be both perfect and unique, you need more than talent: luck, ruthless patience, and the disposition of the reader all play their part. Which is all to say that Milorad Pavic’s Dictionary of the Khazars, which I finished last year after failing to get through it for more than a decade, comes closer than any novel I know to laying out a series of increasingly improbable formal challenges and triumphing on every level, assuming that you’re willing to read it on its own terms.

Dictionary of the Khazars, as its title implies, is a dictionary—or, more precisely, three dictionaries with some prefatory material and two appendices—in which the entries can be read in any order. (There’s also the small point that the book comes in two versions, male and female, that differ in a single crucial paragraph, although it’s not until you get to the final page that you understand why.) You can read the entire book straight through, if you like, or you can read parallel entries in the three different sections, or you can follow the text from one cross-reference to the next. Characters mentioned briefly in one entry receive full treatment in another; you can read the end of one story before finding the beginning or middle; and throughout, there’s the teasing sense that you’re on the verge of uncovering the answer to a puzzle revolving around the fate of the Khazars, a tribe of Central Asian nomads that vanished shortly after their conversion to a neighboring religion, either Christianity, Islam, or Judaism. The fact that Pavic sets all these enigmas and expectations in motion and then actually resolves them is stunning enough: at first glance, the novel seems chaotic, but it’s really a perfect crystal, and it answers all the questions it raises. It’s even more miraculous that the journey is so beautiful, witty, and moving. It’s possible that I reacted to the last few pages so strongly because of the role that this book has played in my own life, as it followed me from one set of shelves to another, waiting patiently to be discovered. But I wouldn’t be writing this if I didn’t think it might hold the same magic for you, too.

“He sensed that something inside him had changed…”

(Note: This post is the twenty-sixth installment in my author’s commentary for The Icon Thief, covering Chapter 25. You can read the earlier installments here.)

Act breaks are hard. They’re hard enough, in fact, that it’s tempting to think that we can do without them altogether, on the principle that the division of a story into three distinct acts is a convention of lazy or formulaic writing. Artists who attempt to structure their works in more complex or intuitive ways should be commended for it, especially if they can pull it off—although it’s important to note that even a movie like Cloud Atlas organizes its separate narrative strands in what looks, when you stand back, an awful lot like a conventional beginning, middle, and end. And there’s a good reason for this. The three-act structure is the most powerful storytelling tool ever developed, and when it’s properly deployed, you don’t even notice it: you’re only aware of being drawn along by a narrative that moves with its own inevitable momentum. In a good movie, in particular, the act breaks should be invisible; if they aren’t, it’s the sign of a script that has been assembled according to one of those mechanical plans, so beloved by aspiring screenwriters, that put the central dramatic question on page three and the inciting incident on page ten. That kind of structure doesn’t do anyone any favors.

In reality, the act breaks in a story are more like the stakes and string that the architect Christopher Alexander describes in his wonderful book The Timeless Way of Building. Before you start construction on a house, you go out to the site and lay down its outlines with some twine and a few bits of wood. You adjust their position, a foot here, a foot there, until the layout looks more or less right. The result isn’t a house, or even a true plan, but it’s an essential first step. And the act breaks in a story are the stakes that you use to guide yourself as you begin to plot the story in greater detail. Whenever I begin any substantial writing project, I generally have a sense of what the two or three major turning points of the narrative will be, even before I’ve fleshed out the plot or characters. Even if you don’t outline to the extent that I do, it’s useful to at least have those big moments in mind. If nothing else, it’s a way of keeping yourself sane: I’ve often found myself lost in the shapeless middle section of a novel, but take comfort in the fact that there’s some good stuff just around the corner. And laying down the stakes for a few important moments helps you navigate the rough patches on the way to your next destination.

So where should the stakes go? They should go, well, where the stakes are greatest—at points in which the narrative has been fundamentally changed by some new crisis or development that is organic to the shape of the story itself. In The Icon Thief, the nature of my first big act break was clear from the beginning. The story so far had been structured around a heist in which Ilya, my thief, would steal the painting Study for Étant Donnés. At the end of Part I, he obtains the painting, but finds himself betrayed and on the run, with the picture still in tow. I don’t know whether this development took anyone by surprise—looking over it again now, I have a hunch that it telegraphs itself a bit too clearly—but that’s less important than the conflict it establishes, which will play out over the rest of the novel. Ilya has defined himself as a thief who is working, in some small way, against the system that abandoned him. When he discovers that, in fact, he’s been a part of that system all along, it shifts the terms of the story and transforms him from a skilled professional into a man driven by self-preservation and, ultimately, revenge. Which is the kind of meaningful change required to propel the characters, and the reader, into the long second half of the novel.

And much of the impact of this moment is, for lack of a better word, typographical. At the end of Chapter 25, the reader sees Ilya on the run, racing for his life through a darkened field with everything he believed in ruins—and then turns to a blank page, followed by the stark pair of epigraphs for Part II. The fact that the reader has to turn the page to see that a new section is beginning is an accident of layout: later, in Part III, and for both major act breaks in City of Exiles, the new section appears on the facing page of the previous one, which means that the reader can tell, out of the corner of his or her eye, that a new phase in the story is beginning. That’s fine, but I’m still tickled by the way the reader needs to turn a page here to see that we’ve reached the end of the novel’s first movement. As I said above, in a movie, this transition should be invisible, but there’s something weirdly satisfying, at least for me, when a novel manages to gather all of its narrative threads together in one moment of crisis as its first section visibly ends. Of course, picking up the action again in the second half presents problems of its own. But that’s a topic for another day…

The glorious fiasco of Cloud Atlas

A few months ago, I wrote a piece for Salon on whether there was such a thing as a New Yorker feature curse. I was largely inspired by the example of John Carter, in which the magazine’s highly positive profile of director Andrew Stanton was followed shortly thereafter by a debacle that deserves its own book, like Final Cut or The Devil’s Candy, to unpack in all its negative glory. Judging from the response, a lot of readers misunderstood the piece, with one commenter sniffing that I should read Carl Sagan’s The Demon-Haunted World before spreading so much superstition. My point, to the extent I had one, was that the New Yorker curse, like its counterpart at Sports Illustrated, was likely a case of regression to the mean: magazines like this have only a limited amount of feature space to devote to the movies, which means they tend to pick artists who have just had an outstanding outlier of a success—which often means that a correction is on the way. And although my theory has been badly tested by Seth MacFarlane’s Ted, which is now the highest-grossing R-rated comedy of all time, at first glance, the recent failure of Cloud Atlas, which follows a fascinating profile of the Wachowskis by Aleksandar Hemon, seems to indicate that the curse is alive and well.

Yet at the risk of sounding exactly as arbitrary as my critics have accused me of being, I can’t quite bring myself to lump it into the same category. This isn’t a movie like John Carter, which was undermined by a fundamentally flawed conception and a lot of tactical mistakes along the way. Cloud Atlas has its problems, but as directed by the Wachowskis and Tom Tykwer after the novel by David Mitchell, it’s a real movie, an ambitious, entertaining, often technically spellbinding film that probably never had a shot at finding a large popular audience. I’m not a huge fan of the Wachkowskis, who over the past decade have often seemed more intelligent in their interviews than in their movies, but I give them and Tykwer full credit for pursuing this dazzling folly to its very end. Cloud Atlas is like The Tree of Life made in a jazzy, sentimental, fanboyish state of mind, and although it doesn’t succeed entirely, under the circumstances, it comes closer than I ever expected. It’s the kind of weird, personal, expensive project that gives fiascos a good name, and it’s one of the few movies released this year that I expect to watch again.

And with one exception, which I’ll mention in a moment, the movie’s flaws are inseparable from its fidelity to the underlying material. I liked Mitchell’s novel a lot, and as with the movie it inspired, it’s hard not to be impressed by the author’s talents and ambition. That said, not all of its nested novelettes are equally interesting, and its structure insists on a deeper network of resonance that isn’t always there. Some of its connections—the idea that Somni-451 would become a messianic figure for the world after the fall, for instance, or that she’d want to spend her last few moments in life catching up with the story of Timothy Cavendish—don’t quite hold water, and in general, its attempts to link the stories together symbolically, as with the comet-shaped birthmark that its primary characters share, are too facile to be worthy of Mitchell’s huge authorial intelligence. (You only need to compare Cloud Atlas to a book like Dictionary of the Khazars, which does keep the promises its structure implies, to see how the former novel falls short of the mark.) And the movie suffers from the same tendency to inform us that everything here is connected, when really, they’re simply juxtaposed in the editing room.

All the same, the movie, like the book, is one that demands to be experienced. There are a few serious lapses, most unforgivably at the end, in which we’re given a new piece of information about the frame story—not present in the original novel—in the clumsiest way imaginable. For the most part, however, it’s fun to watch, and occasionally a blast. Somewhat to my surprise, my favorite sequences were the ones directed by Tykwer, an unreliable director who also offered up one of the best action scenes in recent years with the Guggenheim shootout in The International: he gives the Louisa Rey narrative a nice ’70s conspiracy feel, and the story of Timothy Cavendish, which I thought was unnecessary in the novel, turns out to be the most entertaining of all. (A lot of this is due to the presence of Jim Broadbent, who gives the best performance in the movie, and one of the few not hampered by elaborate but frequently distracting makeup.) The Wachowskis can’t do much with the journal of Adam Ewing, but the futuristic ordeal of Somni-451 is right in their wheelhouse. It’s a movie that takes great risks and succeeds an impressive amount of the time. And as far as I’m concerned, the curse is broken. At least for now.

The mystery of the Khazars

A good book is like a journey, for the reader as well as the writer, and some journeys take longer than others: I just made it to the end of one that took me fifteen years to complete. This book is Milorad Pavic’s Dictionary of the Khazars, but before I talk about it here, I need to explain why I find novels like this so fascinating. According to the Norwegian scholar Espen J. Aareseth, there are two kinds of literature: ergodic, which places nontrivial demands on the reader in assembling or traversing the text, and nonergodic, which asks nothing more of the reader than “eye movement and the periodic or arbitrary turning of pages.” Aarseth defines the former term rather narrowly, but for me, it refers to books that constantly remind you of your process as a reader, often by forcing you to regularly turn pages in both directions. And not surprisingly, such works tend to cast a certain spell over compulsive readers who may not be able to finish them, including me.

As a result, I’ve always been drawn to this sort of story, perhaps because my own fiction is so relentlessly linear. The trouble is that such novels often promise more than they’re finally able to deliver, once all the pieces have been put into place. House of Leaves, for instance, does a terrifying job of hinting at some unspeakable horror lying in wait for readers who can make it all the way through—it’s the only book of the past decade that gave me nightmares—but it ultimately disintegrates before our eyes. Cloud Atlas, as I’ve noted before, is a brilliant piece of writing and imagination, but it finally comes off as a set of nested novelettes with only occasional stabs at larger meaning. Until recently, the only ergodic novel I’d read that really lived up to its promise was Nabokov’s Pale Fire, which does exactly what such a novel ought to do: implicate the reader, inextricably, in the process of its own creation, until it seems less like a book than a place we’ve gone to visit and can never quite escape.

Which brings me to Dictionary of the Khazars. I bought this novel by the Serbian poet Milorad Pavic more than fifteen years ago, intrigued both by its design, which is among the most beautiful I’ve ever seen in a novel, and its unique structure: it’s arranged as a dictionary, or rather three parallel dictionaries, the entries of which can be read in any order, either by following cross-references or simply browsing at random. In the years that followed, I often leafed through the book, to the point where I know certain entries extremely well, but never read more than a third of the material, mostly because it was hard to keep track of how much I had remaining. And while I never forgot it, I don’t think I would have finished it if I hadn’t recently become interested in the historical Khazars themselves, as part of the novel I’m currently trying to write. On a whim, then, along with the more factual sources I was reading, I decided to give Pavic another try, on the off chance he could provide me with some useful ideas.

And what I discovered, much to my astonishment, is that this is one of the few novels of its kind that really lives up to its promise. Dictionary of the Khazars constantly hints at a greater pattern that will be visible to readers who finish the entire book, and amazingly enough, it delivers—when you’ve read the last few pages, the full picture locks into place, and the effect is shattering and unforgettable. The result makes House of Leaves or Cloud Atlas seem half-baked by comparison, and the difference is less one of inspiration than of pure craft: Pavic, who died several years ago, is just as inventive and fanciful as his ergodic peers, but if the book’s individual entries reflect a maddening poetic imagination, its overall structure is ruthlessly logical. (Note that the order of the entries isn’t entirely random: at least one crucial entry and the book’s second appendix aren’t cross-referenced anywhere else, which subtly ensures that the reader will approach them last.) It’s nothing less than one of the best modern novels I’ve read in a long time. And it only took me fifteen years to get here.

A year’s worth of reading

These days, I’m fortunate enough to have more work than I can handle, which also means that I no longer have much time to read for my own pleasure. The past year, in particular, was all business: I had just over nine months to take City of Exiles from conception to final draft, along with a number of other projects, which meant that nearly all my free time was devoted to either writing or research. All the same, I managed to make time to read a number of books that didn’t have anything to do with my work, either in my spare moments, on vacation, or in parallel with writing the novel itself. (Like many writers, I like to read a few pages of an author I admire before starting work for the day, which means that I tend to read books in piecemeal over the course of many weeks or months.) And while I doubt I’ll ever return to being the sort of omnivorous reader I was growing up, it’s still important to me to read as much as possible, both for professional reasons and for the sake of my own sanity.

Much of this year was spent catching up on books that I’d been meaning to read for a long time. The best book I read this year, by far, was The Magic Mountain by Thomas Mann, which seems likely to stand as one of my ten favorite novels, followed close behind by Catch-22, which really does deserve its reputation as the most inventive comic novel of the twentieth century. Turning to slightly more recent books, I was able to catch up on such disparate works as The English Patient, Cloud Atlas, and The Time Traveler’s Wife, all of which I admired. Of these, the two that retain the strongest hold on my imagination are John Crowley’s Little, Big, despite my mixed feelings on reading it for the first time, and J.M. Coetzee’s Disgrace, which strikes me as one of the most perfect of all recent novels. More disappointing were London Fields, Updike’s Terrorist, and, somewhat to my surprise, A Confederacy of Dunces, which I found clumsy and only intermittently engaging, despite its reputation as a classic.

Of books published in the last few years, my reading consisted mostly of nonfiction, despite my nagging resolve to read more contemporary novels. I greatly enjoyed The Immortal Life of Henrietta Lacks by Rebecca Skloot, which is a model of both popular science and investigative journalism. Like everybody else, I bought and read Steve Jobs by Walter Isaacson, which is short on analysis but long on fascination—more a gold mine of material than a real portrait, but still an essential document. I read The New York Regional Mormon Singles Halloween Dance by Elna Baker partly as background material for my novel, but was ultimately won over by Baker’s genuine wit and candor—it’s one of the funniest books I’ve read in a long time. And although The Possessed by Elif Batuman was a little thin, like a selection of essays in search of a theme, it made me curious to see what she’ll do next, given a more substantial project.

As for the coming year, as before, I expect that most of my time will be spent on background reading and research. Still, I have a few other authors I’ve been meaning to try. I’m going to read DeLillo for the first time, probably starting with Underworld, and then the later Philip Roth, beginning with American Pastoral. If I’m feeling really ambitious, I’ll tackle Faulkner, Morrison, and Perec’s Life: A User’s Manual as well. Above all else, I’m going to make a concerted effort to read more contemporary fiction. A glance at the bookshelves in the next room—the property of my wife, who is a much better reader than I am—reveals such titles as A Visit From the Goon Squad, Swamplandia!, and The Magicians, all of which have been beckoning to me for some time now. These days, of course, even my leisure reading has something mercenary about it, as I look for tricks and techniques to borrow or steal. As the year goes on, then, I hope to have a chance to talk more about these books, and if all goes well, I’ll have a few useful things to share, too.

“A multitude of drops”: Thoughts on Cloud Atlas

While I was in Los Angeles over the weekend, I finally finished David Mitchell’s Cloud Atlas, which over the past few years has gradually emerged as a consensus choice for one of the major novels of the decade. It first gained critical attention, and a fervent cult following, both for its striking structure—six nested novelettes, arranged like a series of Russian dolls, with each story commenting obliquely on its predecessor—and the virtuosity of Mitchell’s language and command of genre, which ranges from  dystopian science fiction to thriller to period pastiche. And while I do have some mild reservations about the novel, which is probably unavoidable for book that pushes the envelope so consistently, there’s no doubt that Mitchell is a formidable talent, and an author I’m looking forward to reading for years to come.

The element of Cloud Atlas that I enjoyed most, surprisingly, was its commitment to genre and plot. Despite what other critics have said, I don’t think the tone of the individual stories ever degenerates into simple parody, not even in “Half-Lives: The First Luisa Rey Mystery,” which some have called a satire of an airport novel. For my own part, I feel that Mitchell loves and respects his sources too much to dismiss them so easily. If anything, “Half-Lives” reads more like a tribute to The Crying of Lot 49 by way of a 70s thriller, which turns out to be a surprisingly heady combination, even if it’s likely to be underrated simply because it’s so readable. The same is true of “An Orison of Sonmi-451,” which, until its rather predictable closing twist, qualifies as a near-great science fiction novella, and the work of someone who clearly has great affection for the form.

That said, I also suspect that Mitchell’s fondness for the genres he’s inhabiting prevents him from weaving the novel together more tightly. A work like Cloud Atlas needs to walk a fine line between seeming too tidy or contrived and spinning apart into its separate components, and I think it strays a bit too close to the latter: I wanted more resonance, more jangling, between the constituent parts of the story, and the connections seemed either too obvious (the birthmark that all but one of the central characters share, with the implication that they are reincarnations of the same soul, perhaps on its way to Bodhisattvadom) or nonexistent (as in “The Ghastly Ordeal of Timothy Cavendish,” a segment which, while fun to read, fails to justify its presence). Obviously, this is a matter of taste. But I can’t help but thinking that a writer like John Barth or Nabokov would have given us a more elegantly structured edifice, even if it might have been less true to the genres of the stories themselves.

Still, it’s only been a few days since I finished the novel, and on going back, I’ve already begun to appreciate some of Mitchell’s more subtle associations. I also have a feeling that this is a novel that will gain much on rereading, which is something I plan to do fairly soon—certainly before the movie version comes out. The adaptation that has been announced, with Tom Tykwer and the Wachowskis directing, seems utterly unnecessary: if anything, it should be an anthology piece, with a different director tackling each segment, or at least a virtuoso acting challenge, with the same actors tackling roles in various time periods. Neither, it seems, it going to happen, which makes me skeptical about the outcome. No matter the result, though, we’ll still have Mitchell’s novel, with its richness, its ambition, and its only occasional lapses into tedium or obviousness. It’s a startling hydra of a book, and seems likely to endure for a long time.

What I read when I’m writing

When Norman Mailer was writing The Naked and the Dead, the novel that made him famous at age twenty-five, he had a simple method for getting ready to work in the morning. Mailer says:

I had four books on my desk all the time I was writing: Anna Karenina, Of Time and the River, U.S.A., and Studs Lonigan. And whenever I wanted to get in the mood to write I’d read one of them. The atmosphere of The Naked and the Dead, the overspirit, is Tolstoyan; the rococo comes out of Dos Passos; the fundamental, slogging style from Farrell, and the occasional overrich descriptions from Wolfe.

And Mailer isn’t the only writer who kept a few favorite books on his desk. I imagine that many novelists have books that they keep at the ready for when they feel inspiration starting to flag. Sometimes it’s the same book over the course of an entire career; more often, I suspect, it varies from project to project. In my own case, I start each writing day by reading a few pages of a book that embodies the tone or voice I’m trying to achieve—as if something of the author’s talent will magically transmit itself—and return to it more than once as I continue to work. And rather to my surprise, when it comes to the novels I read while writing, I find myself sticking to a limited, strictly defined circle of books.

As I mentioned yesterday, I generally do a fast, rough draft of an entire chapter first thing in the morning, which usually takes a couple of hours. I’ve found from experience that the best books to read while I’m doing that messy initial version are rich, ripe, stylistically powerful books that encourage my own writing to be a little more florid—qualities that I pare down relentlessly in subsequent revisions, but which are often good to have in a first draft, where the point is to get as many ideas or images onto the page as possible. For me, the ideal author for this purpose is John Updike. Our styles as writers couldn’t be more different, but something in his ornate sentences just puts my brain to work. (It’s the Heist school of writing: I imagine a writer better than I am, then figure out what he would do.)

At the moment, then, I’m starting each morning with a few pages of Updike’s Terrorist. Later in the day, though, when I’m polishing what I’ve already written, I feel that it’s a mistake to read something so dense and mannered, because I run the risk of ending up with mere self-indulgence (a quality to which even Updike himself isn’t immune). For later drafts, it’s better to go with an author whose prose is a little more restrained, clean, and elegant—someone like Ian McEwan, say. While writing Kamera, I worked my way through Atonement, Amsterdam, Saturday, The Innocent, Black Dogs, and The Comfort of Strangers. Right now, since I’ve already read all the McEwan I own, I’m doing something similar with Michael Ondaatje’s The English Patient, which has the kind of spare, classical style that I’m hoping will restrain the worst of my impulses.

There’s a negative side to all this, too. While I’m writing, I avoid books that I think will noticeably infect my style, for better or worse. This includes bad books, of course, but also good novels where the author’s style clashes with mine. I also try to avoid books in translation, reasoning that it’s better to read books by great stylists who originally wrote in my own language. The problem? Since I’m always writing, my reading for the past few years has been extremely constrained. I haven’t read Cloud Atlas, for instance, because I’m afraid of being overly influenced by it, and because I don’t want to read anything in translation, I haven’t gotten around to Mario Vargas Llosa, among many others.

Obviously, this state of affairs can’t stand: as much as I like Updike and McEwan, I don’t want to be stuck with them for the rest of my life. And reading and being influenced by radically different authors is an important part of growing as a novelist. At some point, then, I’ll probably need to rethink this approach. (Although not until I finish this draft.)