Thoughts on an author’s commentary

I don’t know how many authors really enjoy reading their own books. In my case, whenever I start a project, I tend to operate under the illusion that it’s a novel I’m going to enjoy reading for my own pleasure. I write stories that I’d like to see, and ones that I don’t think other authors are writing—not necessarily because they’re so original, but because they’ve been tuned to reflect a certain balance of plot, action, and concept that I happen to like. By the time I’m done, though, I’ve spent so long in the story that I can barely see the words on the page, and going back to read it again has about as much appeal as eating the same dinner for three days in a row. Still, there are times when I’ve reread my work and been sucked in despite myself. Short stories like “Kawataro” or “The Boneless One” tend to hold up better, since they were often the product of a few weeks of work, rather than months or years. I can flip through one of my stories and get lost again in the plot, while a novel is usually inseparable from the memory of the year or so it occupied in my life.

As a result, it was a little strange to sit down to confront a book that used to be a part of me, only to be set aside and revisited long after the fact. Of course, when I started my author’s commentary for The Icon Thief, I didn’t know what the experience would be like. At first, I saw it primarily as a way for me to get some things off my chest: the book is loaded with references, hints, and little jokes, and I wanted to document those elements so they wouldn’t be lost forever. I also wanted to talk about the decisions I made along the way, showing how logical—or otherwise—my choices had been and how many other directions a story can take. Once a novel is published, it starts to seem inevitable, but really, a text covers up countless alternatives, like the layers of the paintings in The Mystery of Picasso, and I wanted to strip away the surface to uncover some of those discarded variants. Finally, I hoped that the book would provide convenient illustrations and examples for some of the topics I wanted to cover here: plot, structure, handling action and exposition, creating suspense.

In the end, it didn’t turn out quite as I intended. I could have simply listed off the references and my own cute stories about how each chapter was written, but I quickly figured out that this wasn’t particularly interesting, especially because many of the visitors to this blog haven’t read the book. It also felt a little self-indulgent, and not in a good way. There’s a reason, I discovered, why most authors leave commentary and interpretation to the critics, even if they get most of it wrong: a novel is a living thing, capable of multiple interpretations, and it can be a mistake for the author to explain what he was thinking in too much detail. Early on, then, I made a conscious effort to frame each post in terms of a larger writing topic, using the chapter as necessary to illuminate the points I was making. Sometimes the chapter itself wouldn’t be mentioned until the last paragraph, if at all, and I rarely did close analysis of the text. (This is partially because I knew I’d only end up wishing that I’d written it differently, and also because I didn’t want to kill the novel for myself any more than I already have.)

The result ended up being more of a monster than I expected. A director’s audio track can be tossed off in a couple of hours, but this commentary took close to sixteen months and something like 30,000 words, or a third of the novel itself. But the most valuable thing about it, at least for me, lay in how it clarified my feelings toward the novel itself. The Icon Thief was my first published book, and like all debut novels, there are things about it that its author would like to change: I’d give more background on the art historical material, for instance, and I’d try to hold the reader’s hand a little more over the first hundred pages. But I’m surprised at how much I still like it. In some ways, it’s an odd, unwieldy novel, and it never fit comfortably into any one category: the cover makes it look like a Da Vinci Code knockoff, but its heart lies with the dry, slightly frigid skepticism of Marcel Duchamp, even as it tries to incorporate big action set pieces and the elements of a crime procedural and international thriller. In the end, though, I think it all sort of works. And that’s really the most any author can ask of his own novel.

“He found himself studying her face…”

Note: This post is the sixty-first and final installment in my author’s commentary for The Icon Thief, covering the epilogue. You can read the earlier installments here. Major spoilers follow for the ending of the novel.)

You can tell a lot about a writer by the way he or she approaches endings. Some novelists, like Stephen King, prefer to dive into a story without knowing how it ends, which allows the action to unfold more organically—and also leaves you with the possibility, which we often see in King, of a rousing, suspenseful story that peters out in a vast anticlimax. Others prefer to have a specific ending in sight, or even to work backward from a conclusion, as John Irving says to The Paris Review: “I love plot, and how can you plot a novel if you don’t know the ending first? How do you know how to introduce a character if you don’t know how he ends up?” My own approach, as in most things, involves trying to have it both ways. I generally start with a decent sense of where the story is going, but I postpone any detailed outlining until I’m ready to begin the last fifty pages or so. With The Icon Thief, I figured out the ending pretty quickly, and it remained virtually unchanged throughout more than a year of rewrites. And then, less than a month before we were scheduled to go out to publishers, I changed it.

The original ending tracks the existing epilogue fairly closely until the final page, although there are a number of important differences. My first version was told from the point of view of Vasylenko, a character we haven’t seen except in passing, as he meets with Lermontov—now on the run—to discuss the latter’s move from London to Moscow. The two men visit the British Museum, where David’s Death of Marat is conveniently on loan from Brussels, then head to Vasylenko’s home in Fulham. Ilya is waiting for them there. And although we suspect that he’s there to kill them both, he’s really working to extract a confession with Powell, who is listening on a wire as he waits outside to make the arrests. Ilya leaves the two men to the police, throws his gun into the Thames, and walks away, apparently liberated at last. (Incidentally, I opened the scene with The Death of Marat mostly because I wanted to discuss an ingenious theory about the painting that I first encountered in Why Are Our Pictures Puzzles? by James Elkins, although it may also have been an unconscious homage to The Eight by Katherine Neville, who, three years later, would go on to provide the cover blurb for Eternal Empire.)

As endings go, I thought it was pretty good, even if the final beat owed a lot to the last scene of Michael Clayton—a movie I’ve raided for inspiration more than once. Later, though, after the rest of the novel had been revised, I found that the ending no longer worked. The greatest single change to the plot, as I’ve mentioned before, was to have Ethan die at Lermontov’s hands. Once that change had been made, the dynamic of the ending, a hundred pages later, was all wrong. Lermontov had to face the consequences for killing such an important character, and the one who most deserved to take revenge was Maddy. I don’t think I realized this right away; it was more an intuitive sense that the balance of the conclusion was flawed. Once I figured this out, the logic of the scene was fairly straightforward, and I wrote it in less than a day. The revised version is told from Lermontov’s point of view, an important fix, and now it’s Maddy who is working with Ilya to tie off the loose ends. Justifying her involvement required a bit of thought, and I’m still proud of my solution, in which Maddy is able to track Lermontov down based on his purchase of an unusual picture frame from the House of Heydenryk, the owner of which later contacted me to thank me for mentioning his company in such a positive light.

Strangely enough, this radically altered ending, which changes the dynamic of Maddy’s entire journey as a character, didn’t require a great deal of revision for the rest of the novel, although obviously scenes that read one way in the original version acquire a different meaning now. But that small decision ended up affecting the books that followed in fundamental respects. When I wrote The Icon Thief, I wasn’t thinking in terms of a series, and I was content to leave Maddy where we last see her—compromised to some extent by her revenge, yes, but also free to make a life for herself in a way that Ilya is not. Even after I knew that I’d be writing at least one more book with some of the same characters, I wanted to keep Maddy out of it, and she doesn’t appear at all in City of Exiles. (If nothing else, I felt that she deserved a break.) Much later, though, I began to see that her story wasn’t finished, and I found myself curious to see where she ended up after Ilya left her alone on that street in Fulham. The result was Eternal Empire, which in some ways was an attempt to work out some of the implications of Maddy’s last, fateful decision. And the answers I found weren’t always what I expected…

This is the last installment in my author’s commentary for The Icon Thief, which I began over a year ago. Tomorrow, I’ll be looking back over the experience and reflecting a bit about what I’ve learned along the way.

“It’s always the other ones who die…”

Note: This post is the sixtieth installment in my author’s commentary for The Icon Thief, covering Chapter 59. You can read the earlier installments here. Major spoilers follow for the ending of the novel.)

For some reason, my novels tend to end in hospital rooms. Both The Icon Thief and City of Exiles conclude with one character paying a visit to another in recovery, during which most of the unanswered questions in the story are addressed. To some extent, this is an artifact of the way these novels are constructed: the penultimate chapters tend to be heavy on action, with the players sustaining a certain amount of damage, and there isn’t a lot of time in the heat of events to resolve any of the plot’s remaining mysteries. And although it’s best for a novel to end as soon after the climax as possible, there’s also room for a bit of falling action and consolidation. Practically speaking, of course, these scenes should be as short as possible, a rule that I’ve followed fairly well in two out of three novels—I think the hospital visit in City of Exiles runs a little long. (If I’m being honest, I should also confess that I’ve been influenced by the final chapter in Red Dragon, which uses an important character’s recuperation in the hospital to tie off a number of crucial plot points.)

The last chapter of The Icon Thief, not counting the epilogue, has to walk a particularly fine line. Powell’s final speech to Maddy, who is recovering in the hospital after the events at the Philadelphia Museum of Art, essentially tears down the entire novel: the Rosicrucians don’t exist, Maddy and Ethan were suffering from a chemically induced paranoia, and she broke into the installation for no rational reason. I know that this revelation troubled some readers, along with at least one editor, who expressed an interest in acquiring the novel if I could rewrite the ending so that the Rosicrucians were real—which would have meant turning it into another story entirely. Still, in order for this to seem like anything but an enormous cheat, I had to put something inside Étant Donnés for Maddy to find, but it had to play by the rules. It had to be plausible, consistent with what I knew about Duchamp, expressive of the novel’s themes, and evocative enough to compensate the reader for the extended trick the novel had played. And for most of the writing process, I had no idea what this object would be.

My memory of when the answer hit me is oddly specific. I was standing in the appraisal line at the Strand in New York, waiting to resell a few used books, when it occurred to me that Maddy could find a pawn from the chess set that Duchamp had carved for himself in Buenos Aires. The pawn—which can be seen here—seemed like an ideal object for a number of reasons. It was small and easily concealable, so it could have remained unnoticed in Étant Donnés for decades and also lie clutched in Maddy’s hand, unseen, after her departure from the museum. It was symbolically resonant, yet nonspecific, so it could sustain any number of readings. And it tied in with many of the novel’s themes and touchstones: chess, of course, but also Through the Looking-Glass, with its sense of entering a strange world, a mirror image of our own, in which a pawn that makes it to the end of the board can become a queen. (Incidentally, the pawn may not have been Duchamp’s own handiwork: he seems to have carved the chessmen himself, but left the pawns to be turned by a local woodcarver, a technicality that I didn’t think was worth mentioning.)

If this novel has one message, it’s that when all is said and done, it’s enough to survive. As I’ve said before, I’m drawn to conspiracy fiction because it seems to get at something close to the heart of how we experience the world. We’re always telling stories to ourselves about history and our own lives, and we have a tendency to find patterns that aren’t really there. If Maddy’s journey means anything, it’s because the Rosicrucians were secretly her way of dealing with her own failures and disappointments: it’s easier to accept life’s reverses if we sense that there’s a guiding hand, even a sinister one, controlling it behind the scenes. The pawn reminds us that there’s a dignity in simply making it across the board, even if the contest itself lacks any logic, like the moves in Carroll’s looking-glass game. And in the original draft, I had intended to leave Maddy here. Later, of course, the story took a turn that I hadn’t anticipated. Next week, I’ll be finishing up this commentary with a look at the epilogue, in which we discover that Maddy’s story is far from over…

“Before her stood the wooden door…”

Note: This post is the fifty-ninth installment in my author’s commentary for The Icon Thief, covering Chapter 58. You can read the earlier installments here.)

One piece of advice I’ve learned to share with aspiring writers is that if you aren’t sure how to end a story, take the scene you like best—the one you’re absolutely dying to write—and restructure the plot so that it serves as your climax. This may take a bit of tinkering, since you’ll often be tempted to put the big scene as early as possible, if only because you know you’ll actually get to write it. Really, though, endings count for so much that you need to save the best for last. A reader’s opinion of a story will largely turn on how satisfied he or she is by how it concludes, and a novel that unfolds beautifully for three hundred pages won’t survive a failure of nerve in the last thirty. In the case of my own novels, I usually know what the ending will be, at least in general terms, soon after I get the initial idea. The process of writing a novel is so uncertain and unpredictable that it helps to have a destination in mind: when I’m stranded in second-act problems and trying to get out of a jam, it helps to know that I have an ending that will work if I can manage to bring it off.

Of course, it’s one thing to know in broad strokes what the climax will be, and quite another to put it into narrative form. For The Icon Thief, as I’ve noted before, I knew that the novel would end with Maddy breaking into the installation of Étant Donnés at the Philadelphia Museum of Art, but I didn’t know what she would find there; I only knew that there had to be something, or else the entire story would seem like one long cheat. I also didn’t know how that moment would tie in with the machinery of the larger plot. Twenty pages from the end, I still had a lot of material to tie off, and for the sake of narrative momentum, I knew that I’d have to stage what followed—Sharkovsky’s attempt on Maddy’s life, Ilya coming to the rescue, and Ilya’s final escape—as close to the installation itself as possible. Trying to cover all of this in a way that seemed surprising and logical within the considerable constraints that the location presented was a real headache, and it took me a long time to make it work.

In the end, as usual, it was the location itself that provided the answers I needed, and it wasn’t until I spent a few hours at the museum, repeatedly walking over the same ground, that the pieces fell into place. And I’m still proud of much of what happens here. I like the little MacGyver trick, involving a fragment of a porcelain spark plug, that Maddy uses to get past the tempered glass in the installation. The moment when Sharkovsky—and the reader—thinks that he’s killed Maddy, only to realize that he shot the dummy inside by mistake, may stretch credulity a bit, but I enjoyed the effect so much that I kept it in. And Ilya’s final escape through the window in the Duchamp gallery, which I told you we’d see again, is a nice touch of badassery. (This moment, incidentally, involves one of the novel’s few intentional cheats: I don’t think it would actually be possible for Ilya to escape through this window, which is made of bulletproof glass, in the manner in which he does here. By the time I realized this, though, I’d already written the scene, and after some thought, I decided to let it stand, with a nod to the rule of cool.)

The result is the single longest chapter in the novel, as well as one of the few that switches between multiple perspectives, cycling from Sharkovsky to Maddy to Ilya. I hope it feels like a satisfying conclusion; it’s certainly one of the few chapters that I can read again for my own pleasure as if it had been written by someone else. But the passage that sticks with me the most is the final beat between Maddy and Ilya, in which she silently asks him to spare Sharkovsky’s life. It’s an important moment for both of them: it conveys the essential difference between these two characters, points a way forward for Ilya to leave behind his violent past, and lays the groundwork for the epilogue’s closing twist. And we’ll revisit this moment again. At the climax of Eternal Empire, the final novel in the trilogy, I harken back to it, but both Maddy and Ilya have charged a great deal in the meantime. And it’s not until then, at the very end of the series, that we understand what that exchange of glances really meant…

“But something else was involved…”

Note: This post is the fifty-seventh installment in my author’s commentary for The Icon Thief, covering Chapter 56. You can read the earlier installments here.)

When I first began working on The Icon Thief, I knew that I’d be walking a fine line. For reasons that I’ve discussed before, I’ve always been drawn to conspiracy fiction, but it’s generally been of the skeptical variety, like Foucault’s Pendulum or The Illuminatus Trilogy, which raises as many questions as it answers. In many cases, the conspiracy at the heart of the novel is revealed to be the product of the protagonist’s imagination or paranoia, and even seemingly unambiguous events can be read in any number of ways. This kind of thing can sometimes feel like a tease for the reader—or an attempt to have your cake and eat it, too—but I think it’s intuitively closer to how I suspect the world really works: the answers we seek aren’t always straightforward, our preconceptions shape what we choose to see, and the search for a overarching explanation can ultimately turn into a perverse form of idealism. So while I didn’t think I’d be capable of writing a straight conspiracy thriller in the Dan Brown manner, I wanted to retain as many of the pleasures of the genre as possible while breaking it down as gently as I could.

What I realized early on, as my notes from the period indicate, was that I essentially had to construct three different plots for the same novel. The first plot would be a conspiratorial fantasy that would allow me to indulge in my love of historical arcana: the Rosicrucians, the Bolsheviks, the Vehmgericht, Acéphale, the Black Dahlia murder, the intersection of art and the occult at Monte Verità, and more, all centered on the mysterious figure of Marcel Duchamp. Lying beneath it would be a more skeptical reading that would explain away the intricate web of conspiracy I’d constructed—accurately enough, I might add—as a product of coincidence, overinterpretation, and misguided ingenuity. Finally, and most crucially, would be the real conspiracy, one that would frame the events of the story in more realistic terms, but which would also be striking and compelling in its own right. (Attentive readers will notice that this is basically the structure that Umberto Eco employs in Foucault’s Pendulum, although he does it at much greater length and with several additional layers of deception and interpretation.)

In the end, this final level of reality ended up revolving around the looting and disappearance of fine art in Europe at the end of World War II, a story fascinating enough to drive an entire thriller in its own right. My primary sources here were the nonfiction works The Rape of Europa and The Lost Museum, the latter of which was where I first heard the story of the art collector Paul Rosenberg, whose collection may well have ended up in a secret warehouse run by Russian intelligence. Chapter 56 of The Icon Thief, in which the true outlines of the plot are laid out at last, is one of my favorite chapters in the entire novel, and one of the few that I can still read happily for my own pleasure. My only quibble with it is that, yes, Reynard does confess to his role awfully quickly, but as I’ve said elsewhere, sometimes you just need to get on with the plot. (If you’re really interested in trivia, I can reveal here that the history of my fictional Study for Étant Donnés, as well as the description of its provenance markings, is based on Courbet’s Nude Reclining by the Sea, which hangs in an adjacent wing to the Marcel Duchamp gallery at the Philadelphia Museum of Art.)

Looking back, I’m happy with the triple plot I constructed, but I’m also aware that this approach may have cost me a few fans. Readers who weren’t interested in conspiracy fiction at all might have taken one look at the cover of the novel—which certainly looks a lot like a Dan Brown knockoff—and concluded that it wasn’t for them, while those who were looking for a straightforward conspiracy thriller might have felt cheated by the revelation that much of the book is a mislead. I tried my hardest to construct a story that struck a happy medium, although I’m aware that such a strategy always leaves readers hanging to either side. But I’m not sure I had much of a choice. The moment I decided to base my novel on Duchamp, the ultimate skeptic, I knew that I had to honor his refusal to be confined to any one interpretation, however colorful or intriguing it might be. I can’t say that I know how Duchamp himself would have reacted to the uses to which I put his work, but I’d like to think that he’d at least be amused. And I hope he’d be willing to forgive me for what I’m about to do to him next…

“A message here that she was supposed to see…”

Note: This post is the fifty-sixth installment in my author’s commentary for The Icon Thief, covering Chapter 55. You can read the earlier installments here.)

I’m proud of the novels and stories I’ve published, but if they all have one limitation, it’s that they aren’t comfortable with the idea of a story—or even a scene—in which nothing much happens. This isn’t to say that they’re packed exclusively with action: even Eternal Empire, which probably has more straightforward action scenes than the previous two novels combined, takes a little while to ramp up. But in most respects, there’s always something happening in these books. There are plot points to cover, information to convey to the reader, characters who need to get from point A to point B. The plots are invariably complicated, and most of them were cut down considerably from their original length, which means that each page carries more than its share of story. This is entirely intentional: I like dense, layered novels, and I enjoy seeing how far I can push complexity within the bounds of the genre I’ve chosen. But it’s still an approach that limits the kinds of stories I can tell or moods I can evoke. And although I’m well aware of this, I’m still some distance away from being comfortable with scaling it down.

In the entry on Yasujirō Ozu in his Biographical Dictionary of Film, David Thomson diagnoses this tendency in a beautiful passage I’ve quoted here before:

[S]o many American films are pledged to the energy that “breaks out.” Our stories promote the hope of escape, of beginning again, of beneficial disruptions. One can see that energy—hopeful, and often damaging, but always romantic—in films as diverse as The Searchers, Citizen Kane, Mr. Smith Goes to Washington, Run of the Arrow, Rebel Without a Cause, Vertigo, Bonnie and Clyde, Greed, and The Fountainhead. No matter how such stories end, explosive energy is endorsed…Our films are spirals of wish fulfillment, pleas for envy, the hustle to get on with the pursuit of happiness.

Needless to say, that’s an impressive list of movies, and many of our best recent films—from the work of Christopher Nolan to Pixar—have been predicated on similar principles. Yet it’s often the stories that find time for silence and emptiness that linger the most in the reader’s imagination, and you’d have trouble finding a truly empty moment in any of my novels.

Well, maybe there’s one. In Chapter 55 of The Icon Thief, after Maddy arrives in Philadelphia, she walks across the bridge and heads for the museum on foot. It’s probably the least eventful chapter in the entire novel: there’s a tiny bit of plot, as she stops into a hardware store to pick up the items she’ll need to break into Étant Donnés, but for the most part, we’re alone with Maddy and her thoughts. And I like the result a lot. It’s based closely on my own visit to the city, in which I followed a route much like the one Maddy walks here, taking notes as I went. (Although the hardware store is a fictional one, introduced after I arrived at the museum and realized how difficult breaking into the installation would actually be.) It’s one of my favorite memories from writing this novel, especially for the moment when I paused outside the museum, taking in its layout, and noticed the same curious detail that Maddy does. The museum is laid out symmetrically, with identical east and west wings, with only one anomalous element: a single glass pane, with no corresponding window on the other side, that looks into the gallery devoted to Marcel Duchamp.

It’s the last really calm scene in the entire novel, as we prepare to enter its closing sequence of revelations and confrontations. Even here, though, the machinery of the plot isn’t entirely out of sight, and it’s likely that I felt justified in indulging in a quiet moment here because I knew—as does the reader—that Sharkovsky is waiting to follow Maddy into the museum as soon as she arrives. And the more I look at this chapter, the more it seems to hint at a way forward for the rest of my work. If the reader accepts the scene, it’s because the silence is charged with a form of anticipation, one that wouldn’t be there if it weren’t for the previous events of this very busy novel. Finding the right balance between activity and stillness is a narrative problem that I still haven’t cracked, for all my thoughts about craft, but for a few pages, I feel as if I got it right. At the time, of course, I wasn’t thinking in those terms: I just wanted to pause and focus on the location where the climactic action of the novel would take place, making sure that the reader, along with Maddy, noticed the window of the Marcel Duchamp gallery. We’ll be seeing that window again…

“He saw a word in his mind’s eye…”

Note: This post is the fifty-fifth installment in my author’s commentary for The Icon Thief, covering Chapter 54. You can read the earlier installments here.)

I still have the notebook page on which I began jotting down ideas for the novel that eventually became The Icon Thief. At that point, I hadn’t yet acquired the good habit of dating all my notes, but from context, I seem to have started work on the story just over five years ago. As a result, the page is a time capsule of both my thoughts while roughing out the novel and my writing process at the time. In most ways, my approach hasn’t changed all that much, and the ideas I sketched out here are surprisingly close to what the novel eventually became. Here’s a sample:

Three levels of plot: supposed order, alleged order, and real order.
Themes: paranoia, information overload, vision/eyesight
New York art world; intersection between art + finance

This is followed by a long list of potential plot points or ideas from the real world, some of which ended up being crucial to the story that resulted, while others were eventually discarded. Neither Marcel Duchamp nor Étant Donnés are mentioned until the fourth page of notes, at which point I’d been working on the idea for several weeks. And on the top of the first page is the title of the novel I had in mind: Camera.

Regular readers of this blog will know that I’ve rarely had much luck with the initial titles of my novels, and they’ve invariably had to be changed, usually with only a few weeks left before delivery of the manuscript. I have a weakness for opaque, suggestive titles that have more than one meaning, while publishers tend to be happier with titles that clearly signal what the book is going to be about. Consequently, I need to walk a fine line, and I’m very happy with the title The Icon Thief. Still, I do miss Camera. I’d wanted to write a novel with that name for a long time, although the original conception survives only as a shadow in what was eventually published: the initial plot, which I now think was probably too ambitious for my talents at the time, was about a man piecing together the reasons behind a loved one’s suicide by looking at the photos on her camera after her death. There’s a touch of this in the finished novel, as Maddy tries to figure out the clues that Ethan left behind, and my research into the life of the photographer Diane Arbus informed some elements of Maddy’s personality. Really, though, its presence in the story is more like that of a ghost, or a double exposure, dimly visible behind the plot’s convolutions.

The other inspiration for the title was “Camera” by R.E.M., possibly their saddest and most mysterious song, and one with a particularly haunting backstory. When I wrote that title at the top of the page, I didn’t have a plot in mind yet, but I certainly had a tone I wanted to capture, as well as a handful of themes that had always fascinated me: the gap between what we see and what we think we understand, the tendency for images to be misinterpreted, and the ambiguity of the photographic medium itself. These themes were radically transformed in the final product, and perhaps that’s the way it should be. But the working title achieved its purpose. It allowed me to focus my thinking, emphasizing some themes in preference to others, and at one crucial point, it also informed me that I was on the right track. Early in the process, I realized that the Russian chemical warfare program would be a part of the story, since it allowed me to unite several key themes—conspiracy, paranoia, Russia itself—into one convenient thread. And I still remember the strange thrill I felt when I learned that Laboratory 12, the notorious poison laboratory of the secret services, had also been known as Kamera.

Kamera, then, was the title under which the novel went out to publishers, and that’s how it was sold. And it’s instructive, at least to me, to go back over the story to see how it reads with its original title in mind. An ambiguous title is a sort of clue to the reader, a hint to keep an eye out for information that might otherwise seem unimportant, and in that light, a sequence like that of Chapter 54 would read altogether differently. We’ve already witnessed the end of Anzor Archvadze, dying in the hospital with a case of toxic epidermal necrolysis and barely managing to force out his last words: “Camera. Camera.” It’s not until several chapters later that Powell sees the words for what they really are. In the novel as it stands, it’s a good scene, but it would have been even better in the original version, as the true meaning of the title locked into place. Kamera, of course, means chamber, so the working title served triple duty: it was meant to evoke the poison program, the various roles that cameras and photographs play in the narrative, and the chamber of Étant Donnés itself. All this was lost in the final version. And although I’m mostly pleased by the way it turned out, I can’t help but miss what was there before…

“Tzaddikim knew how to be patient…”

Note: This post is the fifty-fourth installment in my author’s commentary for The Icon Thief, covering Chapter 53. You can read the earlier installments here.)

One of the first things a writer needs to realize is that it can be a mistake to base a character too closely on yourself, or to have the plot of a novel track literal events from your own life. Part of this lies in the importance of detachment: when you’re writing about yourself—or a thinly disguised surrogate—or relating incidents that really happened, it can be hard to maintain the necessary objectivity. A reader who doesn’t know you personally can’t be expected to take the same interest in the details of your inner life, at least not before the material has been refined and rethought, and it’s easier to do this when you depart enough from the facts to make their implications seem new. Much of the creative process consists in searching for metaphors or analogies for your own experience, which allow you to deal with what concerns you while regarding the result with a clear eye. In The Spooky Art, for instance, Norman Mailer advised young New York writers to deal with the events of 9/11 indirectly, keeping the emotional core while shifting it into another context:

There must be five hundred young writers in New York who had a day of experience that was incomparable—nothing remotely like that had ever happened before in their lives. And it’s likely that some extraordinary work will come out of it. Hopefully, not all of it about 9/11. If you never write about 9/11 but were in the vicinity that day, you could conceivably, in time to come, describe a battle in a medieval war and provide a real sense of such a lost event. You could do a horror tale or an account of a plague. Or write about the sudden death of a beloved. Or a march of refugees. All kinds of scenes and situations can derive ultimately from 9/11. What won’t always work is to go after it directly. That kind of writing can be exhausted quickly. And the temptation to drive in head-on is, of course, immense—the event was traumatic for so many.

And an experience doesn’t need to be traumatic to lend itself to fictional transmutation. Nearly every choice I’ve made as an author—and writing is really just a series of choices—can be traced back to something in my own history or personality, transformed into something very different that still reflects its hidden origins.

Take Ilya’s religious background. I knew from early on that one of the primary characters in The Icon Thief would be Jewish, and it’s hard to think of any one decision that had a greater influence on the novels that followed: the art world and conspiracy elements that dominate the first installment are gradually toned down, but Ilya’s background and his ambivalence about the two sides of his personality—the Scythian and the Tzaddik—are central to the trilogy, and I don’t think the question is fully resolved until the last page of Eternal Empire. At first, like Wolfe’s Mormonism, this was a detail that I introduced almost at random, merely because I thought it seemed promising: I liked the idea of a hit man who read the Sefer Yetzirah, and I knew that it would allow me to bring in a lot of material that I’d always found interesting. And although these themes never quite come to the forefront of these novels, they’re always there in the background, providing insight into Ilya’s character and a kind of counterpoint to the main action, with its recurrent motifs of interpretation, history, and exile.

Most of all, it allowed me to approach aspects of my own inner life from an unexpected angle. If there’s one theme that I seem condemned to revisit endlessly in my own fiction, it’s the problem of interpretation, of how we find meaning in texts, stories, and the world around us. I’m not the first writer to be drawn to Jewish models as a lens for examining these issues: Borges, among others, has done more with this tradition than I ever could. Still, in creating Ilya, I found that I was inventing a figure who was oddly like myself, as different as we are in most external respects. Like me, he’s drawn to texts and traditions of exegesis, like the midrashim and the cabala, both because of the inherent beauty they possess and because they stand in contrast to what we can and can’t understand about the world around us. The world may be a text, but it pushes back against us in ways that we don’t encounter on the printed page, and just because we’re good at kind of interpretation doesn’t make us good at the other. Ilya’s struggle to come to terms with the way his world works, and with the contradictions of his own personality, gave me a way of dealing with my own. And as the novel draws to its climax, we’re about to find out who Ilya really is…

“It was two hours to Philadelphia…”

Note: This post is the fifty-third installment in my author’s commentary for The Icon Thief, covering Chapter 52. You can read the earlier installments here.)

Four years ago, I took a bus from New York to the Philadelphia Museum of Art to see Étant Donnés for the first time. I’ve mentioned before that I like to tackle each part of a novel in turn, focusing on researching, outlining, and writing one section at a time while leaving the rest relatively undefined. Here, though, I was taking this approach to its extreme. At that point, I’d already been working for more than nine months on the novel that would later become The Icon Thief, a story that depended enormously on interpretations of Marcel Duchamp’s enigmatic final masterpiece. I’d written solid drafts of Parts I and II, which spent a lot of time speculating on the work’s history and meaning, and I knew that Part III would climax at the doorway of Étant Donnés itself. Yet although I’d studied photographs and diagrams of the installation, read countless critical studies, and even paid an exorbitant amount of money for a reproduction of Duchamp’s original Manual of Assembly, I’d never gone to see it in person. There’s no particular reason for this; it’s only two hours by bus, and most sane writers would have made this pilgrimage early on, probably before a word of the rough draft had been written.

When I boarded the bus that day, then, it’s fair to say that I was in a state of considerable apprehension. I was excited about seeing the installation at last, but part of me also worried that I’d discover something during my trip that would ruin my plans for the novel’s conclusion. (As it happens, I did stumble across one inconvenient fact at the museum that forced me to rethink the logistics of the ending, but I’ll deal with story when I come to it.) But there was a reason I’d waited so long. I don’t generally talk about character and its creation in mystical terms: I spend a lot of time thinking about my characters, particularly about their objectives and specific actions, but I’m usually content to keep them at arm’s length—which in my case is probably a good thing. When I do research on location, for instance, I try to regard the scene through the eyes of the primary character, but I’m also thinking as a writer, noting down ideas, retracing my steps, and looking for useful details or bits of business. If I’d gone to Philadelphia earlier in the process, that’s the detached mode in which I’d have been operating, and it’s possible that I wouldn’t have been thinking of my characters at all.

On the day I finally did go, however, I was in a very different state. I’d been living with the character of Maddy Blume for a long time—even longer than I’d spent working on the novel itself—and I knew deep down that it was important for me to spend this last trip as much in her head as possible. I’ve noted elsewhere that every novel is secretly about the process of its own creation, and in this case, I had good reasons to identify myself with Maddy: we’d both been obsessed with Étant Donnés for a long time from a distance, and much of my own research process ended up in the novel itself, refracted through her point of view. She worked at a firm whose offices resembled those of my old company, she lived on my block in Brooklyn, and when I envisioned her violent struggle with Sharkovsky, I staged it to take place within inches of my own desk. As a result, it was easy for me to put myself in Maddy’s shoes. I wasn’t being stalked by a killer, but I was being followed by something equally insidious: an unfinished novel that I suspected would rise or fall based on what Maddy could do at that museum.

Not surprisingly, many of the small details in Chapter 51 of The Icon Thief—as well as many of the chapters that followed—reflect my experience that day. And in retrospect, I’m glad that I waited to go. One of the wisest pieces of advice on creativity I know comes from the great film editor Walter Murch, and it’s a point that I frequently repeat to myself:

Each stage leaves a residue of unsolved problems for the next stage—partly because the particular dilemma you’re facing cannot be solved in terms of the medium you’re working in right then…It would be deadly if you did solve all the problems in the script—you do not want to be asking for the gods’ help at every stage—because then everything subsequent would be a mechanical working out of an already established form…

In this case, the unsolved problem in the story happened to coincide with the mystery within the plot itself, which strikes me as a good way of attacking the conclusion of a novel that had previously been planned and outlined almost to a fault. Instead of approaching this trip with a writer’s objectivity, I was going to the museum, like Maddy, in a state of nervous anticipation. And neither Maddy or I knew what to expect…

“Then she saw that there was no way out…”

Note: This post is the fifty-second installment in my author’s commentary for The Icon Thief, covering Chapter 51. You can read the earlier installments here.)

In theory, a novel should unfold as neatly as a proof in mathematics, with the plot emerging from a sequence of logical objectives and actions arising from the protagonist’s central problem. In practice, of course, it isn’t quite as straightforward. A manuscript in progress is a complex system, with elements on the smallest level invisibly affecting the largest. An author will often start with a handful of scenes or moments he wants to write, structuring the rest of the story—including the motivations of the central characters—so the plot will advance along a path that he happens to find interesting. There’s nothing wrong with this: I imagine that nearly every book contains scenes that have less to do with rigorous narrative economy than with what the author feels like writing at the time. Usually, these preconceived goals change along the way as well, and the resulting plot is the product of an ongoing process of action and reaction. Writing a novel isn’t a straight line: it’s more of slalom. And in the end, if you’re lucky, you’ll get to where you want to arrive without falling down on the way.

The Icon Thief went through many radical transformations from its initial conception to its final form, but I knew from the very beginning that it would end with Maddy physically breaking through the door of Étant Donnés at the Philadelphia Museum of Art. If nothing else, this was a striking, memorable conclusion, and if there’s one thing I’ve learned from experience, it’s that if a really good ending suggests itself, you should do whatever you can do steer the story in that direction. And I generally won’t start serious work on a novel at all until I have a decent climax in mind. (In City of Exiles, I had two—the incident on Chigorin’s plane and the final chase in Helsinki—and Eternal Empire similarly builds to an ending that I’d roughed out on my very first page of notes.) The trouble was getting Maddy to that point in a way that would seem inevitable. The scene as written, which I’ll discuss in more detail within the next couple of months, is one that arouses strong reactions from readers: there’s no conventional violence, at least not yet, but to see a work of art desecrated in such a visceral way is hopefully a little shocking. But it wouldn’t work at all if the reasons behind it didn’t make sense.

Chapter 51 of the novel is eventful in its own right, but its real purpose lies in preparing the reader for the climax that will occur seven chapters later. As a result, the real challenge lay in the amount of ground it had to cover. In less than seven pages, I had to leave Maddy convinced that the conspiracy against her life was real; that the answer lay inside Étant Donnés; and that the only way to save herself was to go to the museum and see what was inside the installation with her own eyes. I also needed time for her to be attacked at home by Sharkovsky, fend him off, learn that Ethan was dead, and see Ilya watching as she fled her apartment. All these moments are important in themselves, but they’re really designed to propel her into the novel’s endgame. Whether or not it works is something that I’m hardly prepared to judge, but if nothing else, I’d say it achieves its purpose within the logic of the story, whether the reader believes in the Rosicrucian conspiracy or suspects that Maddy’s paranoia may have another cause. But if this succeeds, it’s only because I’ve taken pains as the author to stack the odds.

In constructing the beats of the scene itself, I was largely inspired by the climactic scene in Rear Window, in which Jimmy Stewart fights off an intruder using the tools of a photographer’s trade, a gimmick, as Hitchcock rightly observes in his interview with Francois Truffaut, that is really nothing more than canny screenwriting. And many of the other details—Maddy hiding in the closet, Sharkovsky seeing the burning cigarette—were consciously introduced an excuse for me to play with the toys that this kind of scene provides. The idea that Maddy would use replicas of Marcel Duchamp’s readymades—the shovel, the bottle rack—to fend off her attacker is a little cute, but I like it. The Icon Thief is a fairly cerebral novel, and when I look back at it now, I wonder if it might not be too clinical: I wouldn’t change it in any fundamental way, but there are times when I worry that its devotion to a clockwork plot gets in the way of more immediate pleasures. That’s why staging this knockdown brawl between Maddy and Sharkovsky was so satisfying. And although Sharkovsky is out of commission for now, they still have one last confrontation in store…

“The old man walked along the Avenue of the Americas…”

Note: This post is the fifty-first installment in my author’s commentary for The Icon Thief, covering Chapter 50. You can read the earlier installments here.)

Novels are curious creatures. Even if you think of yourself as a gardener, not an architect, in practice, a novel can’t be grown like a living thing: instead, it’s assembled out of pieces that might have been written and conceived weeks or months apart, even if they end up lying side by side in the finished draft. In general, the task of the writer is to make these seams as invisible as he can. One of my primary concerns, especially during the rewrite, is to structure the story so that it reads as much as possible like a single continuous piece of narrative. There may be a lot of chapters—and in my case, they often switch between different points of view—but if the plot has been properly constructed, the chapter breaks feel less like interruptions than necessary beats in the novel’s rhythm. Film editors know that any cut the viewer notices is a bad one, and at best, their work is invisible. For the most part, that’s a rule that writers should follow as well: unless you’re going for a particular experimental effect, the story should feel like it was cut from one piece of cloth, even if it was really stitched together from countless shreds and patches.

There’s one exception to the rule, though. In theory, there’s no reason why a novel has to be subdivided into more than one major section, and indeed, many excellent thrillers—the early novels of Thomas Harris, for instance—consist of an uninterrupted sequence of short chapters. A section break, as opposed to a chapter break, is practically a violent event in itself: instead of following the action smoothly to the next scene, the reader is confronted by a blank page, a Roman numeral, and a stark epigraph. But this can also be very useful. I’ve written before about the purely typographical impact of epigraphs and section breaks, which signal that a new phase of the narrative is beginning, and I’d argue that this is one place where the reader can and should be aware of a rift in the continuity. (For an example in a very different context, notice how Mitch Hurwitz still fades to white for the act breaks in the new season of Arrested Development. There’s no commercial break, but it’s still a nice way of punctuating a plot development and structuring the viewer’s attention.)

Such a powerful tool should be used sparingly, of course, and I often get a little impatient with novels that break the action into four or more sections—maybe because I’m so fond of the rule of three. It also requires a little more trouble than an ordinary chapter break. I’ve found that it helps to think of it less as a break than a hinge. A hinge, as this diagram helpfully points out, consists of three parts: the central pivot, vividly known as the knuckle, and a wing to either side to join the door and the frame. A section break operates in much the same way, and it often requires more than one chapter to set up correctly. If you look at the section breaks in The Icon Thief, City of Exiles, and the forthcoming Eternal Empire—especially the ones between Parts II and III of each book—you’ll see that they consist of about three chapters. One chapter consolidates the information from the section we’ve just read; another provides connective and expository material to set up the section break; and the last serves to propel the plot forward to the next stage of the game. And without the chapters that came before it, this final chapter wouldn’t work nearly as well.

Chapter 50 of The Icon Thief is best understood as the knuckle chapter of this sequence. It’s a relatively quiet scene, following Sharkovsky—whose point of view appears for the first time in the narrative—as he meets Lermontov and gets the order to kill Maddy. On the last page, it also reveals that Ilya has been keeping an eye on the art gallery, and it ends with him following Sharkovsky to his next destination. To borrow a term from television writing, the bulk of the chapter is concerned with laying pipe, and it’s interesting mostly in how it gives the reader the information necessary to understand the action to come. It also gives us some breathing room after the revelations of the chapter before it. (Originally, this scene included a big reveal of its own—the disclosure of Reynard’s true role in the story—but I cut it in a subsequent rewrite, sensing that it would work better at a later point in the novel.) In theory, you could take this chapter out and convey the essential points in some other way, but as the knuckle between the two parts of a hinge, it plays an important role. Without it, the door would fall out of its frame. And we’re about to see what lies on the other side…

I’ll be appearing on The Afternoon Shift with Niala Boodhoo on WBEZ Chicago Public Radio at 2:00 pm today to discuss the business of writing fiction. You can listen to the broadcast online here.

“Powell stared silently through the glass…”

(Note: This post is the fiftieth installment in my author’s commentary for The Icon Thief, covering Chapter 49. You can read the earlier installments here.)

For a certain kind of novelist, there’s an enormous temptation to base one’s characters on recognizable people, and many stories gain nearly all of their interest from the perception that they’re thinly veiled depictions of real public figures. As Dean Koontz points out in his dated but valuable book Writing Popular Fiction, works by the likes of Harold Robbins or Jacqueline Susann are compelling largely because we think we can guess who these rich, glamorous, oversexed characters are supposed to be, and we’re more likely to take the author’s portrait at face value precisely because the names have been changed: the novel implicitly promises to tell it like it is, without fear of libel, at least for readers who are clever enough to fit names to faces. Irving Wallace went even further, spelling out his sources in the text itself—and often on the back cover copy. As I’ve mentioned before, in a novel like The Plot, Wallace isn’t simply content to create a character based on Christine Keeler, but blandly tells us that her scandal was “ten times more exciting than the old Profumo affair.”

While this can be an effective fictional device, a lot of novelists resist it, and for good reason. Norman Mailer, in his afterword to Harlot’s Ghost, explains that his decision to incorporate real people into the narrative using their proper names arose from a desire to avoid this kind of phony authenticity:

It was obvious, therefore, that one would have to give Jack Kennedy his honest name…One could only strip him of his fictional magic by putting a false name on him; then the reader’s perception becomes no more than, “Oh, yes, President Fennerly is Jack Kennedy—now I will get to learn what made Jack Kennedy tick.”

As a result, Mailer uses the actual names of important characters like Howard Hunt, Allen Dulles, and Bill Harvey, knowing that the reader will naturally be more critical of how these men are portrayed, thinking, “That isn’t my idea of Howard Hunt at all.” And it’s also likely that Mailer, in writing in what amounts to an epic spy novel, was encouraged by the conventions of suspense fiction, in which real names are often used to give the action an air of verisimilitude. Frederick Forsyth, for example, populates his books with such historical figures as Kim Philby and Simon Wiesenthal, many of whom were still alive when these novels were written, allowing him to blur the line between fiction and reportage—which is a large part of his work’s appeal.

In The Icon Thief and its sequels, I’m operating in a similar mode, and I’ve occasionally run into the problem of whether or not to use the real names of living people. (I’m much less concerned about historical figures, whom I tend to name freely, even as I indulge in other forms of speculation or invention.) President Putin never appears directly in these books, but he’s frequently mentioned, and I decided long ago that it would be absurd to refer to him by any other name. I thought seriously about placing a real energy company at the center of the plot of City of Exiles, but I finally chickened out, reasoning that a fictional version would give me more narrative freedom in later installments. And for a long time, I considered making Garry Kasparov a major figure in the second novel. In the end, I didn’t, although there isn’t much doubt about which legendary chess grandmaster Victor Chigorin is supposed to represent. I changed the name partly to give me more flexibility in constructing the story, and also because I felt uncomfortable subjecting Kasparov to what ultimately happens to Chigorin, which isn’t pretty.

Besides, it’s usually more interesting when characters diverge from their original inspirations. I’ve mentioned before that Maddy and Ethan were loosely based on the real art world couple of Teresa Duncan and Jeremy Blake, although I doubt that many people would have made the connection. In Chapter 49, however, when we finally learn what happened to Anzor Archvadze—who has been missing in action for much of the novel’s second half—I imagine that more than a few readers were immediately reminded of Alexander Litvinenko. The two cases are very different, of course: Litvinenko died of radiation poisoning, while Archvadze is dying of toxic epidermal necrolysis, which bears a greater resemblance to another mysterious death in Russia. Still, I hope that readers do think of Litvinenko, not so much in order to capitalize on the parallels to a real event than out of a desire to remind them of how much like a novel the truth can be. Litvinenko’s death was often compared to something out of a spy thriller, but it was horribly real. And as farfetched as Archvadze’s fate might seem, reality is far stranger…

“When Maddy went to see Reynard…”

(Note: This post is the forty-ninth installment in my author’s commentary for The Icon Thief, covering Chapter 48. You can read the earlier installments here.)

For every published page of a novel, there’s usually a page or more of deletions, dead ends, and wrong turns. This makes sense: even when you work from a fairly detailed plan, as I generally do, you’re always figuring out important elements of the story as you go along, and it often isn’t until after you’re finished that you even know what the book is about. What’s more surprising is how quickly the old material vanishes down the memory hole. I’ve since gotten better about not overwriting my rough drafts, but when I wrote The Icon Thief, I was still cutting close to half of the total word count from first version to last. This meant losing scenes, digressions, subplots, and a lot of lovingly researched background. Needless to say, these were all good cuts, to the point where I don’t even remember what’s missing when I read the novel again. And whenever I think about a scene that was cut or radically revised—as was the case with Ethan’s death—it’s a very strange feeling, as if these events had taken place in some alternate but monetarily plausible universe that now seems hard to imagine.

Chapter 50 of The Icon Thief is one of the few sections where I vividly remember cutting something, mostly because it was a detail that I enjoyed but removed after overwhelming opposition from my early readers. It’s a quiet but important scene of the sort that takes place at least once in every conspiracy novel: the moment when the protagonist takes her concerns to a sympathetic outside party, only to be told that it’s all in her mind. (As usual, Foucault’s Pendulum has the best version of this scene I’ve ever read.) In this case, Maddy, shaken by her last argument with Ethan, goes to Reynard, her boss, to tell him about the plot they think they’ve discovered. Reynard listens, concerned, and then reasonably takes down her argument point by point. There’s no evidence that Lenin ever crossed paths with the Dadaists; Aleister Crowley was nothing but a fabulist; John Quinn, who knew both Crowley and Duchamp and was rumored to be the spymaster for the former, may have been precisely what he appeared to be, a patron of the arts with interesting friends.

And so on. The reader, of course, knows better than to take this sort of rational explanation at face value. But here’s the funny thing: Reynard is absolutely correct. All the points he makes are valid ones—they’re actually very close to my own feelings about the conspiracy theory my novel invents—and when he says that Maddy and Ethan may simply have been imposing a false pattern onto history, he’s right, although the reasons behind their paranoia are a little more complicated. At the time, though, this is far from clear, at least if I’ve done my work correctly. And any reader with a sense of the genre knows that it’s often the man who calmly suggests that you lie down and think things over who later shoves you into an unmarked van. Reynard isn’t quite on that level, but he’s certainly not telling the whole truth, and the challenge in writing this scene was to allow him to serve as a voice of reason without the reader suspecting that he’s in on the plot. After a bit of fiddling, I came up with this moment, which occurs as Maddy is about to leave his office:

“It’s all right.” Reynard walked her to the office door, laying a hand on her shoulder. “If you see Ethan, tell him that after all is said and done, I still want him here. And I still want you.”

Maddy looked into Reynard’s eyes. He did not remove his hand. For an inexplicable instant, she felt something pass between them, and knew with sudden certainty that he was going to take her in his arms.

In the end, the moment passed, leaving her unsure of what had happened, if anything had happened at all…

Now, I liked this a lot. I’d conceived it as a form of sleight of hand, a way to deflect readers from the possibility of Reynard’s guilt by distracting them with this awkward moment. The trouble is that it worked too well. Nearly every reader who read the novel in manuscript objected to this passage, saying that it made Maddy seem like the kind of woman who gets romantically involved with every man in her professional life. If only one reader had raised the issue, I might have ignored the note, but in this case, the consensus was strong enough that I finally cut these paragraphs just before we went out to publishers. And I still regret it a little, although I have no doubt that it was the right call. As it stands, I have a feeling that many readers will suspect that Reynard is up to no good, but I’m okay with it. We may have a hunch that Reynard is involved, but we don’t know why. And there are still a few big surprises in store…

“Ethan went into the gallery…”

(Note: This post is the forty-eighth installment in my author’s commentary for The Icon Thief, covering Chapter 47. You can read the earlier installments here. Massive spoilers follow—you’ve been warned.)

In many ways, this is the central chapter of The Icon Thief. It’s the scene that gets mentioned to me the most often when I’m asked about the book, and it clearly had the greatest impact on readers. It’s also one of the few sections that I go back and read when I’m trying to convince myself that I actually wrote a decent first novel. (Most days, I feel pretty good about the whole thing, but like all writers, I cycle through varying degrees of enthusiasm for my own work.) When I first started writing this author’s commentary, this was the chapter I looked forward to discussing the most. It certainly seems to have shocked a lot of people. And the strangest thing about this chapter, which now seems so crucial to the development both of The Icon Thief and of the novels that followed, is that it wasn’t part of the plot as originally conceived. If the surprise here works, it’s partially due to the fact that I didn’t know it was coming until very late in the game: as with the revelation of Karina’s true killer, I wrote most of the novel with one plan in mind, only to switch it at the last minute, which bakes an organic form of misdirection into the story itself.

As I’ve mentioned before, the Maddy and Ethan storyline was largely inspired by the real case of Teresa Duncan and Jeremy Blake, the New York art world couple whose lives ended in paranoia and a baffling double suicide. In most respects, Ethan isn’t much like Jeremy Blake, but I’d always been haunted by the accounts of Blake’s final walk into the sea, and in the first draft, Ethan dies in much the same way. He and Maddy have both grown increasingly paranoid, largely as a result of their unwitting exposure to a neurological agent at the party several days before, and in the end, they turn on each other as well. After Ethan accuses Lermontov, Maddy’s mentor and former employer, of being part of the plot, she leaves his apartment in a rage. The next day, Ethan takes a train to Far Rockaway, leaves his wallet and keys on the beach, removes most of his clothes, and walks into the water. But we don’t see it happen. Maddy receives a call from the police telling her that her friend is dead—her number was the last one dialed on Ethan’s phone. And that’s how his story ends, even as hers is still several steps away from its ultimate resolution.

This version of the story persisted throughout more than a year of rewrites. It’s possible that I clung to it for longer than I should have, if only because I liked the idea of Ethan’s senseless death and its connection to the novel’s original inspiration. At some point, however, my agent made the case that it wasn’t a very satisfying way of writing out such an important character. My first solution was to dramatize his suicide, rather than leaving it offstage, and the result was a fairly strong chapter. (At least, I think it was fairly strong—I haven’t read it in years.) My agent still pushed back, though, saying that the fact of his suicide itself had inherent narrative problems. At this stage, remember, we’d been revising this novel for a long time without going out to publishers, and the last thing I wanted was to change the plot in a drastic way. After mulling it over, however, I began to see a possible way out, and I wrote my agent the following:

After his final argument with Maddy, Ethan, brooding over the situation, decides that he can only convince her of his theory by proving that Lermontov is involved. He walks around the city for hours, trying to build up his resolve, then leaves a note at Maddy’s house and goes to Lermontov’s gallery. Ethan doesn’t really expect to find Lermontov there, but he does. He introduces himself, lays out what he’s found, and demands that Lermontov tell him the truth about the Rosicrucians.

And Lermontov kills him.

Needless to say, that’s what eventually happens, and I think the result is the best scene in the book. In my note to my agent, I pointed out that this change solves a number of problems at once: it offers us a more compelling death scene for Ethan, gives Maddy a more urgent reason to believe that her life is in in danger, tightens the screws on Ilya—who will potentially be framed for the murder—and transforms Lermontov into a more imposing villain. (Interestingly, it’s only after reading over the note again today that I remember that I briefly considered having Reynard, Maddy’s boss, kill Ethan instead, which would have been even more out of the blue, but probably unworkable.) And it shifted the terms of the rest of the novel in ways I only gradually began to realize. At first, the chapter stood more or less on its own, with the remainder of the story proceeding along the same track as before. Eventually, though, I realized that I had to fully confront the implications of this scene. In the original version, the novel ends with the arrest of Lermontov and Vasylenko in London, with Ilya working with Powell to take them down. Reading it over again, however, I saw that this ending no longer worked. Lermontov had to be forced to pay a greater price. And Maddy was the only one who could do it…

“Sharkovsky slid through the narrow opening…”

(Note: This post is the forty-seventh installment in my author’s commentary for The Icon Thief, covering Chapter 46. You can read the earlier installments here.)

I’ve written before about how research for a novel is less about factual accuracy than about finding material for dreams. In particular, it’s a valuable source of specificity. When you first start writing a story, your ideas tend to be vivid in certain areas and amorphous in others, and research is one way of acquiring a useful stash of facts, images, and concrete details—the building blocks out of which all good fiction is assembled. Reading books is no substitute for firsthand observation, of course, but at its best, it can supplement and enrich what you can take in through your own experience. Painters know that you often can’t see what’s right in front of your eyes unless you know what you’re looking for, which is why the formal study of anatomy and perspective is so essential. And although it may sound backward in principle, in practice, it’s often not until you’ve done a bit of work in a library that you’re prepared to take in the specifics of the world around you.

When I started writing The Icon Thief, for instance, I knew that much of the novel would unfold in Brighton Beach, since the world of Russian immigrants and mafiosos was central to the story I had in mind. This inevitably meant that I’d need to incorporate the details of the neighborhood itself into the plot. Just as Hitchcock knew that a movie set in Holland would need to include tulips and windmills, I knew that I had to incorporate the amusement parks and furniture stores of Coney Island: anything less would be a waste of material. I never had the chance to write the amusement park chase of my dreams, but otherwise, I think I made good use of the locations that the setting afforded. This was partially the result of many days spent exploring the beach and the surrounding streets and buildings, including one memorable trip to a steam room in Sheepshead Bay, but I also owed a great deal to some serendipitous secondary sources.

I knew, for example, that I would have to go under the boardwalk. Early on, I came up with the image of a woman’s headless body preserved in the sand beneath the boards, and although I didn’t know who she was or how she tied in with the rest of the story, it was an image I wanted to keep, which required a lot of surreptitious legwork. I spent the better part of several days walking on the boardwalk, studying the area underneath and trying to figure out exactly how you’d deposit a body there. I probably could have figured most of the details out on my own, but I also lucked out by finding a piece by Michael Wilson of the New York Times that described the boardwalk’s recent history, and how the space under the boards—which used to be open, walkable, and a popular spot for the homeless—had been reclaimed by the sand after the Army Corps of Engineers extended the beach. Sand, I learned, had blown in and been caught by the newly installed fences at the rear, until finally it was all the way up to the boards themselves.

This piece of information was vital, since it gave me a timeline for the dumping of the body, which could only have been brought to where it was found at a time when that section of the boardwalk was clear. Even more crucially, it taught me how to see. Going back to the boardwalk, I noticed for the first time how certain kinds of fencing allowed the sand to blow through, leaving areas that were clear enough for a man to walk upright, while sections only a few yards away were impassable. This information might have been obvious to anyone with a good pair of eyes, but in my case, it was that initial bit of research that allowed me to see the sand for what it really was. This paid off again in Chapter 46, when Ilya and Sharkovsky make their escape from the raid at the club by utilizing another gap, an area under a drinking fountain that had been deliberately kept clear to allow maintenance crews to reach the plumbing. It was there all along, but it was only because my research had taught me to think about the sand that I saw it. And otherwise, I don’t know how Ilya would ever have gotten away…

“Or should I call you the Scythian?”

(Note: This post is the forty-sixth installment in my author’s commentary for The Icon Thief, covering Chapter 45. You can read the earlier installments here.)

As I’ve mentioned before, no work of art has had a greater influence on my own fiction, at least on a practical level, than the movie L.A. Confidential. The novel is extraordinary as well, of course, and there are moments and scenes, like the last stand of Buzz Meeks, that I’ve revisited countless times. Yet it’s the movie that sticks in my head, both for its surface pleasures of action and atmosphere and for its deeper structure. Something about its story of three rival cops whose lives intersect at crucial moments appealed to me at once: it’s the best illustration I know of how a multiple plot can become greater than the sum of its parts, until it seems to encompass an entire world. It opens up possibilities of contrast, juxtaposition, and shifting perspectives, and when the pieces come together at last, it’s with an almost musical satisfaction. As a result, this kind of tripartite plot has been central to each of the novels I’ve written, although I’ve since come to see the film’s example as rather misleading: most stories lend themselves best to a single point of view, and there’s a reason why a movie like this only comes around once a decade or so.

But when I look back, I find that I’ve also misremembered or deliberately distorted the film’s structure in my own imagination. I’ve always thought of it as a movie that starts with its three main characters far apart, only to bring them inexorably together, but this isn’t exactly true. In fact, two of its three major characters share just one scene. On the night of Bloody Christmas, Jack Vincennes sticks his head into Bud White’s office and says: “You better put a leash on your partner before he kills somebody.” Then he leaves without waiting for a response. As far as I can recall, that’s the only time Bud and Jack share the same frame, and Bud doesn’t even reply. Like the silences in Shakespeare, it’s a striking omission, and one that raises a lot of questions. This is a dense, crowded movie that finds time for countless fruitful pairings among its five or six most important players—Bud and Lynn, Ed and Dudley, Bud and Dudley, Ed and Jack, and finally Ed and Bud—and the fact that Bud and Jack aren’t among them is revealing in itself. And it’s quite possible that Brian Helgeland and Curtis Hanson, for all their ingenuity, just couldn’t figure out what these two men would have to say to each other.

There’s a similar hiatus in The Icon Thief, which is a novel that owes a great deal to L.A. Confidential in its construction, even if the movie’s influence is otherwise hard to see. My investigator, Alan Powell, spends most of the novel unraveling a complicated criminal conspiracy with the thief Ilya Severin at its center, but if you don’t count their brief chase at the New York County Courthouse, Powell and Ilya only appear together once. It’s in Chapter 45, in the basement of the Club Marat, as Ilya emerges from the restaurant office with Sharkovsky as a hostage. Powell is there already, of course, along with a squadron of law enforcement officers, and in the standoff that follows, the two men exchange a line or two. But it’s Powell’s supervisor who ends up doing most of the talking, and in any case, the scene quickly moves to the next stage, as Ilya works out the logistics of his escape. And that, incredibly, is it. By the time the next chapter begins, Ilya and Powell have been separated once more, and they don’t cross paths again. These are two of the book’s three most important characters, and their only real encounter lasts for less than a page.

This wasn’t originally how it was supposed to happen. In fact, in my first draft, Powell and Ilya reunite on the final page. The story of how the epilogue was revised at the last minute, with enormous consequences both for this book and for the ensuing series, is one I’ll tell at the proper time. As it stands, though, the fact that Ilya and Powell don’t otherwise interact deserves an explanation. The first reason is that Ilya is most interesting when he remains something of an enigma, and whenever he’s clearly seen by another character, it diminishes that mystery—a problem I’d be forced to confront more seriously in City of Exiles. The second reason is a technical one: this is a book about a chase, and by definition, the pursuer and the pursued don’t often end up in the same room. But the third reason is the most important. What I didn’t understand about multiple plots when I began this book, and started to figure out only after I’d written several drafts, is that they’re most convincing when a piece is removed. A plot like this works to the extent that it evokes something larger, a world in which the stories intersect beyond the margins of the page, and if each piece connects too neatly with every other, that illusion is broken. In the end, Powell and Ilya go their separate ways. But they’ll meet again in another book…

“He entered the club, leaving the door open…”

(Note: This post is the forty-fifth installment in my author’s commentary for The Icon Thief, covering Chapter 44. You can read the earlier installments here.)

“Although there is no substitute for merit in writing,” E.B. White notes in The Elements of Style, “clarity comes closest to being one.” This is true of all aspects of fiction, from character to dialogue, and it’s especially true of action scenes. An action sequence in a novel is a kind of musical number in which all the basic notes of storytelling are hit with extraordinary focus and concentration, and there’s little room for error. It’s a war between the two central aspects of reading: the impulse to linger and the need to keep going. Most good fiction is written on the premise that the reader will occasionally pause to reflect on a complex sentence, read a paragraph over again, or even turn back a few pages for clarification or context. An action scene, by contrast, is one that begs to unfold in real time, with the act of reading coinciding as much as possible with the rhythm of the action itself. Writing a scene that satisfies these requirements while remaining stylistically consistent with the rest of the novel is a real challenge, and the key, as White points out, is clarity—although I doubt he was thinking of chase scenes or gunfights when he wrote those words.

A novel isn’t a movie, of course, and it can be dangerous to look to film for examples of how to stage action: movies have tools at their disposal, like montage and intercutting, that don’t come easily to fiction. But film still provides some useful illustrations. I’ve noted before that my favorite action scenes of recent years—the Guggenheim shootout in The International, the Burj Khalifa climb in Mission: Impossible—Ghost Protocol, and the opening chase in Drive—were all worked out on the page, rather than in the editing room. They unfold in real time and proceed from one logical beat to the next, and there’s rarely any doubt about what the characters are thinking, although the full meaning of the scene in Drive isn’t revealed until the very end of the sequence. They’re cleanly photographed and cut together in a way that isn’t afraid to hold on a shot, and each depends intimately on spatial relationships, both in the physical geography of the set and within the frame itself. All these decisions have their common origin in a concern for clarity. The audience is grounded at every moment, and we’re never confused, as we often are by Michael Bay, about what is happening on the screen.

In short, a good action scene, either in fiction or in film, serves as an intense microcosm of the virtues of good storytelling itself. Like the story as a whole, it’s best when it’s built on a sequence of clear problems, which the protagonist succeeds or fails at solving in some logical order. Clarity is always important, but never more so than when the action needs to move at a fast pace, and the use of clean, vivid prose is the literary equivalent of the composed shot and spatially coherent editing. Each sequence has a clear beginning, middle, and end: it’s no accident that the three scenes I’ve mentioned above could be taken out of their surrounding movies and enjoyed in their own right, as many of us tend to do when we’re watching them at home. And just as finding the right rhythm for an action scene in a movie can come down to the addition or removal of a few frames, the action in a novel needs to be written and revised with particular care, long after the author’s own excitement has begun to fade, until the logic is crystal clear in the moment to a reader encountering it for the first time.

When I wrote The Icon Thief, I was still figuring most of this out, and although it took me a long time to formulate these rules, I managed to follow them intuitively, at least most of the time. Chapter 44, for instance, is as close to a conventional action scene as you’ll find in the book, and it’s only in retrospect that I can see how the objectives follow logically one after the other, as determined by the geography of the setting at the Club Marat. Ilya emerges from the darkness under the boardwalk behind the club, takes out the busboy, slips through the back door, makes his way up the corridor to the downstairs office, incapacitates the restaurant manager, arms himself, and places a call to Sharkovsky for a meeting, knowing that he’ll come to the office first. The only bloodshed comes a few pages later, when Ilya shoots Misha, but as far as I’m concerned, that’s only a coda to the real source of interest, as we follow Ilya’s thoughts and decisions from one point of cover to the next. There’s no question that I was inspired by the movies I loved, especially the final confrontation in Michael Mann’s Thief, and the action here unfolds in the novelistic equivalent of one continuous shot. And it isn’t over yet. As Ilya and Sharkovsky are about to discover, they aren’t alone…

“When Powell and Wolfe arrived at the club…”

(Note: This post is the forty-fourth installment in my author’s commentary for The Icon Thief, covering Chapter 43. You can read the earlier installments here.)

In some ways, the novel is an unwieldy, slightly unnatural form of storytelling. A poem, short story, or play arises directly from the oral tradition: it can be told aloud in a few minutes or an hour, and listeners can easily remember most of the important plot points. Even epic poetry, which goes on for much longer, usually boils down to episodes that can be condensed or expanded according to the needs of the audience, strung together like beads on a string. (We can still see this structure of our surviving text of the Iliad, which preserves the full version of certain episodes while reducing others to only a few lines.) The average novel, by contrast, presents a story that is too complex to be held in the mind all at once, even by the author. As I’ve mentioned before, it’s a structure that evolved from the physical characteristics of printed books themselves, which allow readers to turn pages both ways, so that elements introduced in the first chapter can return to play an important role near the end—a form of setup and payoff that doesn’t exist in the oral tradition. And although the novel seems natural now, it’s really a recent development in the history of how we tell stories.

That’s why it’s important to acknowledge its limitations as well as its strengths. On the one hand, a novel rarely achieves the kind of crystalline perfection that we see in poetry or short fiction, and when it does, it may seem artificial or unreal, as John Gardner observes of Madame Bovary. A novel, as Henry James said of Tolstoy, often ends up being a loose, baggy monster, and in order for it to feel like an accurate representation of life—as well as a pleasurable experience for the reader—it can’t pitch every page at the same level of intensity. Instead, it’s a series of convergences and divergences, of rising and falling action, and it requires time and patience for its full impact to be felt. On the other hand, its size and relative complexity allow it to achieve effects that aren’t possible in shorter forms. It can methodically establish themes, motifs, and story elements that will pay dividends at a later time, and when it works, the effect can be almost symphonic, as threads that have been independently established come together at last.

This may seem like a roundabout way of getting to The Icon Thief, which even I’m willing to admit is a very modest example of the novel form. But like most first novels, it stands both as a story in itself and as a kind of laboratory in which a writer is figuring out his craft for the first time. When I wrote the first draft, I was in my late twenties, and although I’d written one unpublished novel already, I still had a lot to learn. As a result, the book sometimes feels like a sandbox in which I was testing out various approaches to telling this kind of extended story. Although the result is clearly a product of its genre, it also allowed me to think about narrative in a way that paid off when it came to my second and third books, as well as the ones I hope to write in the future. Suspense, in particular, seemed like a way to explore these tools in their purest state, as action foreshadowed, promised, and delivered. And one thing that fascinated me from the very beginning was how a novel can use its own intricacy of construction, which allows for more building blocks than other forms, so that the events of the plot are inextricable from the structure of the book itself.

As Chapter 43 begins, for instance, we’re entering a point in the novel where the structure of the story serves almost a character in itself. Three distinct groups of characters—Powell and his partners in law enforcement, Sharkovsky and his men, and Ilya himself—are converging on a common location, the club in Brighton Beach, that has already been established in detail, both within the narrative itself and in what amounted to a direct briefing to the reader. The next few chapters will narrate the ensuing developments from multiple perspectives, often moving back and forth slightly in time. This was both a technical solution to the problem of treating simultaneous action and a way of binding the scenes more closely together, and none of it would mean as much if the foundations hadn’t been laid much earlier. By now, if I’ve done my work properly, the reader knows something about Powell, Wolfe, Ilya, and all the others, and has some idea of how each character will react to the violent events that the structure itself implies. My one regret, which is also inherent to the novel form, is that the reader can tell that we aren’t quite at the real climax yet: we have well over one hundred pages to go. And there’s a lot still left to come…

“A solitary figure stood near the basketball courts…”

(Note: This post is the forty-third installment in my author’s commentary for The Icon Thief, covering Chapter 42. You can read the earlier installments here.)

In theory, a writer can get to know the characters in his fiction better than anyone in his own life, himself included. I’m often struck by how little information I really have about people—even close friends—I’ve known for years: I’ve generally only spent time with them in one particular context, and I have a hunch that I’d be startled to see what they’re like at the office, say, or with their own families. Go a little further back, to childhood or adolescence, and the picture is even less clear. In fiction, by contrast, we can learn everything about a person, at least in principle. A quick search online turns up countless forms and questionnaires designed to help authors brainstorm every detail of their characters’ appearance, habits, past, and inner life, from their eye color to their favorite hobbies to how they really feel about their parents. It’s more scrutiny than many of us even devote to ourselves, at least at any one time, and the question of how much of this material an author requires to invent fictional but convincing men and women is an issue that every writer needs to confront.

In my own case, I’ve been inclined leave many of these questions unanswered. Part of this is a practical consideration: as I’ve mentioned many times before, I’m not a fan of backstory, and although I spend a fair amount of time thinking about each character’s personal history before I start writing, I know that very little of this material will end up in the finished novel. Human beings, both in fiction and in real life, tend to be more vividly defined by their needs in the moment, and while what happened to them in the past can inform their present wants and decisions, we very rarely embody the full sum of our life stories: one quality tends to predominate over the next from one minute to another, and certain threads of our personality grow in importance while others wither away. As a result, I’ve never tried to create systematic backstories for my characters. Instead, I work out which elements bear most urgently on the present moment and leave the rest in shadow, which is why, for instance, I go into detail about Maddy’s failed attempt to start an art gallery, but still don’t know the names of her parents.

But there’s also a more intuitive aspect to this approach. Protagonists, as William Goldman observes in Which Lie Did I Tell?, must have mystery, and they’re often more interesting to the extent that the author withholds crucial information. Part of this has to do with the way we identify with characters in fiction: David Mamet has pointed out that it’s better to say “A hero on a white horse” than “A tall hero on a white horse,” because the more we add unnecessary detail, the harder it is for readers to imagine themselves in the protagonist’s place. (This is one reason why I’m especially resistant to detailed descriptions of a character’s appearance, to an extent that sometimes frustrates my circle of initial readers.) It’s possible, of course, for a writer to leave important elements to implication in the finished work, while privately knowing a great deal about the characters. But I’ve found that I’m happiest when certain aspects remain a mystery to me as well, even as I obsessively think about each character’s objectives from scene to scene.

In Chapter 42 of The Icon Thief, for instance, Ilya is standing near a playground in Brighton Beach, preparing to go after the men who betrayed him, when he sees a little boy and his mother:

“Daniel!” the boy’s mother shouted. At the sound of the name, Ilya’s head gave an involuntary jerk…Watching the boy rejoin his mother, Ilya thought of the many transformations that he had undergone since he had last referred to himself by that name.

This is the only indication in the entire novel, and in the two books that follow, that Ilya was once called something else, or that the name he goes by now is one that he chose for himself. There’s obviously a story here, and it’s possible that when I wrote this passage, I was laying in a hint for something I intended to develop later on. As it turns out, I never did. And if you were to ask me who Ilya was and what he was doing before he was thrown into Vladimir Prison, I’m not sure I could tell you, although there are clues scattered throughout the series. Keeping these details hidden, even to myself, turned out to be a kind of insurance policy: Ilya, I realized, grew more interesting the less he was shown, and by keeping his past deliberately undeveloped, there was no risk that I’d err in revealing too much. What really matters, far more than where he came from, is what he wants now. And at the moment, he wants revenge…

“Gentlemen, we have a warrant…”

(Note: This post is the forty-second installment in my author’s commentary for The Icon Thief, covering Chapter 41. You can read the earlier installments here.)

One of the few things I know about writing is that less is usually more, and that a story is generally effective in proportion to how much the author can leave out. As overstuffed as my novels tend to be, I’m always trying to pare back elements like backstory and personal description, to the point where my advance readers often beg me to put them back in. A lot of things, I’ve found, are best left to implication, although it takes a lot of revision and feedback to find the right level of clarity. Still, there are always places where it’s necessary to spell things out. When a story contains a lot of complicated action, for instance, it’s often useful to brief readers on what they’re about to see, which can be allowed to unfold more impressionistically when the crucial moment comes: it’s fine if your characters are confused or uncertain, but that’s rarely an emotion you want in the reader, unless you’re trying to achieve it on purpose. (The best example I know of this kind of advance grounding is the computer simulation of the sinking ship in Titanic, a movie whose shrewdness of construction has been frequently underestimated.)

And giving your reader a game plan for how the action is supposed to unfold can be particularly useful in suspense. A good thriller is all about anticipation, and there’s a peculiar satisfaction in being given just enough information on what’s about to take place to look forward to the action to come—and especially to see how it deviates from what the characters are expecting. In describing the scene in The Godfather where Michael shoots Sollozzo and McCloskey, David Thomson talks about “the sinister charm of action foreseen, spelled out, and finally delivered,” and when properly done, it’s one of the most useful tools in a writer’s arsenal. Hence the moment in any decent heist movie in which the logistics of entering the mansion, disabling the security system, and cracking the safe are lovingly described in advance, which stands as one of the few instances when exposition builds the suspense, rather than destroying it.

Suspense, as Thomson points out in his discussion of Inception, has a lot in common with comedy, which is also built on anticipation and surprise, and at its best, this approach embodies a classic piece of comedy advice: “Tell them what you’re going to do. Then do it. Then tell them what you did.” In a thriller, though, this last step might be better described as “Then tell them what really happened.” Because spelling out the coming action carries an additional charge of irony and tension. A sophisticated reader—which is to say anyone who has seen a movie or two—is well aware that nothing ever goes entirely as planned: a properly constructed caper film, for instance, will withhold the most essential information until the big score itself begins to unfold, as in Ocean’s 11, which means that any initial description of the plan is really just a list of things that can go wrong. And as always, it’s best to acknowledge this, and play off the reader’s knowledge of the genre, rather than trying to fight against it.

Chapter 41 of The Icon Thief, for example, is largely taken up by one of my favorite categories of this kind of exposition: the police briefing in advance of a raid. Louis Barlow, the FBI assistant special agent in charge, spends several pages describing what will take place when they finally raid Sharkovsky’s club in Brighton Beach, and because of the considerations I’ve mentioned above, I give more space to this speech than I might have done elsewhere in the novel. In fact, this is one of the few chapters that was significantly expanded in the rewrite, as it became clear to me that I needed to lay out the impending action as clearly as possible—and if I’ve done my work properly, the reader will appreciate it on several levels at once. It creates anticipation for the scene to come; it provides a kind of map for following the action itself, which will ultimately unfold across multiple points of view; and, best of all, it allows the reader to wonder what, exactly, is going to go wrong. Because as you can probably guess, this raid isn’t going to go exactly as planned…