“But some things can’t be undone…”

Note: This post is the sixty-second—and final—installment in my author’s commentary for Eternal Empire, covering the epilogue. You can read the previous installments here.

How do you end a series that has lasted for three books and more than a thousand pages? To some extent, no conclusion can be completely satisfying, so it makes sense to focus on what you actually stand a chance of achieving. There’s a reason, for instance, that so few series finales live up to our hopes: a healthy television show has to cultivate and maintain more narrative threads than can be resolved in a single episode, so any finale has to leave certain elements unaddressed. In practice, this means that entire characters and subplots are ignored in favor of others, which is exactly how it should be. During the last season of Mad Men, Matthew Weiner and his writing team prepared a list of story points that they wanted to revisit, and reading it over again now is a fascinating exercise. The show used some of the ideas, but it omitted many more, and we never did get a chance to see what happened to Sal, Dr. Faye, or Peggy’s baby. This kind of creative pruning is undoubtedly good for the whole, and it serves as a reminder of Weiner’s exceptional skill as a showrunner. Mad Men was one of the most intricate dramas ever written, with literally dozens of characters who might have earned a resonant guest appearance in the closing stretch of episodes. But Weiner rightly forced himself to focus on the essentials, while also allowing for a few intriguing digressions, and the result was one of the strongest finales I’ve ever seen—a rare example of a show sticking the landing to maintain an impossibly high standard from the first episode to the last.

It’s tempting to think of a series finale as a piece of valuable real estate in which every second counts, or as a zero-sum game in which every moment devoted to one character means that another won’t have a chance to appear. (Watching the Mad Men finale, I found myself waiting for my favorite supporting players to pop up, and as soon as they had their scene, I couldn’t help thinking: That’s the last thing I’ll ever see them do.) But it can be dangerous to take such a singleminded approach to any unit of narrative, particularly for shows that have thrived on the unpredictable. My favorite example is the series finale of Twin Peaks, which wasn’t even meant to end the show, but provided as perfect a conclusion as any viewer could want—an opinion that I’ll continue to hold even after the new season premieres on Showtime. Instead of taking time to check in with everyone in their huge cast, David Lynch and Mark Frost indulge in long, seemingly pointless set pieces: the scene in the bank with Audrey, with the decrepit manager shuffling interminable across the floor to get her a drink of water, and especially the sequence in the Black Lodge, which is still the weirdest, emptiest twenty minutes ever to air on network television. You can imagine a viewer almost shouting at the screen for Lynch and Frost to get back to Sheriff Truman or Shelly or Donna, but that wouldn’t have been true to the show’s vision. Similarly, the Mad Men finale devotes a long scene to a character we’ve never seen before or since, the man at the encounter group who ends up inspiring Don’s return to humanity. It might seem like a strange choice, but it was the right call: Don’s relationships with every other character were so burdened with history that it took a new face to carry him over the finish line.

I found myself dealing with many of the same issues when it came to the epilogue of Eternal Empire, which was like the final season of a television series that had gone on for longer than I’d ever expected. Maddy and Wolfe had already received a sendoff in the previous chapter, so I only had to deal with Ilya. Pragmatically, the scene could have been about anything, or nothing at all. Ilya was always a peculiar character: he was defined mostly by action, and I deliberately refrained from detailing large portions of his backstory, on the assumption that he would be more interesting the less we knew about his past. It would have been easy to give him a conclusion that filled in more of his background, or that restored something of what he had lost—his family, a home, his sense of himself as a fundamentally good man. But that didn’t seem right. Another theme that you often see in series finales, particularly for a certain type of sitcom, is the showrunner’s desire to make every character’s dreams come true: the last season of Parks and Recreation, in particular, was a sustained exercise in wish fulfillment. I can understand the need to reward the characters that we love, but in Ilya’s case, what I loved about him was inseparable from the fact of his rootlessness. The novel repeatedly draws a parallel between his situation and that of the Khazars, the tribe of nomads that converted to Judaism before being erased from history, and I once compared him to the tzaddikim, or the unknown men and women for whose sake God refrains from destroying the world.  Above all else, he was the Scythian, a wanderer of the steppes. I chose these emblems intuitively, but they clearly all have something in common. And it implied that Ilya would have to depart the series as he began it: as a man without a country.

What we get, in the end, is this quiet scene, in which Ilya goes to visit the daughter of the woman who had helped him in Yalta. The woman was a bride of the brotherhood, a former convict who gave up her family to work with the thieves, and her daughter ended up as the servant of a gangster in Moldova, five hundred miles away. Ilya gives her some money and her mother’s address, which he hopes will allow them to build a new life together, and then leaves. (The song that is playing on the girl’s cassette deck, incidentally, is Joni Mitchell’s “Cactus Tree.” This might be the nerdiest, most obscure inside joke of the entire series: it’s the song that appears in a deleted epigraph in the page proofs of Gravity’s Rainbow, before Thomas Pynchon removed it prior to publication. I’d wanted to use it, in some form, since The Icon Thief, and the fact that it includes the word “eternity” was a lucky coincidence.) It all makes for a subdued conclusion to the trilogy, and I came up with it fairly late in the process: as far as I can remember, the idea that there was a connection between the women in Yalta and Moldova didn’t occur to me until I’d already outlined the scenes, and this conclusion would have been an equally late addition. And it works, more or less, even if it feels a little too much like the penultimate scene of The Bourne Supremacy. It seemed right to end the series—which was pointedly made up of big, exaggerated gestures—on a gentle note, which implies that reuniting a parent and her child might be an act of greater significance than saving the world. I don’t know where Ilya goes after this, even though I spent the better part of four years trying to see through his eyes. But I suspect that he just wants to be left in peace…

“The need for change is there…”

Note: This post is the sixty-first installment in my author’s commentary for Eternal Empire, covering Chapter 60. You can read the previous installments here.

Vladimir Putin is still here. I type these words not because we need to be reminded of that fact—I can’t think of another foreign political leader whose shadow has loomed so ominously over a peacetime presidential race—but to consider what it means. When I began writing The Icon Thief, more than eight years ago, Putin was ostensibly on his way out: he was ineligible to run for a third term, so the reigns of power were passed to Dmitry Medvedev, his chosen successor. Instead, Medvedev appointed him prime minister, and a few years later, Putin was back in the presidency, as if he’d never been gone. It isn’t hard to imagine him pulling the same trick forever, or for as long as his health holds out, which might be for quite some time. He’s only in his early sixties now, which is practically his young adulthood compared to some of the decrepit Russian leaders of the past, and he’s in what he takes pains to assure us is peak physical condition. It’s a situation that ought to keep most of us up at night, but it’s also a boon to suspense novelists. As I once pointed out, Putin’s name is the most evocative word in the lexicon of the modern thriller: it calls up an entire world of intrigue and implication, allowing a novel to do in a few sentences what might otherwise require five pages. As a rhetorical device, it isn’t just confined to fiction, either. Putin wouldn’t be evoked so often in this election if he didn’t have such a powerful hold over our imaginations, and recent events have only confirmed, as I’ve said from the beginning, that nothing that a writer can invent about Russia can possibly compare to the reality.

Incorporating a contemporary or historical political figure into a thriller is nothing new, of course. The gold standard was set, as it was in so many other things, by Frederick Forsyth, who built The Day of the Jackal around an assassination attempt on Charles de Gaulle, and who gave prominent speaking parts to Margaret Thatcher in several of his later novels. It’s a trick that grows stale when a writer uses it too often, as Forsyth sometimes does, but its easy to understand its appeal. For a certain kind of thriller, the story is less about something that could happen than about what might be happening right now, or that has already happened without our knowledge. Such novels often set up a sliding scale of verisimilitude, starting with big, obvious figures like Putin, working their way down through historical figures or events that aren’t as familiar, and finally entering the realm of pure fiction. Even if you’re reasonably conversant with current events, you can have trouble telling where fact leaves off and invention begins, especially when the novel starts to show its age. (For instance, I have a feeling that most contemporary readers of The Day of the Jackal aren’t aware that the opening sequence, which depicts a failed attempt on de Gaulle’s life, is based on fact—an interesting case of a novel outliving the material that it once used to enhance its own credibility.) Ideally, the transition from someone like Putin to the fictional characters at the bottom of the pecking order should be totally seamless, at least in the moment. We know that Putin is real and that most of the other characters aren’t, but in some cases, we aren’t sure, and the overwhelming fact of Putin himself serves to organize and enhance the rest of the story.

Eternal Empire is literally framed by Putin, both in terms of how the novel was conceived and of how it was finally published. It opens with an epigraph from Rachel Polonsky’s Molotov’s Magic Lantern, which describes how Putin asked to have a fragment of the polar seabed brought back to him as a nod to the underground kingdom of Shambhala, and it ends with an excerpt from a New York Times article from December 10, 2011, which describes the abortive protests that flared up that year against the Putin regime. As I’ve mentioned elsewhere, the entire novel unfolded like a paper flower from those lines in Polonsky’s book, and it isn’t hard to see why they struck me. In juxtaposing the steely figure of Putin, the ultimate pragmatist, with the gauzy myth of Shambhala, it encapsulates the tension that defined the rest of the series, which in many ways is about the collision between practical spycraft and the weirder elements that have a way of impinging on the rational picture. (As Powell says to Wolfe of the Shambhala story: “That doesn’t sound like the Putin I know.”) The closing epigraph attracted me for many of the same reasons. Its image of protesters with white flowers and ribbons was derived from an actual event, but it could easily stand for something more. A white flower can mean just about anything, so it wasn’t hard for me to tweak the story so that the protests seemed to emerge from the Shambhala plot. And the entire narrative was timed to culminate at this moment, which would serve as the visible eruption of the forces that my characters had spent the entire book marshaling in secret.

Now that five years have passed, the image that concludes the trilogy, of Maddy watching the protesters on television, feels very different in tone. The protests themselves are little more than a footnote, and Putin’s hold on power has never been stronger. Since the plot hinges on a plan to change Russian politics from the inside, the historical outcome might seem to undermine the whole story. I’m not sure it does, though. Maddy notes that Tarkovsky has bought himself “a few years” to prepare, which might well mean that his plan is underway even now—although I doubt it. More pragmatically, the characters observe, both here and in the epilogue, that most attempts at reform are crushed, and that a revolution is more likely to die than to endure. (You can picture me typing those lines, more than three years ago, as a way of hedging my bets.) But if there’s a thread that runs through all these novels, it’s the importance of small, private victories in the face of the indifference or hostility of larger systems. I began the series with a conspiracy novel, which is a genre that implicitly raises the issue, even in its pulpiest incarnations, of the relationship between the individual and the impersonal forces to which he or she is subjected. All three books conclude on a similar note, which is that we can try to get glimpse behind the mask, if only for a moment, and then return to the more achievable task of establishing what little order we can in our own lives. It isn’t much of an answer, but it provides just enough consolation to see us through, both in a novel and in the real world. Putin survives, as I suspect I always knew he would. But so do Wolfe and Maddy. And that’s how their story ends…

“To spare another man’s life…”

Note: This post is the sixtieth installment in my author’s commentary for Eternal Empire, covering Chapter 59. You can read the previous installments here.

Why do our villains always have to die? Roger Ebert says somewhere—I haven’t been able to track down the exact reference—that he’d be happier if a movie ended with the hero sealing the bad guy’s fate with a few well-chosen lines of dialogue, followed by a closeup of the bastard’s face as he absorbs his predicament. And there’s no question that this would be much more satisfying than the anticlimactic death scenes that most stories tend to deliver. It’s safe to say that if a book or screenplay goes through the trouble of creating a nice, hateful antagonist, it’s usually for the sake of his ultimate comeuppance: we want to see him pay for what he’s done, and hopefully suffer in the process. In practice, the manner in which he ends up being dispatched rarely lives up to the punishment we’ve mentally assigned to him in advance. For one thing, it’s often too fast. We want him to perish at a moment of total recognition, and the nature of most fictional deaths means that the realization is over almost before it begins. (This may be the real reason why so many villains are killed by falling from a great height. It leaves the hero’s hands relatively clean, however illogically, and it also allows for at least a few seconds of mute astonishment and understanding to cross the bad guy’s face. The story goes that during the filming of Die Hard, director John McTiernan let Alan Rickman drop a second before he was expecting it. Rickman was understandably furious, but the look he gives the camera is worth it: there are few things more delicious than seeing him lose that mask of perfect, icy control.)

All things being equal, it’s best to allow the villain to live to deal with the consequences. But there are also situations in which a death can feel dramatically necessary. I’ve never forgotten what Robert Towne once said about a similar plot point at the end of Chinatown. Originally, Towne had wanted the movie to conclude on an ambiguous note, but he was overruled by Roman Polanski. Years later, Towne said:

In hindsight, I’ve come to feel that Roman was probably right about the ending, that I don’t think that what I had in mind could have been done; that an end with that ambiguity and ambivalence that I had in mind simply could not satisfactorily be done as the tag to a movie with that much complexity; the end had to have a level of stark simplicity that at the time I thought was excessively melodramatic. Roman rightly believed that the complexities had to conclude with a simple severing of the knot.

Chinatown, of course, ends with anything but the villain getting what he deserves, but the principle is largely the same. In some respects, it’s a matter of contrast. A story that consists of one act of violence after another might benefit from a more nuanced ending, while one that teases out its complexities would go out best with a stark, sudden conclusion. I’ve always preferred the brutally abbreviated last scene of The Departed to that of Infernal Affairs, for instance, because that twisty, convoluted story really needs to close with a full stop. As De Niro says at the end of Casino: “And that’s that.”

And a villain’s death can be necessary in order to close off the story completely: it’s like scorching the end of a nylon rope to prevent it from unraveling. Death is nothing if not definitive, and it can seem unfair to the viewer or reader to leave the narrative open at one end after they’ve come so far already. The decision as to whether or not to spare the villain is a tricky one, and it can be determined by forces from much earlier in the narrative. In his director’s commentary for Mission: Impossible—Rogue Nation, Christopher McQuarrie says that they spent countless drafts trying to figure out ways for Ethan to kill Solomon Lane, only to find that none of the results seemed satisfying. The reason, they discovered, was that Lane hadn’t done enough to make Ethan hate him in particular: it just wasn’t personal, so it didn’t need to end with anything so intimate as a fight to the death. A story’s internal mechanics can also push the ending in the other direction. The original draft of The Icon Thief, which persisted almost until the book went out to publishers, had all three of the primary antagonists surviving, and in fact, Maddy even asks Ilya to spare Sharkovsky’s life. In the rewrite, I realized that Lermontov had to die to balance out the death of another character earlier in the novel, which in itself was a very late addition, and that Maddy had to be the one to take that revenge. This kind of narrative bookkeeping, in which the writer cooks the numbers until they come out more or less right, is something that every author does, consciously or otherwise. In this case, it was a choice that ended up having a huge impact on the rest of the series, and it influenced many other judgment calls to come, to an extent that I’m not sure I recognized at the time.

Chapter 59 of Eternal Empire, for example, is maybe the bloodiest sequence in the entire trilogy, in emotional impact if not in raw body count: it includes the deaths of two major characters and a fair amount of collateral damage. I get rid of Asthana, whom I liked so much that I kept her around for an entire novel after I originally planned to dispose of her, and Vasylenko, whose presence has haunted the series from the start. Looking back on it, I’m pretty happy with Asthana’s swan song, which consists of a complicated set of feints and maneuvers against Wolfe. It’s fair to both characters, and it gives Asthana a second or two to process how she’s been outsmarted. (I wasn’t thinking of Arrested Development, but it’s hard for me to read it now without imagining Asthana saying to herself: “I’ve made a huge mistake.”) But I’m not particularly pleased by how I handled Vasylenko’s death, which is too bad, since by all rights it ought to be the climax of all three books. In some ways, I wrote myself into a corner: there’s really no plausible way to keep Vasylenko alive, or to extend his confrontation with Ilya for longer than a couple of paragraphs, and in my eagerness to write a definitive ending to the series, I may have rushed past the moment of truth. In my defense, the chapter has to provide closure for multiple pairs of characters—Ilya and Maddy, Ilya and Wolfe, Ilya and Vasylenko, Wolfe and Asthana—and I do what I can to give each of them the valediction they deserve. If I had to do it over again, I might have toyed with switching Asthana and Vasylekno’s final scenes, in order to close the novel on a position of greater strength, but this probably wouldn’t have been possible. The Icon Thief ended with Maddy asking Ilya to spare another man’s life; Eternal Empire had to conclude with her asking for the opposite. They don’t end in the same way. But Maddy isn’t the same person she was when we started…

“But we need to work together…”

Note: This post is the fifty-ninth installment in my author’s commentary for Eternal Empire, covering Chapter 58. You can read the previous installments here.

I don’t have the numbers to back this up, but I have a hunch that most professional writers rarely go back to reread their own work. In an interview with The Paris Review, the novelist François Mauriac puts his finger on why revisiting a published story can be such an unpleasant experience:

I only reread my books when I have to in correcting proofs. The publication of my complete works condemned me to this; it is as painful as rereading old letters. It is thus that death emerges from abstraction, thus we touch it like a thing: a handful of ashes, of dust.

The more you unpack this statement, the more insightful it becomes. Reading one of your published stories is like reading an old letter in several ways: it confronts you with the image of yourself when you were younger, it makes your mistakes more visible in hindsight, and it shows you how insidiously the present has turned into the dead past. It’s the fossilized remnant of a process that used to be alive, and as soon as a work of art is locked into its final form, you see all kinds of problems with it. This isn’t necessarily because you could do any better now, but because a story on the page always seems less interesting than it did in your head. When you’re experiencing the work of other writers, you rarely dwell on how else it might have been done, but when you’re reading your own stuff, it’s hard to think about anything else.

This kind of estrangement from a work to which you devoted so much time and energy is unbearably sad—or it would be, if the writer didn’t immediately move on to the next thing. And it explains why the rare story that you can enjoy for its own sake becomes so precious. Usually, it’s something that came fairly easily, as if you were simply transcribing a moment of inspiration that descended from somewhere higher up, or rose from the depths of the subconscious. Isaac Asimov called it “writing over my head,” saying: “I occasionally write better than I ordinarily do…When I reread one of these stories or passages, I find it hard to believe that I wrote it, and I wish ardently that I could write like that all the time.” (Asimov said that he cried whenever he reread the ending of his own story “The Ugly Little Boy.”) Alternatively, you can feel safely detached from one of your own works if you were operating as an artist for hire, without much of a personal stake in the result, but did your job at a high level of technical proficiency. Steven Spielberg has said that the only one of his movies that he can watch with his kids as if he hadn’t directed it, rather than remembering what it was like on the set each day, is Raiders of the Lost Ark. You can see why: it was George Lucas’s baby, and what Spielberg brought to the project was a matchless eye and a useful degree of distance from the material. And I’m not surprised that the result delights him as much as it does me.

When it comes to my own work, there’s almost nothing that I can read now for my own pleasure. Occasionally, like Mauriac, I’ll need to correct page proofs, and I always have to gather my courage a bit: you’re strictly limited in the number of changes you can make, and you can’t imperceptibly massage the text in the way you can when you’re fiddling with a draft in Word. Reviewing proofs shortly after you’ve finished a story is even wore than reading an old letter—it’s like encountering an ex-boyfriend or girlfriend soon after a breakup, when you realize that you’ll never be able to take back what happened. (Not every writer feels this way, and some, like James Joyce, notoriously rewrote entire sections of the manuscript in galley form. But I’ve always assumed that making extensive changes at this stage will only introduce unforeseen complications, so I try to restrict myself to altering a word or a punctuation mark here and there.) Even after my feelings have cooled and a story sits on the shelf like a dead thing, it’s hard for me to look at it again: it’s like being confronted with your irrevocable life choices all at once. And if I had to make a list of the bits and pieces of my fiction that I wouldn’t mind reading again, it represents a tiny slice of the whole: maybe “The Boneless One,” most of “Kawataro” and “The Whale God,” the second half of “Ernesto,” the closing summation in The Icon Thief, and the plane crash and tunnel chase in City of Exiles. That’s about it.

In most of these cases, I was writing over my head, either because I was following up on a good idea that seemed to come out of nowhere, or because I was able to subordinate myself to the mechanics of a plot that I’d already set in motion. And of all the pages I’ve published, Chapter 58 of Eternal Empire might be my favorite—which is to say, if you forced me to pick something to read again, it’s the one I’d probably chose. It isn’t the most complex or difficult thing I’ve written: once I knew that Wolfe and Ilya would team up to take down a dacha full of gangsters and save Maddy, it was mostly just a matter of not screwing it up. But I had a great time writing it, and I still have a good time reading the result. The confluence of names I mentioned above is part of the reason why: it’s one of the few occasions when I felt that I was writing fanfic for my own creations, not because I was indulging myself, but because it combined characters for a payoff that I never would have imagined when I wrote the first book in the series. It’s obviously indebted to scenes like the shootout at the Victory Motel in both the novel and the film versions of L.A. Confidential, and if Wolfe at the climax of City of Exiles slipped into the Clarice Starling of The Silence of the Lambs, she’s closer here to the Starling of Hannibal. It’s the finest moment for my favorite character in the trilogy, which is reason enough for me to like it. Throughout this entire author’s commentary, I’d been looking forward to writing about it, but now that I’m here, I find that I don’t have much to say except that I think it’s pretty damned good. And I’m going back to read it again now…

“This had never been a game of chance…”

Note: This post is the fifty-eighth installment in my author’s commentary for Eternal Empire, covering Chapter 57. You can read the previous installments here.

Earlier this week, in my discussion of Michael Cimino and The Deer Hunter, I managed to avoid mentioning its single most famous—and controversial—plot point. Here’s what William Goldman had to say on the subject in Adventures in the Screen Trade:

Does anyone remember, say, the last part of Deer Hunter? Saigon is going up in flames, and Robert De Niro…is out of service and back in Pennsylvania. He hears about his old buddy, Christopher Walken, who’s still back there…Do you know what Walken has been doing all this time? He’s been playing that game of Russian roulette with real bullets. (The Russian roulette ploy was made up by the movie’s creators, by the way; it didn’t happen in reality.) For months and months, Walken has been taking on all comers in this loony tunes Russian roulette, and…he’s undefeated, untied, and unscored on.

It would take a computer a while to give the odds against that happening, but never mind, because now we’re into the confrontation scene. De Niro versus Walken at Russian roulette. If you looked at the billing of the picture on your way in, did you ever doubt who was going to win?

Obviously, De Niro survives and Walken dies. Goldman concludes that The Deer Hunter, for all its trappings of realism, is ultimately “a comic book movie,” and he adds: “What Deer Hunter told me was what I already knew and believed in: No matter how horrid the notion of war, Robert De Niro would end up staring soulfully at the beautiful, long-suffering Meryl Streep.” And while the film’s Russian roulette sequences are far from its only implausible element, they’ve always served as a focal point for the movie’s critics, both because of their air of racism and because they were invented by the screenwriters out of thin air. What really fascinates me, though, is that these scenes were actually the seed of the entire story, and they came before Vietnam, Pennsylvania, or anything else. The producer Michael Deeley had bought a script called The Man Who Came to Play about games of Russian roulette in Las Vegas, which he called “a very clever piece of writing,” and it was rewritten by Cimino and his collaborator Deric Washburn to take place during the war. You could almost say that these scenes, as arbitrary as they seem in relation to the real Vietnam experience, are what is truly essential, and the rest—all that loving atmosphere at the steel mill and the wedding and the deer hunt and Chopin’s Nocturne—is incidental. And despite my mixed feelings about the movie, I have to concede that Cimino’s fundamental instinct, which was that the Russian roulette element would provide a spine strong enough for him to tell literally any story he wanted, was brilliant. As Roger Ebert, who liked The Deer Hunter far more than I did, wrote, it becomes “the organizing symbol of the film.”

Elsewhere, I’ve said that discovering this kind of narrative trick can feel like stumbling across a new industrial process, and I don’t think that’s ever been more true than it is here. Russian roulette, as a tool for generating suspense, is a writer’s dream: it’s infinitely expansible and compressible, meaning that it can be used to fill thirty seconds of screen time or serve as the motor that drives an entire third act, and it requires a minimum of setup. I’ve often suspected that the whole legal procedural genre sprang up around the fact that a jury delivering its verdict is the most foolproof scene in all of drama: even if the outcome is foreordained, when the foreman passes the folded note to the judge and the defendant is asked to rise, there’s always an increase in tension and anticipation. The trouble—if you’re a writer with the right amount of laziness, which is just another word for the pragmatic use of your limited resources—is you can’t just jump into a verdict scene without any preparation. It requires a fair amount of work to get there. Russian roulette, for better or worse, is a self-contained component: you can slide it in almost anywhere and it works, if only on the most primitive levels of the brain. It delivers violence, or the threat of it, at an unpredictable time in a structured way. I can’t think of anything else in fiction or real life that comes even close to it. Fortunately, perhaps, it’s the sort of thing that can only be done once on this kind of scale. As much as I dislike The Deer Hunter, I almost feel that Cimino deserved to win Best Picture, if only because he recognized the opportunity that the device presented and capitalized on it before anyone else ever could.

Of course, this hasn’t prevented other opportunistic writers from occasionally making use of it. (Among other things, it provides the backbone for the final act of my favorite episode of The X-Files.) And I resort to it here, in Chapter 57 of Eternal Empire, for all the reasons that I mentioned above. Pragmatically, the scene could be about anything or nothing: Maddy has been brought back by her enemies to the isolated dacha in Sochi, and the chapter’s only function is to crank our concern for her safety up to as high a pitch as possible, in roughly five pages, before Wolfe and Ilya storm the compound. A scene like this has to walk a fine line, and I do what I can to give Maddy as much agency as I can, as she tries to turn her captors against one another. But when Vasylenko takes out his revolver and removes all of the cartridges except for one, suddenly it’s all business, and you can almost sense me, as the writer, looking ahead to the next chapter and seeing that I have only a page or two to get my point across. It helps, obviously, that we’re in Russia itself, and Vasylenko’s brief excursus on the history of the game—which I lifted from James H. Billington’s The Icon and the Axe—goes a long way toward justifying it in my eyes. And Maddy’s final revelation, which is that none of this has been a game of chance, is really a character’s glimpse of her author. Like Cimino, I’ve rigged the game to get her here. In the end, the scene works, and Vasylenko doesn’t even need to pull the trigger. That’s the beauty of it. And it’s also why it still makes me a little uneasy…

“In the lights of the cameras…”

Note: This post is the fifty-seventh installment in my author’s commentary for Eternal Empire, covering Chapter 56. You can read the previous installments here.

Few creative choices are so central to the writing process as the selection of a point of view, but it’s often a haphazard, instinctive decision. Unless you’re working in an overtly experimental mode, you’re usually stuck with the first or third person, which isn’t as straightforward as it sounds. It helps to visualize your set of options as a scatter plot, with the dots growing denser around two blobs that we call the first- and third-person point of view—although the boundaries are fuzzy, and there’s a wide range of possibilities within each category. When a writer begins a story, he or she usually selects a point of view from the start, but it’s only in the act of writing itself that the style settles into a particular spot on the spectrum, which can be further refined at the revision stage. The first person is slightly more limited in scope, which is why an author like Henry James, who called it “the darkest abyss of romance,” could claim that it was inherently unsuited to the novel. But it clearly has its uses, and there are even signs that genre readers have come to prefer it. It opens up delicious possibilities for unreliable narrators, who threaten to become a cliché in themselves, and the intimacy that it creates, even if it’s an illusion, can encourage a greater identification with the protagonist. (Hence the fact that the Nancy Drew series switched from the third person to the first about a decade ago, which feels like a sign of the times.)

I made the choice long ago to write my fiction in the third person, and it has remained pretty much in place for everything I’ve published, for both short stories and novels. (The one exception is my story “Ernesto,” which is set to be reprinted soon by Lightspeed: I gave it a first-person narrator as an homage to Holmes and Watson and to discourage me from attempting a bad imitation of Hemingway.) By design, it’s a detached style: I never dip into interior monologue, and even strong emotions are described as objectively as possible. For the most part, I’m comfortable with this decision, although I’m also conscious of its limitations. As far as I can recall, I arrived at it as a form of constraint to keep certain unwanted tendencies in check: these novels are violent and sometimes implausible, and I developed a slightly chilly voice that I thought would prevent the action from becoming unduly hysterical or going out of control.  I wanted it to be objective, like a camera, so that the reader would be moved or excited by events, rather than by the manner in which they were related. Looking back, though, I sometimes wish that I’d modified my approach to give me the option of going deeper into the protagonist’s thoughts when necessary, as Thomas Harris sometimes does. By keeping my characters at arm’s length, I’ve limited the kinds of stories I can tell, and while I don’t mind staying within that range, it also means that I didn’t devote time to developing skills that might be useful now.

That said, I still prefer the third person over the first, and I especially like how it can be imperceptibly nudged in one direction or another to suit the demands of the story. This comes in handy when you’re writing what amounts, in places, to a mystery novel. When you’re working in the first person, it can be hard to conceal information from the reader without it feeling like a gimmick or a cheat—although a few authors, like Agatha Christie, have pulled it off brilliantly. The third person allows you to pull back or zoom in as necessary to manage the reader’s access to the plot, and when you’re working in an omniscient mode that allows you to move between characters at will, you can even cut away entirely. These tricks have been baked into the third person as we’ve come to accept it, so a reader, ideally, will accept such shifts without thinking. (It’s possible to take this kind of switching too far, of course, which is why I try to stick with a single point of view per chapter, and I’m never entirely happy with my attempts to cycle between characters within a single scene.) When an author’s style is inherently objective, we aren’t likely to notice if it retreats or advances a little, any more than it registers when a movie cuts from a medium long shot to a medium shot. And if I’ve remained faithful to that style, it’s largely because it’s more flexible than it seems, and its gradations don’t tend to call attention to themselves.

There’s a good functional example of this in Chapter 56 of Eternal Empire. The first two pages are unusual in that they’re effectively told from nobody’s point of view: they relate a series of events—the explosion of the shadow boat, the movements of reporters, the arrival of the evacuees on shore, and the withdrawal of three unidentified figures to a distant part of the quay—as if recounting them in a news dispatch. (In fact, this is literally what is happening: a big chunk of the section is described as if it were being seen by a viewer on a newscast. If I repeatedly mention the camera crews, it’s to provide an artificial viewpoint through which to narrate the action. The lack of a central character is disguised because a camera has taken its place, which isn’t a tactic that can be extended indefinitely, but it works well enough to get me to the second page.) The reason is obvious: I don’t want to reveal that the three men who have detached themselves from the crowd are Orlov, Ilya, and Tarkovsky, whose fate up to this point has been left up in the air. This wouldn’t work at all in the first person, and if it works here, it’s because I’ve established a style that allows, when the plot calls for it, for the removal of the characters entirely. Very little of this was conscious, but it was all built on a choice of tone that I made two novels earlier, on the hunch that it would lend itself to the kind of story I wanted to tell. A paragraph or two later, we’re back in Ilya’s head. And if I’ve pulled it off properly, the reader should never notice that we left it at all…

“If she was going to run, it had to be now…”

Note: This post is the fifty-sixth installment in my author’s commentary for Eternal Empire, covering Chapter 55. You can read the previous installments here.

In general, an author should try to write active protagonists in fiction, for much the same reason that it’s best to use the active voice, rather than the passive, whenever you can. It isn’t invariably the right choice, but it’s better often enough that it makes sense to use it when you’re in doubt—which, when you’re writing a story, is frankly most of the time. In The Elements of Style, Strunk and Write list the reasons why the active voice is usually superior: it’s more vigorous and direct, it renders the writing livelier and more emphatic, and it often makes the sentence shorter. It’s a form of insurance that guards against some of the vices to which writers, even experienced ones, are prone to succumbing. There are few stories that wouldn’t benefit from an infusion of force, and since our artistic calculations are always imprecise, a shrewd writer will do what he or she can to err on the side of boldness. This doesn’t mean that the passive voice doesn’t have a place, but John Gardner’s advice in The Art of Fiction, as usual, is on point:

The passive voice is virtually useless in fiction…Needless to say, the writer must judge every case individually, and the really good writer may get away with just about anything. But it must be clear that when the writer makes use of the passive he knows he’s doing it and has good reason for what he does.

And most of the same arguments apply to active characters. All else being equal, an active hero or villain is more engaging than a passive victim of circumstance, and when you’re figuring out a plot, it’s prudent to construct the events whenever possible so that they emerge from the protagonist’s actions. (Or, even better, to come up with an active, compelling central character and figure out what he or she would logically do next.) This is the secret goal behind the model of storytelling, as expounded most usefully by David Mamet in On Directing Film, that conceives of a plot as a series of objectives, each one paired with a concrete action. It’s designed to maintain narrative clarity, but it also results in characters who want things and who take active measures to attain them. When I follow the slightly mechanical approach of laying out the objectives and actions of a scene, one beat after another, it gives the story a crucial backbone, but it also usually leads to the creation of an interesting character, almost by accident. If nothing else, it forces me to think a little harder, and it ensures that the building blocks of the story itself—which are analogous, but not identical, to the sentences that compose it—are written in the narrative equivalent of the active voice. And just as the active voice is generally preferable to the passive voice, in the absence of any other information, it’s advisable to focus on the active side when you aren’t sure what kind of story you’re writing: in the majority of cases, it’s simply more effective.

Of course, there are times when passivity is an important part of the story, just as the passive voice can be occasionally necessary to convey the ideas that the writer wants to express. The world is full of active and passive personalities, and of people who don’t have control over important aspects of their lives, and there’s a sense in which plots—or genres as a whole—that are built around action leave meaningful stories untold. This is true of the movies as well, as David Thomson memorably observes:

So 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.

One of the central goals of modernist realism has been to give a voice to characters who would otherwise go unheard, precisely because of their lack of conventional agency. And it’s a problem that comes up even in suspense: a plot often hinges on a character’s lack of power, less as a matter of existential helplessness than because of a confrontation with a formidable antagonist. (A conspiracy novel is essentially about that powerlessness, and it emerged as a subgenre largely as a way to allow suspense to deal with these issues.)

So how do you tell a story, or even write a scene, in which the protagonist is powerless? A good hint comes from Kurt Vonnegut, who wrote: “I don’t praise plots as accurate representations of life, but as ways to keep readers reading. When I used to teach creative writing, I would tell the students to make their characters want something right away—even if it’s only a glass of water. Characters paralyzed by the meaninglessness of modern life still have to drink water from time to time.” This draws a useful distinction, I think, between the two functions of the active mode: as a reflection of reality and as a tool to structure the reader’s experience. You can use it in the latter sense even in stories or scenes in which helplessness is the whole point, just as you can use the active voice to increase the impact of prose that is basically static or abstract. In Chapter 55 of Eternal Empire, for example, Maddy finds herself in as vulnerable a position as can be imagined: she’s in the passenger seat of a car being driven by a woman whom she’s just realized is her mortal enemy. There isn’t much she can plausibly do to defend herself, but to keep her from becoming entirely passive, I gave her a short list of actions to perform: she checks her pockets for potential weapons, unlocks the door on her side as quietly as she can, and looks through the windshield to get a sense of their location. Most crucially, at the moment when it might be possible to run, she decides to stay where is. The effect is subtle, but real. Maddy isn’t in control of her situation, but she’s in control of herself, and I think that the reader senses this. And it’s in scenes like this, when the action is at a minimum, that the active mode really pays off…

“I need you to get me some information…”

Note: This post is the fifty-fifth installment in my author’s commentary for Eternal Empire, covering Chapter 54. You can read the previous installments here.

“You know, like [James] Bond doesn’t have scenes with other men,” the screenwriter Jez Butterworth once told The New Yorker. “Bond shoots other men—he doesn’t sit around chatting to them. So you put a line through that.” To be fair, this isn’t entirely true: Casino Royale, which is probably the finest installment in the whole canon, has an entire second act that consists of little except for Bond chatting in a room with other men. But I understand his point. Few of the Bond films have what we might describe as conventionally good scripts, but they remain useful as a kind of laboratory for a certain sort of film writing, produced under conditions of high pressure. The history of the series, its basic formula, and the need to please a star and a handful of production executives who are answerable to nobody else all create an incentive to cut everything that doesn’t enhance the brand. If it doesn’t fit, you put a line through it. Butterworth’s track record here isn’t perfect—he worked on both Skyfall and Spectre, with notably mixed results in the latter case—but it’s still worth listening to what he has to say. And when you look at the Bond films through the lens of removing whatever isn’t central to what the franchise represents, it goes a long way toward explaining some of their more inexplicable moments.

Take the uranium bullets in Skyfall. If you’ve seen the movie, you probably remember that after Bond’s return from the dead, he removes a fragment of a slug from the wound in his shoulder and hands it over to the lab for testing. The results reveal that it’s made of depleted uranium, which only three assassins in the whole world are known to use, and a glance at their photographs allows Bond to narrow it down to one. Setting aside the fact that being struck by such a bullet should have cut Bond in half, or that MI6 evidently failed to perform any such analysis on the hundreds of spent rounds at the scene, it seems rather careless for a hired killer to leave such a distinctive calling card. You could invent a rationale for this—perhaps the assassin deliberately wants to put his signature on every hit—but it still takes us out of the movie for a few seconds. When you apply Butterworth’s rule, though, you can start to see the reasoning behind it. The plot point with the bullet feels a lot like an attempt to compress what used to be two beats into one. If the lab had managed to narrow down the universe of possible killers using, say, passport tracking, and then had layered the ballistic analysis on top to drill down even further, we probably wouldn’t have thought twice about it. But it would have meant a longer scene of Bond chatting with a man, so it had to go.

The trouble, obviously, is that when you push this kind of compression too far—or at the wrong moment in the story—the audience is likely to object. As a rule of thumb, viewers or readers tend to be more willing to follow the hero through a few intermediate steps of reasoning in the first act than in the third, so the kind of shortcut that Skyfall presents here might well have gone unnoticed in the last twenty minutes of the movie, when we’ve been conditioned to expect the narrative to take a logical leap or two for the sake of momentum. (In this connection, I always think, for some reason, of the penultimate scene of Die Hard With a Vengeance, in which John McClane figures out the villain’s whereabouts using the words on the bottom of an aspirin bottle. We buy it, sort of, because we’re so close to the climax, but it wouldn’t have worked at all earlier on.) A thriller is engaged in a constant balancing act between plausibility and forward movement, and the terms of the equation shift based on where you are in the plot. It isn’t just a matter of what you do, but of when you do it. In general, you can get away with greater gaps in the logic when the surrounding action is furious enough to drown out any implausibilities that would seem glaring if the characters were conversing in a quiet room. Which, in fact, goes a long way toward explaining why Bond can’t be shown chatting with other men: with every such scene, the plausibility of the narrative, which is already so tenuous, comes closer to collapsing entirely.

You can clearly see this principle at work in Chapter 54 of Eternal Empire. Wolfe has just arrived in Sochi in the aftermath of the drone attack, and in order for the plot to proceed, she needs to figure out the location of the launch site with nothing but the information she has at her disposal. The steps in her deductive process are, I think, fairly plausible. The attackers would have wanted to stay off the satellite networks; the drone would have been controlled through line of sight; given its size and the number of rockets it fired, it would have needed enough room for takeoff, or maybe even a pneumatic launcher; and it would have required privacy and a high level of security. Glancing at a map, she concludes that the assault must have been launched from a dacha to the north of the port. Powell, on the phone, says that he’ll pass along whatever he finds, and we later learn that she identified the correct location on her third try. Looking back at the scene from the distance of a few years, I think that I compressed her chain of reasoning just enough, especially because there are so many competing forces at this stage that are propelling the narrative forward. It’s particularly instructive to compare it to the similar series of deductions that Wolfe makes in Chapter 9, which end with her finding a body in East Acton. In that case, instead of unfolding in a couple of paragraphs, it occupies a few unhurried pages—which was just right for that point in the story. The closer you are to the end, the faster it has to be. And you always have to keep your target in sight…

“The deckhand cut the engine…”

Note: This post is the fifty-fourth installment in my author’s commentary for Eternal Empire, covering Chapter 53. You can read the previous installments here.

I think it’s fair to assume that most writers these days—fiction and otherwise—do much of their research online. Yet this seems to make some people uncomfortable, as if it were a violation of the implicit contract between author and reader. One possible analogy is to hiring a consultant or an expert witness: you’d feel cheated if you paid somebody for their time and expertise, only to find that he or she was simply googling the answers to all your questions. Indeed, there’s a strong case to be made that the ease of online research circumvents the lengthy period of absorption in a subject that makes true understanding possible: when you get the answer too quickly, you run the risk that your brain won’t be ready for it. I’m willing to believe this, but I have a problem with the idea that the accessibility of so much information somehow renders the novelist less special. As Jonathan Franzen notoriously said to The Guardian: “When information becomes free and universally accessible, voluminous research for a novel is devalued along with it.” But Franzen writes novels that all but advertise the tremendous amount of legwork they required, which implies that it isn’t whether or not you do research that matters, but what kind it is. The novelist’s role, it seems, is to add value to the data that we have at our disposal, and typing a query into a search engine doesn’t cut it.

This is true enough, but it misses the more important point, which is that a novelist can add value even to just the sort of information that we’re likely to come across on the Internet. As anyone who has spent any time researching a complicated issue online has learned, it isn’t simply a matter of clicking on the first search result that you see: you often have to pull together facts from many places, while constantly monitoring the credibility of your sources, and the result is a pattern that didn’t exist before, even though all of the pieces were theoretically available to everyone. This kind of assemblage poses problems of its own, which is part of the reason I spent so much of The Icon Thief using that process to invent a conspiracy theory: it’s what it does best. But when kept under control, and in the right context, that kind of exercise can be immensely valuable as a way of generating beautiful ideas. (Most of the fiction I’ve published in Analog wouldn’t exist if it weren’t for the tools that we have for combining and associating bits of isolated data into a surprising shape.) A novelist can also perform a useful public service in the form of a deep dive—the singleminded pursuit of material to fill out a gap in the narrative. Sometimes this means spending weeks researching something that will end up as a couple of sentences, or even get cut altogether. And a writer is obsessive enough to invest that level of attention on the most trifling of details.

For instance, I spend a lot of time downloading and reading user manuals. Usually, this is because there’s an arcane piece of hardware in a story that I need to accurately describe, at least to the extent of knowing the right names for all the parts or the actions required for every step in the process. Poking around a little online, once I’ve managed to figure out which search terms to use, often yields useful technical material, especially when I add “filetype:PDF” to the query, and I’ll sometimes spend whole days poring over manuals and spec sheets to make sure I’m getting everything right. The danger, of course, is that the story can turn into a sort of user’s manual itself, in which the author is unwilling to part with any of the factual background that he has so laboriously acquired. (I’m keenly aware of this problem, because it’s particularly troublesome in science fiction and suspense, which are the two genres in which I’ve done the most work.) But it helps to keep the true purpose of this kind of research in mind. As I’ve said many times before, this isn’t about accuracy, which is an incidental benefit; it’s about providing material for dreams. The kind of step-by-step description we find in a user manual, in particular, can serve as a useful backbone or scaffold for the action. If every scene, as I’ve said elsewhere, can be structured as a series of objectives, a manual is nothing less than a one-act play with a beginning, middle, and end, and it climaxes when the user triumphantly completes the repair or installation.

Obviously, it’s possible to take this sort of thing too far, and we’ve all had the experience of reading a story that fails to keep the research where it belongs—in the background. But when properly utilized, it can give shape to the more important things taking place up front. Chapter 53 of Eternal Empire, for example, is all about the evacuation of the sinking yacht after it has been attacked by a drone, which is the kind of big set piece that can quickly degenerate into a bunch of disconnected fragments. Being able to accurately describe, say, the procedure for lowering a lifeboat was a huge deal for me. If you read the scene closely, you see that the characters are almost always doing something, even while more important narrative material receives the most emphasis. It’s a little like the bits of business, like lighting a cigarette, that actors use to do something with their hands, except that every tiny action feeds back to the throughline of the scene, which is to escape safely from the yacht. Much of it came from the lifeboat and sea safety manuals I found online, which briefed me on the vocabulary I needed to get the sequence onto the page. If the reader notices it, I haven’t done my job, but without it, the scene would just lie there. And it’s the kind of research that would be prohibitively difficult if it weren’t for the combination of easy accessibility and the writer’s willingness to spend ungodly amounts of time on it. Writers can and should read user manuals. After all, they may be the only ones who ever will…

“They never would have given up so easily…”

Note: This post is the fifty-third installment in my author’s commentary for Eternal Empire, covering Chapter 52. You can read the previous installments here.

I frequently discuss action sequences on this blog, both because I enjoy thinking about them and because they’re a place in which all the familiar challenges of good writing rise to an unusual pitch of intensity. This might not seem like the case, when we look at how most action scenes in movies are made: many are all but outsourced to the stunt team and second unit crew, and your typical screenplay will often just state that a fight or a car chase ensues without attempting to block out the individual beats. Yet this is almost always a mistake. As I’ve noted here before, my favorite action scenes of recent years—notably the ones in Mission: Impossible—Ghost Protocol, The International, and Drive—all have one thing in common: they seem to have been worked out in detail on the page. An action sequence isn’t a good place for the screenwriter to abdicate responsibility; if anything, it’s the opposite. When I talk about the importance of structuring a plot as a series of clear objectives, the primary reason is to keep the reader or viewer oriented while we focus on the dialogue and the characters and the atmosphere and everything else that made us want to write the story in the first place. A sequence of objectives is the backbone that, paradoxically, gives the writer the freedom to indulge himself. And if that’s true of writing in general, it especially applies to action, in which narrative clarity is all too vulnerable to being swallowed up by sound and fury.

In fact, when we talk about great action scenes, we’re usually talking about the clarity of their writing, often without even knowing it. By now, it’s a critical cliché to complain about the visual grammar of modern action movies, in which an otherwise straightforward sequence is cut into countless tiny pieces of film shot using a shaky camera. (The classic example has quickly become the moment in Taken 3 that uses fifteen cuts to show Liam Neeson jumping over a fence.) In almost the same breath, we usually add that one of the few directors who can do it properly is Paul Greengrass, and that his use of the technique in the Bourne movies has inspired countless imitators to do the same thing less well. This is true enough—but it misses the real point, which is that these scenes work mostly because we know what Bourne is doing and why. This isn’t to understate the sheer technical facility required to take all those brief flashes of the action and assemble them into something coherent in the editing room. But it’s the script, which lays out the situation and the big blocks of the scene in a logical sequence of decisions, that allows for so much visual chaos and excitement. If anything, the editing style obscures the clean lines of the story, which are more obvious in a scene like the lovely opera house set piece in Mission: Impossible—Rogue Nation, a gorgeous example of an action sequence that unfolds almost novelistically in its series of logical complications.    

I can’t help but think of this when I go back to look at Chapter 52 of Eternal Empire, which is one of the few sequences anywhere in my work that I can enjoy without reliving the act of writing it. In part, this is because it’s been long enough since I wrote it that the details have started to blur—although I do remember being nervous about it. I knew from the start that this would be the action centerpiece of the entire novel, if not the whole trilogy, that it would have to cover a lot of plot points in a limited space, and that it hinged on the accurate depiction of a complicated event, in this case an attack on a megayacht by a drone. (I faced a similar challenge in City of Exiles, when it came to describing the sabotage of a private plane and its subsequent crash.) As usual, I started by gathering up all the information I could find on the subject, with the assumption that I could structure the ensuing scene around whatever facts I had available. In the end, many of the beats and much of the language in this chapter came from a little book I found called Megayacht: True Stories of Adventure, Drama, and Tragedy at Sea. None of the incidents it described exactly matched the situation I was writing, but I was able to cobble together enough in the way of persuasive color to construct what I thought would be a convincing naval disaster. And one story in particular caught my eye: the account of a yacht caught in a storm that had to push the helicopter off its upper deck to avoid being tangled up in the wreckage.

When I first read it, I made a note of that idea for a number of reasons. First, it was an exciting sequence, and the book told it with enough circumstantial detail that I knew I could put a version of it into my novel without having to invent too much else. (The rest was filled in with manuals and technical specifications, and I learned more than I ever wanted about tiedown straps and lashing points.) Second, it gave me a few nice images, my favorite of which is the sight of the helicopter sinking into the water, its navigational lights all going up at once as the circuits shorted out, so that it glowed like a ghost in the sea—an image taken directly from the account in Megayacht. Third, and most important, it gave me a sequence of objectives around which I could build the rest of the chapter. What matters, after all, isn’t the helicopter, but what Ilya is thinking and feeling at the time, and by giving him a concrete task to accomplish, I established a clear direction for a chapter that might otherwise have degenerated into a mishmash of furious action. The attempt to push off the helicopter goes badly, of course, and in my original outline, I had Laszlo, the bosun, simply caught in the wreck as it fell overboard. That didn’t seem all that satisfying, especially since it depended on a moment of uncharacteristic incompetence, and as I was working on the chapter, it occurred to me that the drone should turn back and smash itself like a kamikaze into the yacht. That’s what finally happens, and I still love it. But it wouldn’t have occurred to me at all if I hadn’t put together the other pieces first…

“She stared at the burning ship…”

Note: This post is the fifty-second installment in my author’s commentary for Eternal Empire, covering Chapter 51. You can read the previous installments here.

A few weeks ago, an animated video was released online that showed the sinking of the Titanic rendered in real time—all two hours and forty minutes of it. Watching it in shorter stretches is a surprisingly compelling experience, no matter where you start it playing. You’re struck, above all, by just how little seems to happen for so long, until the horrifying ending, in which the bulk of the ship is sucked underwater in less than ninety seconds. It’s riveting, but it’s also a rebuke to conventional narrative. And the Titanic story presents other problems as well. It’s hard to maintain clarity and momentum on any ship sinking at night. You’re dealing with the constrained geography of a big ship itself, in which it can be hard to keep track of where people are and what they’re doing, much of which has to be described using an unfamiliar vocabulary. Darkness and the sheer number of moving bodies make it even harder to maintain a clean line of action. And the way a nautical disaster unfolds resists the natural pacing of a story, with a moment of disaster followed by long stretches of nothing, as the ship sluggishly lists or sinks. Sometimes, as in the case of the Costa Concordia, there isn’t even a clear end point: it just lies there crippled in the water. Everything either happens so fast that it can’t be grasped or so slowly that even the participants start to feel impatient. If I had to write a script about it from scratch, I’m not entirely sure how I’d structure it. Maybe as a love story?

As I’ve noted here before, James Cameron’s Titanic may not have aged well when it comes to some of its more granular qualities, like dialogue, but it’s unsurpassed at staging and choreographing huge blocks of action, and it does it so invisibly that it’s easy to underestimate how hard it really was. It brilliantly solves all the problems I mentioned above, and it does such a good job of keeping the audience informed and oriented throughout that I’d be tempted to use it as a case study in a screenwriting class. The decision to show us an animated visualization of the ship’s destruction at the beginning of the movie, for instance, with a character from the present explaining the physics behind each stage, is a masterstroke: it delivers exposition and foreshadowing all at once, and it allows us to keep track where we are during the sinking without thinking about it twice. Victor Garber’s lovely scene with the ship’s blueprints accomplishes much the same thing. (I’d even argue that the love story itself, and most of its associated plot points, originated a solution to the problem of how to fill those crucial hours of sinking time in which not much of anything else is happening.) When I began work on Eternal Empire, I was forced to think about all of this more systematically. I was writing a novel about a yacht that had to sink, which meant I had to crack all the narrative issues that it presented—and I ended up watching Titanic itself more than once.

Once I knew that the ship had to be destroyed, I turned my attention to how to do it, which was driven both by practical concerns and by the kind of story I wanted to tell. The easiest way to sink a yacht is to approach it in a smaller vessel and detonate a shaped charge against the hull, as used with such devastating effect against the U.S.S. Cole. I toyed with using a remote-controlled boat, but I quickly discarded the idea, mostly for aesthetic reasons. It just didn’t seem interesting or spectacular enough, and I sensed that it would be hard to stage it in an exciting way. An unmanned boat coming up against a yacht at night didn’t give me much in the way of striking imagery, or even a perspective from which to describe the events, unless I broke one of my own rules as a writer and wrote it from the point of view of an objective observer. Since an attack like this would go unnoticed until the moment the explosives detonated, I just didn’t see a way of structuring it, properly, as a series of escalating beats, and I also thought that it would take too long for the reader to figure out what had happened. Basically, I wanted to be able to provide just the right amount of tension before all hell broke loose: enough to make Maddy aware of the inevitable disaster, but not so much that it dissipated the shock. And I wanted it to be visually memorable.

I finally decided to use a drone. In my defense, I should point out that I wrote this sequence at a time when a drone strike wasn’t a total cliché. (Looking back at my notes, I seem to have finished writing this chapter just a week or so before seeing The Bourne Legacy, which used a drone in a similar way. I was probably troubled by this at the time, but it didn’t stop me from pressing on, which was the right choice: I suspect that most people’s memories of that movie are pretty hazy anyway.) The attack would still be over in the blink of an eye, and it ended up covering less than a page. But the nature of a drone, and the staging of the scene on the water, gave me three useful beats, while an attack by boat would have given me only one. Maddy hears the whine of a rocket just before the shadow boat that accompanies the yacht goes up in flames; she sees the drone circling back in her direction in the sky; and finally she sees another line of white as a second rocket smashes into the yacht itself. The result reads like a continuous rush of action, as it should, but the fact that it can be disassembled into smaller components makes the action just a bit more clear. I had to depart slightly from logic to make it work: in actuality, the yacht should have been the primary target, with the shadow boat coming next, but I wanted Maddy to witness the first explosion before feeling the second. The result, I think, is a sweet little scene. And the attack isn’t over yet…

“There’s something we need to talk about…”

Note: This post is the fifty-first installment in my author’s commentary for Eternal Empire, covering Chapter 50. You can read the previous installments here.

Suspense is usually the most linear of genres, but a lot of thrillers include exactly one flashback. You know the one I mean: it comes near the end, just after the big twist, to explain precisely how you were fooled. In a heist movie, it frequently involves the revelation that the plan you thought the protagonists were following was actually something else entirely, and in films that are heavily dependent on fridge logic, it can reveal that much of the movie you believed you were watching was really an elaborate mislead. At its best, as with the unforgettable flashback that occurs two-thirds of the way through Vertigo, it can singlehandedly justify the whole concept of flashbacks in general; at its worst, in a movie like Now You See Me, it can leave you asking why you bothered taking any interest in the plot at all. And these reveals seem to be becoming more common, as the need to find new variations on old surprises has caused such plots to become ever more convoluted and implausible. (We’re at a point now where a single flashback scene isn’t enough: we’re treated to entire flashback montages, replaying what seems like half of the movie from a different point of view. When handled well, as in The Illusionist, this sort of thing can be delightful, but it can also leave a viewer feeling that the film hasn’t played fair with its obligation to mislead us with what it shows, rather than what it omits.)

This sort of flashback is obviously designed to save a surprise for the end of the movie, which is where we’ve been conditioned to expect it—even if some violence has to be done to the fabric of the narrative to put the reveal in the last ten minutes, instead of where it naturally occurred. This isn’t a new strategy. Jack Woodford, the pulp writer whose instructional book Trial and Error was carefully studied by Robert A. Heinlein, thought that all stories should end with a punch ending, and he offered a very useful tip on how to artificially create one:

A good way to do this is to go ahead and end it with the usual driveling collection of super-climaxes, anti-climaxes and what not that amateurs end stories with, and then go over it, find where the punch ending is, rework the ending so that the anti-climaxes, if there is anything in them at all that really needs to be told, come before the final crux ending.

This is why so many stories contrive to withhold crucial information until the point where it carries the most impact, even if it doesn’t quite play fair. (You frequently see this in the early novels of Frederick Forsyth, like The Odessa File or The Dogs of War, which leave out a key element of the protagonist’s motivation, only to reveal it at the climax or on the very last page. It’s such a good trick that you can almost forgive Forsyth for reusing it three or four times.)

Another advantage to delaying the explanatory scene for as long as possible is that it turns an implausible twist into a fait accompli. I’ve noted before that if there’s a particularly weak point in the story on which the credibility of the plot depends, the best strategy for dealing with it is to act as if has already happened, and to insert any necessary justifications after you’ve presented the situation as blandly as possible. Readers or audiences are more likely to accept a farfetched plot development after it has already been taken for granted. If they had been allowed to watch it unfold from scratch, during the fragile early stages, they would have been more likely to object. (My favorite example is how in the two great American drag comedies, Some Like it Hot and Tootsie, we never see the main characters make the decision to pose as women—we cut to them already in makeup and heels, which mostly prevents us from raising any of the obvious objections.) This explains why the expository flashback, while often ludicrously detailed, rarely shows us the one scene that we really want to see: the conversation in which one character had to explain to the rest what he wanted them to do, and why. Even a classic twist ending like the one in The Sting falls apart when we imagine the characters putting it into words. The act of speaking the plan aloud would only destroy its magic.

I put these principles to good use in Chapter 50 of Eternal Empire, which rewinds the plot slightly to replay a crucial scene in its entirety. Structuring it as a flashback was clearly meant to preserve the surprise, but also to downplay its less plausible angles. For the story to work, Maddy had to reveal herself to Tarkovsky, justify her good intentions, and propose a complicated counterplot, all in the course of a single conversation. I think that the chapter does a decent job of pulling it off, but placing the discussion here, after the effects of the decision have already been revealed, relieves it of some of the weight. The reader is already invested in the premise, simply by reading the events of the preceding chapters, and I hoped that this would carry us past any gaps in the logic. But it’s worth noting that I never actually show the crux of the conversation, in which Maddy spells out the plan she has in mind. Asking a character to fake his death for the sake of some elaborate charade is a scene that can’t possibly play well—which might be why we almost never see it, even though a similar twist seems to lie at the bottom of half of the surprise endings ever written. We don’t hear Maddy telling Tarkovsky what she wants him to do; we just see the results. It’s a form of selective omission that goes a long way toward making it all acceptable. But as the reader will soon discover, the plan hasn’t gone quite as well as they think…

“He wondered if this was something new…”

Note: This post is the fiftieth installment in my author’s commentary for Eternal Empire, covering Chapter 49. You can read the previous installments here.

When you’re taking apart a story to see how it works—whether as a fan, a critic, or an aspiring artist yourself—it’s often necessary to distinguish between two different kinds of narrative elements. One consists of the standard components that appear in the story to satisfy our expectations about the genre: a whodunit, for example, usually has a detective, a killer, a cast of suspects, and a series of clues, and most mystery writers assume that these pieces will be present before they’ve even come up with a premise. The second kind of element is designed to address the unique problems that this story alone presents: the details of character, setting, and incident required to enable the specific plot developments or twists that the author has in mind. And you can’t tell the difference until you’ve read two or more stories in the same genre, just as a scientist might need to examine multiple specimens to distinguish between the traits common to an entire species and individual variation. If you’ve only read one gothic romance, for instance, you have no way of knowing how many of those stock building blocks, from the orphaned heroine to the lonely house on the moor, are there primarily because that’s how the genre has defined itself. And it takes yet another level of sophistication to recognize that most of these elements exist because Charlotte Brontë needed them to tell a particular story in Jane Eyre. You can’t see the pattern until you have at least two examples to compare.

The same principle holds true for criticism—the best way to think about a work of art, after all, is to compare it to something else—and for the creative process itself, particularly when the artist is consciously drawing on more than one previous model. A sequel, even one that departs from the first story in surprising ways, always involves some degree of reverse engineering: the artist looks back at the original and tries to pick out the most important features, which create a sense of continuity between the old and new stories. (This usually, but not necessarily, means keeping some of the same characters, but it can also mean reproducing entire plot points or intangible qualities of tone.) And a third installment can draw upon two different models, which can both confuse and clarify the issue. When you mentally superimpose two stories, it naturally emphasizes the places where they overlap, and for the first time, it becomes possible to figure out which plot elements are essential to the series, rather than incidental solutions to specific problems. This is why an ambitious third installment can serve as a direct continuation of the second while simultaneously circling back to the first. Films like The Dark Knight Rises or The Bourne Ultimatum can struggle to manage those competing chains of meaning, which is why the fourth installment in a series often strikes off in a totally new direction: it becomes mathematically impossible, in the narrative equivalent of Metcalfe’s Law, to keep those connections straight when four or more stories are involved.

But it’s often not until the third installment that we know what the series is really about, since it takes two stories to establish the pattern. When I was trying to figure out the plot of Eternal Empire, I was very mindful of the problem of honoring the precedent set by the first two novels while departing when necessary from the template they had established—if only because it gave me a way to structure a novel that I knew was going to be very complicated. As I wrote in a blog post back when the novel first came out: “In a sense, [Eternal Empire] ends up serving double duty: City of Exiles ends on a cliffhanger that the third novel needed to resolve, but it also reaches further back to the first installment, so the resolutions of these two books essentially unfold in parallel before converging at the very end.” These days, I’m less confident that I pulled this off successfully than I was when the book was released, but I know that I gave it my best shot. And even at the time, I realized, consciously or otherwise, that the best way to keep the plot from spiraling out of control was to copy the structural bones of the previous books wherever I could. As a result, each novel has three acts, cuts between three main characters, covers roughly the same amount of time, and uses a narrative funnel that accelerates the action until the last section encompasses the events of less than a day. These are all solid, reliable thriller techniques, but it wasn’t until I had two prior novels to analyze that I began to understand exactly what kind of story I was writing.

You can see this clearly in the sequence of events that culminates in Chapter 49 of Eternal Empire. Both The Icon Thief and City of Exiles contain a plot twist that occurs at approximately the same point in each novel, involving the sudden death of a supporting character at the hands of an unexpected killer. This isn’t exactly unprecedented in this kind of thriller, and if it works, it’s all in the execution—and I think it works pretty well in both cases. (In fact, the two twists are similar enough that I went to considerable lengths to disguise it. This why the scene in the latter novel takes place in a car, rather than indoors, and it’s told from the point of view of the killer, rather than the victim.) I wanted a big twist in the third novel at more or less the same spot, both because I thought a reader would want it and because it served a useful narrative purpose: it occurs at a point where the story needs a jolt of energy to carry it through the last hundred pages. Because I couldn’t pull the same exact trick yet again, I ended up inverting it: instead of an unexpected death, it would be an expected death that turned out to be something else, with the reveal that Tarkovsky had conspired with Ilya to fake his assassination. It was like taking a piece out of the template, turning it around, and inserting it into the same place again to see if it fit. And it did. But if I hadn’t had two previous novels to study, I might not have known that anything had to be there at all. And I’m a little relieved that this series is over, because I’m not sure I’d be able to do it again…

“Yet she was still a woman…”

Note: This post is the forty-ninth installment in my author’s commentary for Eternal Empire, covering Chapter 48. You can read the previous installments here.

“When I start a play, I’ll think, does it matter if this character is a man or a woman?” David Lindsay-Abaire once said. “And if it doesn’t, I make it a woman.” I do pretty much the same thing. And I’d like to think that we both take this approach for an utterly unsentimental reason: it results in better stories. There’s a tendency for writers, male and female alike, to use male characters as default placeholders, especially in genres that have traditionally been dominated by men. By systematically visualizing women instead—even if it’s nothing more than an initial sketch—you’ve already redirected your thought processes at a slightly different angle, which can only be good for the outcome. Whenever I read stories from the golden age of science fiction, I’m struck by the absence of women, which seems less like a sin than a mistake. It’s hard to think of a story from that era that wouldn’t have been improved by turning half of the men into women, without any other revisions aside from the relevant pronouns, as was done, much later, with Ripley in Alien. And I would have addressed this advice squarely to those pragmatic hacks who were only interested in making a living. There are so few writing rules of any value that a professional ought to utilize anything that works on a consistent basis, and the fact that so many of the women we see in these stories are either love interests or secretaries, even in the far future, feels like a missed opportunity.

There’s even a handy empirical test that you can use to verify this. Take a story from any genre in which the genders of the main characters are mostly irrelevant—that is, in which you could rewrite most of the men as women, or vice versa, while leaving the overall plot unchanged. Now mentally change a few of the men into women. The result, in most cases, is more interesting: it generates registers of meaning that weren’t there before. Now mentally turn some of the women in the original story into men. I’m willing to bet that it has the net opposite result: it actually saps the narrative of interest, and makes the whole thing flatter and duller. If you don’t believe me, just try it a few times. Even better, do it when you’re constructing a story, and see which version you like better. In the book Which Lie Did I Tell?, the screenwriter William Goldman writes:

I remember once being in an office with a studio guy and a couple of people were sitting around, fighting the story. And once of the people said this: “What if they’re all women?” Now the story, as I remember, was a male adventure flick. And the studio guy commented on that—“This is an adventure movie here, how stupid a suggestion is that?” Naturally the writer was finished for that day.

The truth, as Goldman points out, is that it was an excellent idea: “Making them all women opened up the world. I use it a lot myself now.” And that’s all the more reason to do it automatically at the earliest possible stage.

Which isn’t to say that you can just change the names and pronouns and be done with it. This exercise is only useful if you follow through on the implications that come with making a character a woman, especially in a genre like suspense, which defines itself so casually in terms of action and violence. In my novels, you could change most of the women to men without affecting the main outlines of the plot, but there would be a real loss of meaning. In part, this is because I unconsciously situated these characters in worlds in which women face particular challenges. For Maddy, it was the world of art and finance; for Wolfe, of law enforcement; and for Asthana, of thieves and criminals. These tensions are mostly just implied, but I’d like to think that they quietly affect the way we see these characters, who are enriched by the choices they must have made before the story began. In retrospect, this explains, for instance, why Wolfe is so much more interesting than Alan Powell, to whom I devoted a third of The Icon Thief before mostly shelving him in Eternal Empire. Wolfe would have had to prove herself in ways that someone like Powell never would, and it shows, even if it’s unstated. And I have a hunch that my endless struggles with Powell as a character might have been avoided entirely if I’d done the logical thing and made him a woman as well.

There’s another missed chance in this series, and it involves the character of Asthana. The only time I come close to exploring the peculiar position she holds—as a woman of color in a criminal world—is in Chapter 48 of Eternal Empire, in which she enters a house in Sochi occupied entirely by Russian thieves. Her thoughts turn briefly to the fact that she’ll always be regarded as an outsider, and I try to show how she establishes herself in the pecking order by being a little smarter than the men around her. But I don’t do nearly enough. Part of this is simply due to a lack of space, and to the fact that it felt more important to define Asthana in relation to Wolfe. Still, her presence here raises a lot of questions that go mostly unanswered, and I can’t help but feel that I could have touched on them more. (If I were doing it all over again today, I would have remembered what Christopher McQuarrie says about Rebecca Ferguson’s character in Mission: Impossible—Rogue Nation: “They’re not men. They’re women that are not trying to be men…You’re here on your own terms and you’re in a shitty situation created by people in power above you. How do you escape this situation and maintain your dignity?”) If anything, the result would have made Asthana an even more formidable antagonist for Wolfe. And although there’s a showdown coming soon between these two women, the most interesting parts of this story will mostly remain unspoken…

“This was the ending that had awaited him all along…”

Note: This post is the forty-eighth installment in my author’s commentary for Eternal Empire, covering Chapter 47. You can read the previous installments here.

One of my favorite storytelling tricks is the false ending, in which the writer fools us into thinking that we’ve reached a satisfying conclusion, only to pause, regroup, and push forward into something even deeper. The great example here is The Usual Suspects. After listening to Verbal spin his convoluted tale for well over an hour, Detective Kujan turns the tables, bombarding Verbal with a version of events—aided by a barrage of flashbacks over a dramatic underscore—in which Dean Keaton was Keyser Soze all along. It’s a convincing performance, and if you went into the film knowing nothing except that it was supposed to have a famous twist, you might be forgiven for thinking that this was it, even if it wasn’t as good as you expected. Yet the sequence flies in the face of one of the few scraps of objective information that the audience has been given: the very first scene in the movie, in which Keaton dies. And if we temporarily forget this, it’s partially because ninety complicated minutes have unspooled in the meantime, but also because Kujan’s closing argument is assembled to look and sound like the end of the movie. It’s a perfectly decent flashback montage, of the sort that is often used to reveal the solution to a mystery, and we have no way of knowing that the movie is about five minutes away from using an even better montage to blow our minds for real. (The unsung hero here, as I never tire of saying, is editor and composer John Ottman, whose contributions elevate the movie beyond what was there in McQuarrie’s script and Singer’s direction.)

Which, when you think about it, is a surprisingly subtle point. It isn’t the logical consistency of the fake ending that fools us, but the way in which it mimics the visual, rhythmic, and aural conventions of the real endings to which we’re accustomed. We’re subconsciously attuned to how a movie feels as it draws toward its conclusion, and for a fake ending to work, it has to give us the full package, which is more important than whether or not it makes sense. And the absence of such cues can tip us off to the trick prematurely. Zootopia, for instance, has what would otherwise seem like an ingenious fake ending, but the movie rushes past it a little too quickly: if it were the real climax, we’d be savoring it, and the fact that the script treats it in an almost perfunctory way is a clue that we shouldn’t take it seriously. If a movie really wants to trick us, it has to edit that fake ending as if it were the real thing, and in particular, it has to pay close attention to the music, which often tells us what to feel. The score at the end of a movie usually swells to carry us out of the theater, and if many fake endings fail to convince, it’s because they’re too quiet. (I’m surprised at how rarely movies use our knowledge of scoring conventions against us. Movie music often prompts us to feel relieved—as when the score softly creeps in again after a long stretch of silence in which the heroine is exploring the deserted house—and I’d love it if a film gave us a few bars to release the tension, and then the jump scare.)

The fact that movies almost never exploit a fake ending to its logical extent is hard to explain, especially because the medium lends itself so naturally to such a mislead. We know exactly how many pages remain in a book, and we generally have a pretty good idea of how long an episode of a television series will last. With a movie, unless we’re watching it at home and have carefully scrutinized the back of the video box beforehand, we don’t really know how much longer it has to go, and even if we can guess that it’s about two hours, twenty minutes in either direction gives it plenty of room to play with our expectations. (Douglas Hofstadter once jokingly proposed padding out novels with fake pages toward the end, to create the same kind of effect, and I sometimes experience this when a book ends, without my knowledge, with a preview of the next installment in the series.) But if the movies seem reluctant to push that kind of fakeout as far as it can go, it might be because the benefits are canceled out by unanticipated side effects. A really convincing fake ending would have the audience putting on its coats and preparing to exit the theater, only to be yanked back into the story, and that sort of manipulation can easily turn viewers against it. Fooling us into the physiological response created by a real ending might make it impossible for us to respond in that way when the movie actually ends. This might explain why the handful of movies that really sell a fake ending, like Terry Gilliam’s Brazil or James Mangold’s Identity, time it so that it occurs only a few minutes, or seconds, before the real thing, compressing the two into one.

Chapter 47 of Eternal Empire occurs long before the ending of the book, but it includes a narrative fakeout that required me to take many of these issues into account. It’s the culmination of the subplot in which Ilya has been forced to assassinate Tarkovsky, and at the end of the chapter, he appears to do exactly that, shooting Tarkovsky in cold blood in the oligarch’s stateroom. Or at least that’s how it looks. Needless to say, there’s something else going on, and within the next couple of scenes, we’ll be let into the secret plan that has been unfolding in plain sight. When a valued reader gave me notes on the first draft, however, he said that he didn’t buy the scene as written—he knew, somehow, that Tarkovsky was still alive. When I went back to reread the relevant section, I saw my mistake: I had written it as if I knew what was coming. If Ilya had shot Tarkovsky for real, this would have been the tragic endpoint of the entire trilogy: the instant in which his true nature as a killer overtook his attempts to become something more, swept up by circumstances beyond his control. I would have lingered on this moment, which would have been one of the major climaxes of the whole series, and the existing version didn’t give it the attention that it deserved. In the revision, then, I slowed it down, putting in the equivalent of a dramatic orchestral sting to play over Tarkovsky’s apparent death, and I dwelled on it as if the entire book had been building to this passage. Which, in a sense, it had. (The rewrite also gave me my single favorite line in the novel, the description of the yacht as “a masterpiece of foresight and design surrounded on all sides by night.”) Does it work? I can’t say. But at least it has a chance…

“I know who you are…”

Note: This post is the forty-seventh installment in my author’s commentary for Eternal Empire, covering Chapter 46. You can read the previous installments here.

Freedom in writing is a lot like its equivalent in everyday life. In theory, we’re all free actors of ourselves, as Harold Bloom describes the characters in Shakespeare’s plays, but in practice, we’re hemmed in by choices and decisions that we made years ago—or that other people or larger systems have made for us. Similarly, a novelist, who really does have access to limitless possibilities, inevitably ends up dealing with the kind of creative constriction that Joan Didion describes in an observation to The Paris Review that I never get tired of quoting:

What’s so hard about that first sentence is that you’re stuck with it. Everything else is going to flow out of that sentence. And by the time you’ve laid down the first two sentences, your options are all gone…I think of writing anything at all as a kind of high-wire act. The minute you start putting words on paper you’re eliminating possibilities. Unless you’re Henry James.

The difference between writing and life, of course, is that it’s easier for a writer to start over. No matter how much time or effort you’ve put into a project, you can always throw it away and begin again without anyone being the wiser, which is harder to pull off in the real world. Yet many writers stubbornly insist on sticking to what they’ve written, as if they didn’t have any alternative.

And the kicker is that they’re often perfectly right, at least where the first draft is concerned. Earlier this week, I talked about the finite amount of energy that a writer can allocate to the different stages of the creative process, and the strategies that he or she develops to conserve it. Discarding the material you’ve already written is a good way to sap the limited resources you have. More pragmatically, starting over usually amounts to never finishing the story at all: the most common reason that most attempts at a novel peter out halfway through is because the author was unable to live with what he or she had written the day before. A crucial part of becoming a writer lies in learning how to plow ahead despite the shortcomings of the work you’ve done, which often means treating your existing pages as fixed quantities. If nothing else, it’s a helpful strategy for concentrating exclusively on the work at hand: during the first draft, it makes sense to think of what you’ve written so far as inviolate, because otherwise, you’ll be tempted to go back and tinker, when you should be more concerned with the pages you don’t have. Revision requires another mental shift, in which everything is on the table, no matter how much work you’ve invested in it. And that transition—which flips your approach to the rough draft completely on its head—is one that every writer has to master.

But there’s an even more interesting case to be made for preserving the elements you’ve already written, or at least for doing everything you can to work with them before giving up. In the past, I’ve spoken of writing as a collaboration between your past, present, and future selves: it’s too complicated for any one version of you to handle alone, so you set up ways of communicating across time with different incarnations of yourself, aided by good notes. Your existing pages are a message from the past to the present, and they deserve to be taken seriously, since there’s information there that might otherwise be lost. In The Art of Fiction, John Gardner says of a writer working on a story about Helen of Troy:

He begins to brood over what he’s written, reading it over and over, patiently, endlessly, letting his mind wander, sometimes to Picasso or the Great Pyramid, sometimes to the possible philosophical implications of Menelaos’ limp (a detail he introduced by impulse, because it seemed right). Reading in this strange way lines he has known by heart for weeks, he discovers odd tics his unconscious has sent up to him, perhaps curious accidental repetitions of imagery…Just as dreams have meaning, whether or not we can penetrate the meaning, the writer assumes that the accidents in his writing may have significance.

Every writer, I think, will recognize this—but it only works if you trust that your past self knew what he or she was doing. And you’re more likely to come up with useful ideas if you treat that material as irrevocable. Like any constraint, it’s only fruitful if it can’t easily be eluded.

There’s a small but telling example in Eternal Empire of how this works. Way back in Chapter 5, when Maddy visits Tarkovsky’s house for the first time, she sees Nina, the oligarch’s daughter, riding a horse on a polo field, dressed in jodhpurs and a bomber jacket. Nina looks stonily at Maddy for a long second, and then disappears. It provided a nice visual button for the scene, but in order for it to make sense, I had to introduce Nina later on as a character. And I had no idea what role she might play. I had her pop up now and then in the pages that followed, always seen from a distance, just to remind the reader and myself that she still existed. Finally, when I reached Chapter 46—months after writing that first image—I knew that I couldn’t postpone it any longer: from a structural perspective, coming just before the huge set pieces that conclude the novel, it was the last area of calm in which any interaction between Maddy and Nina could take place. I thought more than once about cutting the earlier beat, which only amounted to a few lines, and taking the daughter out of the story entirely. But when I forced the two of them to meet at last, I ended up with what still feels like an essential moment: Nina drops a clue about her father’s plans, as well as to the overall plot, and they feel a brief connection that I knew would pay off in the final act. Without the image of that girl on a horse, which I introduced, as Gardner says, “by impulse, because it seemed right,” none of this would have occurred to me. It took a long time for it to justify itself, but it did. And it wouldn’t have happened at all if I hadn’t acted as if I didn’t have a choice…

“Where all other disguises fell away…”

Note: This post is the forty-sixth installment in my author’s commentary for Eternal Empire, covering Chapter 45. You can read the previous installments here.

Occasionally, a piece of technology appears in the real world that fits the needs of fiction so admirably that authors rush to adopt it in droves. My favorite example is the stun gun. The ability to immobilize characters without killing or permanently incapacitating them is one that most genre writers eventually require. It allows the hero to dispatch a henchman or two while removing the need to murder them in cold blood, which is essential if your protagonist is going to remain likable, and it also lets the villain temporarily disable the hero while still keeping him alive for future plot purposes. Hence the ubiquitous blow to the back of the head that causes unconsciousness, which was a cliché long before movies like Conspiracy Theory ostentatiously drew attention to it. The beauty of the stun gun is that it produces all of the necessary effects—instantaneous paralysis with no lasting consequences—that the convention requires, while remaining comfortably within the bounds of plausibility. In my case, it was the moment when Mathis is conveniently dispatched toward the end of Casino Royale that woke me up to its possibilities, and I didn’t hesitate to use it repeatedly in The Icon Thief. By now, though, it’s become so overused that writers are already seeking alternatives, and even so meticulous an entertainment as the David Fincher version of The Girl With the Dragon Tattoo falls back on the even hoarier device of knockout gas. But the stun gun is here to stay.

Much the same principle applies to the two most epochal technological developments of our time, which have affected fiction as much as they’ve transformed everyday life: the cell phone and the Internet. Even the simple flip phone was a game changer, instantly rendering obsolete all stories that depend on characters being unable to contact one another or the police—which is why service outages and spotty coverage seem so common in horror movies. It’s hard not to watch movies or television from earlier in this century without reflecting on how so many problems could be solved by a simple phone call. (I’m catching up on The People v. O.J. Simpson, and I find myself thinking about the phones they’re using, or the lack thereof, as much as the story itself.) And the smartphone, with the instant access it provides to all the world’s information, generates just as many new problems and solutions, particularly for stories that hinge on the interpretation of obscure facts. Anyone writing conspiracy fiction these days has felt this keenly: there isn’t much call for professional symbologists when ordinary bystanders can solve the mystery by entering a couple of search terms. In City of Exiles, there’s a dramatic moment when Wolfe asks Ilya: “What is the Dyatlov Pass?” On reading it, my editor noted, not unreasonably: “Doesn’t anybody there have a cell phone?” In the end, I kept the line, and I justified it to myself by compressing the timeline: Wolfe has just been too busy to look it up herself. But I’m not sure if it works.

Search engines are a particularly potent weapon of storytelling, to the point where they’ve almost become dangerous. At their best, they can provide a neat way of getting the story from one plot point to the next: hence the innumerable movie scenes in which someone like Jason Bourne stops in an Internet café and conducts a few searches, cut into an exciting montage, that propel him to the next stage of his journey. Sometimes, it seems too easy, but as screenwriter Tony Gilroy has said on more than one occasion, for a complicated action movie, you want to get from one sequence to the next with the minimum number of intermediate steps—and the search engine was all but designed to provide such shortcuts. More subtly, a series of queries can be used to provide a glimpse into a character’s state of mind, while advancing the plot at the same time. (My favorite example is when Bella looks up vampires in the first Twilight movie.) Google itself was ahead of the curve in understanding that a search can provide a stealth narrative, in brilliant commercials like “Parisian Love.” We’re basically being given access to the character’s interior monologue, which is a narrative tool of staggering usefulness. Overhearing someone’s thoughts is easy enough in prose fiction, but not in drama or film, and conventions like the soliloquy and the voiceover have been developed to address the problem, not always with complete success.  Showing us a series of search queries is about as nifty a solution as exists, to the point where it starts to seem lazy.

And an additional wrinkle is that our search histories don’t dissipate as our thoughts do: they linger, which means that other characters, as well as the viewer or reader, have potential access to them as well. (This isn’t just a convention of fiction, either: search histories have become an increasingly important form of evidence in criminal prosecutions. This worries me a bit, since anyone looking without the proper context at my own searches, which are often determined by whatever story I’m writing at the time, might conclude that I’m a total psychopath.) I made good use of this in Chapter 45 of Eternal Empire, in which Wolfe manages to access Asthana’s search history on her home computer and deduces that she was looking into Maddy Blume. It’s a crucial moment in the narrative, which instantly unites two widely separated plotlines, and this was the most efficient way I could devise of making the necessary connection. In fact, it might be a little too efficient: it verges on unbelievable that Asthana, who is so careful in all other respects, would fail to erase her search history. I tried to make it more acceptable by adding an extra step with a minimum of technical gobbledegook—Asthana has cleared her browser history, so Wolfe checks the contents of the disk and memory caches, which are saved separately to her hard drive—but it still feels like something of a cheat. But as long as search histories exist, authors will use them as a kind of trace evidence, like the flecks of cigarette ash that Sherlock Holmes uses to identify a suspect. And unlike most clues, they’re written for all to see…

“Are you still willing to play your part?”

Note: This post is the forty-fifth installment in my author’s commentary for Eternal Empire, covering Chapter 44. You can read the previous installments here.

When you conceive of a story as a kind of puzzle box, one of the most satisfying tricks you can play is to write a scene that can be read in two different ways. At first, it suggests one obvious interpretation—if you’ve done it right, it shouldn’t even raise any questions—but on a second encounter, it says something else, based solely on the fresh perspective that the reader or audience brings to it. The canonical example here is Francis Ford Coppola’s The Conversation. It opens with the paranoid sound expert Harry Caul, played by Gene Hackman, eavesdropping on an illicit meeting in the park between a young couple, Mark and Ann, who are having an affair. Harry has been hired to follow them by Ann’s husband, but later, as he cleans up and edits the tape recording, he hears a line spoken by Mark for the first time: “He’d kill us if he had the chance.” Before long, Harry, who obsessively replays that part of the conversation, becomes convinced that his client is planning to have Mark and Ann killed. Of course, that isn’t what happens, and it turns out in the end that Mark and Ann were planning to murder Ann’s husband. Harry’s interpretation of the recording was wrong: it wasn’t “He’d kill us if he had the chance,” but “He’d kill us if he had the chance,” meaning that they have to kill him first. And it’s only when the audience, along with Harry, glimpses the full picture that the line reveals its real meaning at last.

Which is an amazing feat of storytelling—except that it cheats. Walter Murch, who was left to edit the film by himself after Coppola ran off to film The Godfather Part II, was never able to make the audience understand the true meaning of that critical line of dialogue, and he ultimately hit upon a solution that broke the movie’s own rules. During one take, Frederic Forrest, who played Mark, had flubbed his line reading, inadvertently placing the emphasis on the wrong word: “He’d kill us if he had the chance.” As Murch recounts in Charles Koppelman’s Behind the Seen:

I noted that reading at the time…and filed it away as being inappropriate. But a year later during the mixing of the film I suddenly thought, let’s see what happens if we substitute that “inappropriate” reading with its different inflection into the final reel. It might help tip audiences into understanding what had happened: that the “victims” were really the “plotters.” So I mixed it into the soundtrack in place of the original reading and took the finished film to [Coppola]…I prepared him for the change and wondered what his reaction would be when he heard it. It was a risky idea because it challenged one of the fundamental premises of the film, which is that the conversation itself remains the same, but your interpretation of it changes. I was prepared to go back to the original version. But he liked it, and that’s the way it remains in the finished film.

And it was the right call, even if it was a bit of a cheat. When we look at the books or movies that execute the priceless gag of having a scene appear to mean one thing but turn out to mean another, some degree of trickery is almost always involved. No film has ever pulled it off as beautifully as The Sixth Sense, with its closing montage of moments that we suddenly see in a new light, but on a second viewing, we’re acutely aware of how the script walks right up to the edge of deceiving us unfairly. (My favorite example is Lynn’s line “You got an hour,” which works when we think she’s talking to Malcolm, but not if she’s just telling her son that she’s making some triangle pancakes.) The Usual Suspects cheats even more blatantly by giving us a fake flashback—a gimmick that can be justified by the presence of an unreliable narrator, but which still feels like a lapse in an otherwise elegant movie. It’s also common for a story to omit necessary information, so that the dialogue, while not actively misleading, only gives us part of the picture. You frequently see this in movies like Ocean’s 11 and its sequels, which involve us in the planning of a heist but withhold a few details so that we don’t know what the protagonists really have in mind. In small does, this can be delightful, but it verges on being a cliché in itself, and when taken too far, it violates the implicit contract between the story and the audience, which is that we’ll be allowed to see what the main character does and draw our own conclusions.

Chapter 44 of Eternal Empire represents my own effort in that line, and I’m reasonably happy with how it turned out. The chapter opens at the tail end of what seems like a routine conversation between Maddy and Tarkovsky, then follows Maddy as she goes down to the yacht’s tender bay to meet Ilya, who is evidently preparing for Tarkovsky’s assassination. That isn’t really the case, of course, and I had a good time drawing on the standard bag of tricks for this sort of misdirection. Maddy acts as if she’s scoping out Tarkovsky’s office for the kill, when in fact she’s there to warn him, and her ensuing conversation with Ilya is filled with lines of the “He’d kill us if he had the chance” variety. (“Are we safe?” “If you’re asking if the pieces are in place, then yes, we’re ready.” “And are you still willing to play your part?” “I don’t think I have a choice.”) Looking at it objectively, I’d say that the result does its job with a minimum of jiggery-pokery, although there’s always a touch of cheating—which some readers will hate no matter what—when you don’t reveal everything that your point of view character might be thinking. Fortunately, my usual narrative mode is fairly clinical and detached: I don’t use interior monologue, and I prefer to convey emotion through action, which dovetails nicely with the requirements of a scene like this. The chapter works because it isn’t so far removed from what I normally do as a writer, which allows the characters to keep their secrets. And I’d do it again if I had the chance…

“The mirror shattered into spiderwebs…”

Note: This post is the forty-fourth installment in my author’s commentary for Eternal Empire, covering Chapter 43. You can read the previous installments here.

“I am truly at my happiest not when I am writing an aria for an actor or making a grand political or social point,” Aaron Sorkin said a while back to Vanity Fair. “I am at my happiest when I’ve figured out a fun way for somebody to slip on a banana peel.” I know what he means. In fact, nothing makes me happier than when an otherwise sophisticated piece of entertainment cheerfully decides to go for the oldest, corniest, most obvious pratfall—which is a sign of an even greater sophistication. My favorite example is the most famous joke in Raiders of the Lost Ark, when Indy nonchalantly draws his gun and shoots the swordsman. It’s the one gag in the movie that most people remember best, and if you’re a real fan, you probably know that the scene was improvised on the set to solve an embarrassing problem: they’d originally scheduled a big fight scene, but Harrison Ford was too sick to shoot it, so he proposed the more elegant, and funnier, solution. But the most profound realization of all is that the moment works precisely because the film around it depends so much on craft and clockwork timing to achieve its most memorable effects. If every joke in the movie were pitched on that level, not only wouldn’t we remember that scene, but we probably wouldn’t be talking about Raiders at all, just as most of us don’t look back fondly on 1941. It’s the intelligence, wit, and technical proficiency of the rest of the movie that allows that one cornball moment to triumphantly emerge.

You often see the same pattern when you look at the movies in which similar moments occur. For instance, there’s a scene in Annie Hall—recently voted the funniest screenplay of all time—in which the audience needs to be told that Alvy and Annie are heading for Los Angeles. To incorporate that information, which had been lost when a previous scene was cut, Woody Allen quickly wrote and shot the bit in which he sneezes into a pile of cocaine. It included all the necessary exposition in the dialogue, but as editor Ralph Rosenblum writes in his memoir When The Shooting Stops:    

Although this scene was written and shot just for this information, audiences were always much more focused on the cocaine, and when Woody sneezes into what we’ve just learned is a two-thousand-dollar cache, blowing white powder all over the living room—an old-fashioned, lowest-common-denominator, slip-on-the-banana-peel joke—the film gets its single largest laugh. (“A complete unplanned accident,” says Woody.) The laughter was so great at each of our test screenings that I kept having to add more and more feet of dead film to keep the laughter from pushing the next scene right off the screen…Even so, the transitional information was lost on many viewers: when they stop laughing and spot Alvy and Annie in a car with Rob, who’s discussing how life has changed for him since he emigrated to Beverly Hills, they are momentarily uncertain about how or why the couple got there.

And while the two moments are very different, it’s revealing that in both cases, an improvised moment of slapstick was introduced to crack an unanticipated narrative problem. It’s no surprise that when writers have to think their way out of dilemma, they often turn to the hoariest, most proven building blocks of story, as if they’d briefly written a scene using the reptile brain—while keeping all the other levels of the brain alive and activated. This is why scenes like this are so delightful: they aren’t gratuitous, but represent an effective way of getting a finely tuned narrative to where it needs to be. And I’d also argue that this runs in both directions, particularly in genre fiction. Those big, obvious moments exist to enable the more refined touches, but also the other way around: a large part of any writer’s diligence and craft is devoted to arranging the smaller pieces so that those huge elements can take shape. As Shane Black pointed out years ago, a lot of movies seem to think that audiences want nothing but those high points, but in practice, it quickly grows exhausting. (Far too many comedies these days seem to consist of nothing but the equivalent of Alvy sneezing into the cocaine, over and over and over again.) And Sorkin’s fondness for the banana-peel gag arises, I suspect, from his realization that when such a moment works, it’s because the less visible aspects of the story around it are working as well.

My novels contain a few of these banana peels, although not as many as I’d like. (One that I still enjoy is the moment in City of Exiles when Wolfe trips over the oversized chess pieces during the chase scene at the London Chess Classic.) And while it’s not quite the same thing, there’s something similar at work in Chapter 43 of Eternal Empire, which features nothing less than a knock-down, drag-out fight between two women, one a runaway bride, the other still wearing her bridesmaid’s dress. If I’ve done my job properly, the scene should work both on its own terms and as an homage to something you’d see on a soapy network or basic cable show like Revenge. And I kind of love the result. I like it, in part, because I know exactly how much thought was required to line up the pieces of the plot to get to this particular payoff: it’s the kind of set piece that you spend ninety percent of the novel trying to reach, only to hope that it all works in the end. The resulting fight lasts for about a page—I’m not trying to write Kill Bill here—but I still think it’s one of the half dozen or so most satisfying moments in the entire trilogy, and it works mostly because it isn’t afraid to go for a big, borderline ridiculous gesture. (If Eternal Empire is my favorite of the three books, and on most days it is, it’s because it contains as many of those scenes as the previous two installments combined, and only because of the groundwork that comes with two volumes’ worth of accumulated backstory.) And although there’s no banana peel, both Wolfe and Asthana are falling now, and they won’t land until the book is over…

“The rest of the wedding was a blur…”

Note: This post is the forty-third installment in my author’s commentary for Eternal Empire, covering Chapter 42. You can read the previous installments here.

I’ve frequently written here about the theory that you can classify any given writer as either a gardener or an architect. George R.R. Martin, who obviously places himself in the former category, returns to that premise repeatedly in discussing his work, and it’s been picked up by other writers of speculative fiction: when I attended the World Science Fiction Convention in Chicago a few years back, it came up at nearly every panel I saw. And like most such categorizations, it’s most illuminating when we look at the places where it falls short. A good gardener, for instance, doesn’t just put words down on paper and hope for the best: to keep the process from spiraling out of control, he or she soon develops a set of tactics for managing the resulting pages. If there’s a hidden architecture here, it’s the vernacular kind, which emerges out of the constraints of the landscape, the materials, and the needs of the people who live there. It doesn’t arise from a blueprint, but it depends nonetheless on experience and good tricks. And the self-described gardeners of literature—the published ones, anyway—tend to be exceptionally capable at controlling structure at the level of the sentence or paragraph. If they weren’t, the story wouldn’t get written at all. (Or if you’d prefer to keep the gardening metaphor alive, it’s also a little like the parable of the sower: ideas that fall on rocky soil or among thorns are unlikely to grow, but they yield a hundredfold when sown on good ground. And even if this isn’t architecture, it’s at least a kind of horticulture.)

In a similar way, one of the most counterintuitive aspects of the architectural approach is that all of its careful planning and analysis really exists to enable a handful of moments in which the plan goes away. Making an outline is less about laying down the proper path for the story—which is likely to change in the rewrite anyway—than about keeping on task and maintaining a necessary discipline over many weeks and months. This routine exists both to generate pages and to make sure you’re physically there when an organic, unplanned insight occurs: it’s a kind of precipitate from the solution that the writer has prepared beforehand in the lab. I’ve spoken elsewhere about the importance of rendering time, in which you need to stick yourself behind a desk for a certain number of hours or days before good ideas can emerge. Outlining and writing a logical first draft happens to be a pretty great use of the time between inspirations, and it can be hard to tell whether an idea emerged from the preparatory stage or if the latter was just an excuse to keep working until the former appeared. But it still works, and the fact that useful insights tend to appear only after a stretch of systematic, sometimes tedious effort is the best argument I know for writing like an architect. The idea that you need to prepare obsessively to allow for the unexpected isn’t exactly new: in fact, it’s so familiar that it has inspired some of creativity’s great clichés, from Louis Pasteur’s “Chance favors the prepared mind” to Branch Rickey’s “Luck is the residue of design.” But like a lot of clichés, they’re true.

For the most part, The Icon Thief and its successors were meticulously planned novels: they all called for a lot of research, and their outlines, in some cases, approached the lengths of the finished chapters themselves. But that planning was meaningful mostly to the extent that it enabled about ten minutes of real insight, spread unevenly over the course of three years of work. I don’t know of any better example than that of Maya Asthana. When I started writing City of Exiles, the second book in the series, I was working toward what I thought would be a neat twist: Alan Powell, the hero of the first installment, would turn out to be the mole in his own agency. I wasn’t exactly sure how this would work, but I trusted that I’d be able to figure it out, and I wrote about half the book with that revelation in mind. When it came time to outline the second half, however, I froze up: I just couldn’t see how to do it. Yet I’d already baked the idea of a mole into the story, and I couldn’t bear the thought of discarding those pages. Out of desperation, I cast around for another character who could assume that role. And to my surprise, I found that the only plausible candidate was Asthana, the smart, slightly conceited, but warmhearted agent I’d introduced into the story solely as a sounding board for Rachel Wolfe, my protagonist. But once I recognized Asthana’s potential, I realized that her origins as a purely functional supporting character were a real asset: the reader would be unlikely to see the twist coming—and I think the surprise works—because I hadn’t seen it, either.

And one of the unanticipated dividends of that decision was the wealth of small, almost arbitrary character details that I’d unwittingly bequeathed to myself. Like Wolfe, who was originally a minor character whom I made into a Mormon just to make her a little more distinctive, Asthana had acquired traits and bits of business nearly at random, and now I had a chance to put them to good use. In City of Exiles, for example, I’d established the fact that she was planning her wedding, mostly because it was a thread I could write without much effort—I’d gotten married just a couple of years earlier—and because it seemed consistent with her personality. Once Asthana became the villain of the series, though, and after it became clear that I wasn’t going to be able to resolve her story in the second book, it seemed obvious that her wedding day was going to be a major set piece in Eternal Empire. Again, I could have simply ignored the clue that had been planted, but it felt right, like using every part of the buffalo, and I had a hunch that it would be a good scene. And it was. In fact, the sequence that reaches its climax here, in Chapter 41, as Wolfe realizes that Asthana is the mole while standing up as a bridesmaid during the wedding ceremony, is maybe my favorite thing in the whole novel. (A big part of the challenge was figuring out how Wolfe could stumble across the truth at the wedding itself. The solution, which involves a surprise poetry reading and a clue from John Donne, manages to be tidy and contrived at the same time.) It’s a scene that never would have occurred to me if the pieces hadn’t fallen into place almost by accident. And while I’d never call myself a gardener, it was nice to see one idea finally bear fruit…