David Mitchell, 2004
the twentieth book in the visitor recommendation series;
suggested by Clare Parker, Noah MacAulay, Jordan Busse,
Duncan Stevens, and Bret Mogilefsky
I didn't know anything about this book when I started in, but I had high expectations nonetheless. The vast, vast majority of books in the visitor recommendation series were nominated by one person; a handful got two votes; one got three; and Cloud Atlas, as you can see, got five. I am pleased to say that it lived up to those expectations. I'm also pretty glad that I hadn't heard anything before I began, because the structure of the book is loaded with surprises. All of which I am about to spoil, so mind the sled.
Actually, one of the things that has sometimes come up when I've been working on a movie is the extent to which it's a waste of effort to try to build up to a surprise: "They're not going to be surprised," my boss would tell me, sadly. "The thing you're dramatically unveiling? They'll already have seen it in the commercials." So here's an account of how my guesses about what to expect from Cloud Atlas evolved — and how my expectations were shaped by factors outside the text. I punched it up on my Kindle and started in, and I could see why people thought it might be my sort of thing: it's 1850, we're in the South Pacific, and a God-fearing American notary is writing in the style of the time about his voyage. I love history, I have a whole blog dedicated to paying virtual visits to different parts of the world, and sure, I was up for some ersatz Melville. Especially once I got a ways into it and found that Mitchell was very good at this — the bluenosed narration was amusing, and the story grabbed me right away. The early action takes place on Chatham Island, which I'd heard of but wasn't very familiar with, so I pulled up its article on Wikipedia — which mentioned that Chatham Island was the setting of "the first chapter of Cloud Atlas". Not a huge spoiler, but it did reveal that we would be elsewhere for chapter two! I figured we would be, but now I had extratextual confirmation that I hadn't necessarily wanted!
Then, the notary's narration stops cold — in mid-sentence, even. I clicked to the next page, and there was the title page for the second chapter. If I'd been reading a physical book, I would have taken it for granted that this was deliberate… but since I was reading the ebook version, I worried that it might be a coding glitch! So, I looked at a copy of the printed book, verified that it wasn't a glitch, and continued on. Now we were in 1931, reading the letters of a disowned dandy with a musical bent who has absconded to Belgium to try to rebuild his finances by proposing himself as a secretary and junior collaborator to an elderly composer. This interested me less than the nautical yarn, but it was well-written enough that I was happy to follow the new story for a while. At one point the dandy refers to having found the journal from 1850, explaining the connection between the two stories, and I figured that we would bounce back and forth between the two and that they would end up commenting on each other in some way. But no! Instead, yet a third story begins. This one is set in 1975, and is one of those "paranoid thrillers" popular at the time (stuff like The Parallax View and Three Days of the Condor). We're in California, and an intrepid reporter is investigating a corporate scheme to open up an unsafe nuclear power plant. Her contact is a long-in-the-tooth scientist who happens to be the fellow who received those letters back in 1931. So, I thought, we were just going to keep getting stories in different genres, right up to the present! As it turned out, not quite. Next up was the story of a shady British publisher on the run; one of the manuscripts he's been reading is that selfsame thriller about the nuclear plant. I figured that maybe we were in the '80s now — but then our new narrator namechecks George W. Bush and al-Qaeda, placing us in what at the time of Cloud Atlas's publication was the present. It seemed like this meant that we were looking at a total of four stories. Then, a couple of weird things happened. First, I hit a hyperlink. It purported to take the reader to a page of a book the narrator had published, but it actually just jumped ahead to a section of Cloud Atlas written entirely in dialect. I didn't read it carefully — I wanted to avoid spoilers — but the link had claimed to lead to a passage from an Irish memoir, and I had no reason not to believe it. I thought that maybe this memoir would be one of the stories that make up Mitchell's novel — that now that we had reached the present, we would soon be traveling backward in time through a different set of stories. I dreaded having to slog through that memoir. I hate heavy use of printed dialect. But even more worrisome was that I didn't know what to make of what had just happened. I checked to see whether the print version instructed the reader to flip ahead in the book at that point. It didn't. The ebook was turning out to be a very different experience!
The second thing that happened was that I happened to mention to someone that I was reading Cloud Atlas, and she asked how it was. I said I was enjoying it, and asked how spoilery I should get in telling her why — I mentioned that it was structurally unusual, and asked whether she already knew about this aspect of the book. No, she said, all she'd heard about it was that it was science fiction. Whoops! It hadn't been up to this point, but as soon as she said that, it finally dawned on me that the present wouldn't turn out to be the point at which the book made its U-turn. I was therefore less surprised than I might have been when the fifth segment of Cloud Atlas turned out to be set in a dystopian future. Nevertheless, this was where my enjoyment of the book jumped to a new level. Mitchell had shown me that he could tell (at least the beginning of) a gripping story in all sorts of different genres, but none of the genres up to this point had really been in my wheelhouse. But the Calendar archives are full of dystopian fiction. And then when I caught up to that hyperlink and discovered that the next section was not an Irish memoir but a postapocalyptic story? Back in 2006 I'd read and watched just about every piece of postapocalyptic fiction I could get my hands on. Unfortunately, I had one more extratextual spoiler waiting for me: at one point, trying to locate my place, I accidentally pressed the "next chapter" button a couple of times instead of the "next page" button, and discovered that chapter seven was in fact the conclusion to chapter five and chapter eight was the conclusion to chapter four. So much for that surprise. As a writer, it is weird to think that the effects you carefully build into your text can be so easily undermined by something this dumb.
I actually wrote something structured like Cloud Atlas once. As a teenager, I'd liked Ender's Game enough that I picked up Maps in a Mirror, a collection of Orson Scott Card's short fiction. It was divided into five sections: horror, science fiction, fantasy, stories about religion, and Mormon children's stories. I decided that I would write a novella in each of these genres, except for the Mormon children's story, which would be short. Then it occurred to me to build loose connections between them so that I could pass the collection off as a "composite novel", which I called Angels and Other Monsters. The connections among the stories of Cloud Atlas are also very loose. And while they're all excellent page-turners — again, Mitchell has a skill that a lot of authors of literary fiction lack, namely the ability to set up a compelling situation quickly that makes the reader want to find out what happens — they're also derivative pastiches. For instance, the dystopian tale, "An Orison of Sonmi-451", owes enough to Nineteen Eighty-Four and Brave New World that Mitchell name-checks the authors (though the reference to an equally obvious influence, Soylent Green, is hidden away in the previous chapter). Standing on its own, I would have found "Orison" perfectly readable but not particularly impressive. Luckily, it doesn't stand on its own, and neither do the other stories.
For the past year I've been (very slowly, obviously!) updating my list of favorite songs, because the copy that's online is nearly five years old at this point. Once my revision was complete, I decided to move on to do the albums as well. Evaluating albums is interesting. There's a point at which quality is additive: an album consisting of fourteen solidly good songs isn't a "solidly good" album, it's a great album! But there's also a point at which quality isn't additive. An album that consisted of a hundred songs that were marginally above average would not merit notice for much other than length. And what about when quality is uneven? How does album consisting entirely of solidly good songs compare to one that consists of three all-time classics and ten filler tracks? I'm still trying to figure out how I feel about this. I do have a sense of how Cloud Atlas charts on my silly asymmetrical 0-to-24 scale as of this writing: I'd rate the individual stories, in chronological order, as something like 10, 9, 9, 8, 12, and 12, and the book as a whole as a 17. And it's not just that a variety-pak is a better value than a single flavor; as loosely connected as these stories are, they do cohere into a greater whole. Call it a concept album.
And the concept that jumped out at me is one that's built into virtually any story or set of stories that unfold over this vast a stretch of time: progress. History is often presented as a process of steady evolution, with each year better than the last, give or take a few setbacks: technology becomes more advanced, attitudes become more enlightened, quality of life improves. Cloud Atlas paints a different picture. The first chapter tempts us to be grateful that we've grown out of Victorian imperialism, but the subsequent ones show how little we've grown — from the traces of World War I that haunt the second story, to the troubling combination of corporate greed and nuclear power in the third, to the slavery, cannibalism, and societal collapse that mark the future. That said, I wonder to what extent this corrective is really needed. Historical progress is a concept that's already scoffed at in many corners — both on the right, which is pretty much defined by the notion that society has fallen from some past golden age into a state of degeneracy, and on the left, whose ideology maintains that changes must be made to our current trajectory. This last claim strikes me as clearly true: the increase in economic inequality needs to be turned around, the spike in CO2 levels needs to be checked posthaste, etc., etc. But at the same time, when people like Martin Amis declare that "the twentieth century is unanimously considered to be our worst century yet", it is preposterous bullshit. Yes, the twentieth century had its share of atrocities, but so did every earlier era, as a glance at India in the 1870s or the Americas in the 16th century or Messenia in the 7th century BCE will show. I understand the impulse to be a contrarian in the face of the triumphalism that underlies the way history tends to be taught, but come on. In what earlier era of recorded history would you rather be born, especially if you had to reroll the demographic dice?
Cloud Atlas
David Mitchell, Andy Wachowski, Lana Wachowski, and Tom Tykwer, 2012
After reading the book, I figured I might as well take a look at the movie. As expected, the filmmakers threw the book into a blender. I once worked on a movie project not entirely unlike this, and my boss laid down the law: all story threads have to be shown up front. Audiences, or at least studio executives, can't handle it when they think they're watching Movie A and all of a sudden they're watching Movie Q. Or so I was told. In the case of Cloud Atlas, this decision kills what was one of the book's main strengths: momentum. You need a chance to settle into a story before "What's going on?" can grow into "What happens next?", and the constant shuttling between storylines makes that difficult — and if you do find yourself wondering what will happen next in a given storyline while watching this movie, your answer will come at some unpredictable point in the future rather than on the next page.
Another effect of this change is that the loose connections between stories suddenly become significantly tighter: we see parallel elements right next to each other rather than separated by hundreds of pages. And "see" is the right verb, since it's a movie. One of the weaker elements of the book was that the chief characters are all given a comet-shaped birthmark, as if that tied them together in a more profound way than, y'know, being the chief characters in the same book. It is the sort of thing that works better in a movie, though — or should. I remember that on one movie I worked on, I had the main character put on a supporting character's jacket, and my boss gave that idea the thumbs-up: "It only takes up one line in the screenplay, but it ends up communicating a vast amount," he explained. "We'll see that jacket in every shot for the rest of this act." Whereas in a book, yeah, to achieve the same effect you have to actually mention the jacket a whole bunch of times. Books convey information in serial, movies in parallel. So I was disappointed when the filmmakers didn't take advantage of this. They should have just applied the birthmark to each of the characters and never focused on it, never mentioned it. Just let us notice it, if we do. If we do, we can pat ourselves on the back for catching a cool thing. If not — it's not actually a great loss!
Probably the worst decision the filmmakers make, though, is to use the same cast in each of the six storylines. The issue is not just that using the same small handful of actors to portray the varied casts of the novel requires a lot of makeup to sorta-kinda disguise age, sex, and race swaps — though, yes, this is a $102 million movie that slaps some Play-Doh over a bunch of white actors' eyes and goes, "Look, Koreans!" A bigger issue is that the stunt calls attention to the fact that the actors are actors, that we're looking at performances rather than peering into alternate worlds where events are unfolding. It's hard to get lost in a story when you're constantly being informed that it's time for another rousing game of Spot-the-Tom-Hanks.
One choice that could hardly have been avoided, though, was to simplify the ideas of the novel. For instance, in the movie, the rebel group Sonmi-451 joins is just a rebel group. In the book, the entirety of the action in the dystopian section turns out to be a puppet show — the rebel group Sonmi-451 joins is operated by the government itself in order to "attract social malcontents" and provide "the enemy required by any hierarchical state for social cohesion" — but Sonmi-451 knows this and plays her role anyway, because she knows that in the long run she will be a force for progress by doing so. That the ideas she puts forth, the ideas that the government wants her to put forth as an articulation of heresy, will outlive both her and the government. This is the sort of complexity that I have tried to put into scripts and have had taken right back out again, on the grounds that it's way too abstract for a visual medium, which it may well be. On the other hand, comics are also a visual medium, and Grant Morrison's New X-Men touched on something similar, with its theme about how Magneto's ideas gained a much wider following once he was no longer around to incarnate them. It's weird to end an article on a tangent, but I've always liked Magneto's speech as he and the sixteen million inhabitants of his nation are killed. So, since it's apposite to Cloud Atlas, here's the end of it:
Once, I was a mortal man. Now I am becoming memory, immortal. They must have thought they could silence us forever. Instead we have become magnetic. Unstoppable. Do you understand? Our voices will be broadcast around the world… into space. At the speed of light. At the speed of radio. Our voices traveling without end through the depths of time and space. Beyond this life. And far, far… beyond this death.