Some good questions came up in the comments section of my lengthy Inception critique (“Seventeen Ways of Criticizing Inception“), and I thought it made the most sense to respond to them with a new post. So let’s wade back into Limbo, shall we…
One thing I criticized Nolan for was the number of cuts he made, or the number of shots he used: “Herr Wunderkind Nolan can accomplish in thirteen shots what it takes most directors six to do!” I then proceeded to read the number of shots (13) in Inception‘s opening scene:
1. slow motion pan of waves crashing against a rock
2. slow motion shot of waves
3. close-up of Cobb’s face as he lies on the shore
4. point-of-view shot of a boy on the beach making a sandcastle
5. reverse shot of Cobb’s face
6. pov shot of the boy and a girl making a sandcastle
7. reverse shot of Cobb’s face
8. pov shot of the kids running out of frame
9. shot of a gun being pointed at Cobb’s back
10. reverse shot of a guard standing over him, holding the gun
11. reverse shot of the gun pulling Cobb’s shirt up, revealing his own gun
12. reverse shot of the guard calling to another guard
13. long shot of that other guard
My criticism here is one of economy. Thirteen shots is a lot of busyness in a scene that doesn’t really need it. All Nolan needs to establish at the start of the film is that:
- Cobb has washed up on a beach.
- He thinks he sees his kids.
- An armed guard discovers him.
- The guard is going to take Cobb to a Japanese mansion also on the beach.
If I’d been directing this film (amuse me), and had wanted to keep the same basic idea that Nolan used (fairly doubtful), then here’s how I would have tried it (i.e., this would have been my storyboard):
- Establishing shot of the beach.
- Closer shot of Cobb washed up on the beach.
- Close-up of Cobb. (Click here for an explanation of close-ups, etc.—see the section “Shot Scale”.)
- Reverse shot of his kids (to establish they exist in his POV).
- Reverse close-up of Cobb.
- Close-up of his back. A gun enters the frame to poke him.
- Reverse shot of a guard holding the gun. He looks up and calls to another guard as the shot pans to a mansion on a nearby cliff.
…So, I would have used nearly half the number of shots that Nolan used. I would also have made each shot longer.
But, that all said, I wouldn’t have done it that way.
For one thing, I would have begun with a very tight close-up of Cobb, both for its unsettling effect (audiences don’t expect films to begin with close-ups), and to establish the possible reading that the entire film is going to take place in Cobb’s mind. (This seems to me so obvious that I’m surprised Nolan didn’t do it.) (No, actually I’m not surprised.)
Such an opening would be effective, I think, because it’s uncommon, but note that it isn’t anything particularly innovative. I’m cribbing here from two of my favorite Hollywood films of the past 25 years:
The Portrait of a Lady (1996), directed by Jane Campion:
(See in particular the very end of that clip, around 2:20.) (An aside: I really love how Campion begins this film with portraits of modern-day women. Jane Campion—always thinking.)
Blue Velvet (1986), directed by David Lynch:
Blue Velvet‘s main plot begins with Jeffrey Beaumont (Kyle MacLachlan) discovering a severed human ear (see around 6 minutes in). Much later, toward the end of the film (at the beginning of its coda), Lynch tracks out from an extreme close-up of Jeffrey’s ear, which I read as the suggestion that the entire film takes place inside Jeffrey’s mind (or at least his ear). (Lynch suggests something similar in Eraserhead (1976).)
Talent borrows, genius steals:
- Extreme slow-motion close-up of Cobb’s face. Begin tracking slowly backwards, enough to include spray from the water; maybe have a wave crash around Cobb’s head. Use the soundtrack to establish that he has washed up on a beach.
- Reverse shot of kids (to establish they exist in his POV).
- Reverse extreme close-up of Cobb, continuing that backwards tracking shot (enough so we see some of his body).
- Close-up shot of his back. A gun enters the frame to poke him.
From there, I would probably just cut to the shot of Cobb being dragged into the Japanese mansion; I think the audience can be trusted to understand from the shot including the gun that he was discovered and taken to a nearby building. (I also like the matching of cutting from a shot of his back to another shot of his back.) Not to mention, this more fragmented approach keeps things a bit more disorienting, which I think appropriate for a film like this.
So, four shots, rather than thirteen. And, again, that’s assuming I would have even kept this opening scene. Suffice to say, had this film been mine, I would have changed a lot! But I think it’s pretty easy to demonstrate that, if one does want this opening scene here, there are much more economical ways to execute it.
Let’s look at another scene:
[Note: the original clip I used got taken down, so the times below are wrong, but the shot order is still correct. The section in question begins around 1:10.]
- 0:01 Ariadne medium close-up
- 0:05 Cobb medium close-up
- 0:08 Ariadne medium close-up (somewhat closer)
- 0:13 Cobb medium close-up (somewhat closer)
- 0:20 Ariadne medium close-up; begin slowly tracking toward
- 0:25 Cobb medium close-up (closer) (note the bad match on his leaning backward in the previous shot)
- 0:27 Ariadne close-up; continue to push in
- 0:31 Cobb close-up
- 0:33 Ariadne close-up
- 0:36 Cobb close-up
- 0:39 Ariadne close-up
- 0:41 master (cafe); backward tracking
- 0:43 Ariadne close-up
- 0:47 Cobb close-up
- 0:49 Ariadne close-up
- 0:53 insert shot (coffee cup) (This is such a cliché! Really, a rattling coffee cup, couldn’t Nolan have thought of anything better? I guess he’s seen L.A. Story (1991) (which is a nice little film)—or maybe he hasn’t, since L.A. Story mocks this cliché.)
- 0:55 Ariadne close-up
- 0:56 explosion 1
- 0:58 explosion 2
- 1:00 Ariadne close-up
- 1:02 explosion 3
- 1:05 explosion 4
- 1:07 Ariadne close-up
- 1:08 explosion 5
- 1:13 Ariadne close-up
- 1:14 explosion 6
- 1:16 Ariadne close-up
- 1:17 explosion 7
- 1:19 explosion 8
- 1:22 Ariadne close-up
Does Nolan really need thirty shots here? Sure, the last thirty seconds or so feature some whiz-bang explosions (which is what I imagine most people remember from this scene)—and when most people defend Inception, I imagine they’re defending stuff like these explosions. And, to be fair, the explosions are pretty cool! All thirty seconds of them! I, for one, would have liked to have seen even more of them! (Why is this effect never again used in the film?)
I’ve mentioned this before, but anyone who likes these explosions, do do do check out Michelangelo Antonioni’s tragically underrated Zabriskie Point (1970); he gives you over five minutes of explosions! And they’re much better because they’re real! (Plus: Pink Floyd!)
(See 3:40–8:55 for the good stuff.) …Not to mention, the rest of ZP gives you an orgy, spectacular landscape photography, plenty of action, bizarre dialogue by Sam Shepard, drugs, much more Pink Floyd, and tons of great late-60s fashion. The reason why so many people like 1960s European art cinema is not because it’s foreign or snooty, but because it’s extremely fun and awesome!
Back to Nolan: a nifty half-minute of explosions, I’ll grant him that. But the minute preceding them is nothing but boring, boring shot-reverse-shot, during which Cobb spouts yet more exposition at Ariadne.
And this is essentially my criticism of Inception: that the vast majority of the film consists of inelegant crap like this (well, this, and incoherent shots of people firing automatic weapons at one another). Perhaps we can agree that a film is ultimately about whatever most of its running time depicts? If a scene features thirty seconds of explosions, and sixty seconds of dialogue, then the scene was more about the dialogue than the explosions, right? Well, when you watch Inception, you spend a great deal of time watching close-ups of a constipated-looking Leonardo DiCaprio expositing—not my idea of fun on a Friday night.
How to do this scene differently? For one thing, I would have recorded most or all of Cobb and Ariadne’s conversation in a single medium-long shot that features both actors. (They’re sitting at a café, making it easy to seat them at 45 degrees to one another, facing slightly outward, toward the street.) The longer shot (in both distance and duration) would have allowed Ellen Page and Leonardo DiCaprio the opportunity to really react to one another, as well as to act with their bodies. (In other words, I’d have shot their conversation in a single master shot, with perhaps a couple of insert close-ups to punctuate certain lines or reactions—but then returning to the master.) I’d also have cut as much exposition as possible (or at least revised the dialogue to make it less painful. I’d have given Cobb some personality, for one thing—maybe he speaks more enigmatically? poetically? And I’d have cast someone better to play him—imagine, say, George Clooney in the role, instead—someone genuinely charismatic, whom you trust at first, but gradually realize is deranged. Or Tadanobu Asano! Or Tilda Swinton!).
The explosions at the end can stay mostly the same, I guess—although I would have remembered them later on in the film, and used that effect again! But I would have made one major change: I wouldn’t have cut away for the first explosion. Rather, I’d have the rumbling start during the medium-long master shot of Ariadne and Cobb conversing, then have the explosion happen during the middle of that shot—it would have been so much better! Then, I would have cut away for all the other explosions, before returning to the master shot, where the first slow-motion explosion would still be happening. (I would, of course, dispense with all of Ellen Page’s reaction shots at the end; I think they really gum things down.)
In both film and writing, it’s usually best to say things succinctly. This was standard practice in the classical Hollywood cinema (1930s–1960s), which tended to use editing much more sparingly—but also (and as a direct result) more effectively. (There are of course lots of great reasons why a director might want to cut a lot; this is a rule of thumb.)
For an example, let’s look at the opening of Billy Wilder’s classic Ace in the Hole (1951):
Billy Wilder never did anything wrong. There’s plenty of cutting here (23 shots), including some uses of shot-reverse-shot, and other close-ups. But nothing in the scene is formulaic or static; quite the opposite. (It’s conventional, loosely obeying the dominant style of the time, but it isn’t predictable.) There are lots of tracking shots and pans, accompanied by very strong editing, with plenty of matches on action. Overall, this clip distinguishes itself from Inception by its variety of shots: in addition to the close-up work, there are several wider shots that allow Kirk Douglas to really act (not only with his voice and his face, but his body—and he really takes advantage of it!) The variety of shots—which also run different amounts of time—helps provide tension, as we don’t always know what the next shot is going to be, or when it’s going to come. This matches Douglas’s character’s restlessness very well. It also forces the audience to pay more attention, rather than settling into a slack-mouthed passivity as the film follows some basic formula (close-up, close-up, close-up…, cutting every three seconds). (We also get a very good look at the the newspaper office as a location—and note the ironic interaction between Douglas and the embroidered sign on the wall.)
23 shots in 6 minutes makes for an average shot length of just under 16 seconds. These longer takes allow for a lot of different kinds of shots and effects. But since the 1980s, Hollywood films have been “speeding up,” and ASLs have been getting shorter and shorter. There’s a lot of debate as to why this is so, but one commonly voiced explanation is that since TV tends to be faster-paced than film, filmmakers are trying to make movies look more like TV (always a noble goal). Another common explanation is that people have shorter attention spans these days.
Regardless of the motivating reasons (I’m sure there are more than one), it’s a real phenomenon. Film critics David Bordwell and Kristin Thompson have written a lot about it; for example:
The Departed has calmed Scorsese’s urge to track a bit, but that’s balanced by its over 3200 cuts. The result is an average shot length (ASL) of about 2.7 seconds. Not unusual for an action picture nowadays, but consider where Scorsese started by conning these ASLs:
- Mean Streets 7.7 seconds
- Alice Doesn’t Live Here Anymore 8.0 seconds
- Taxi Driver 7.3 seconds
- King of Comedy 7.7 seconds
- Gangs of New York 6.7 seconds
- The Aviator 3.6 seconds
Like his contemporaries, Scorsese has succumbed to the fast-cut, hyper-close style that has made our movies so pictorially routine, however well-suited they may be for display on TV monitors and computer screens and iPods. In 1990 he seems to have realized that he needed to pick up the pace. Of GoodFellas (ASL 6.7 seconds) he remarked: “I guess the main thing that’s happened in the past ten years is that the scenes [shots] have to be quicker and shorter. [GoodFellas] is sort of my version of MTV. . . but even that’s old-fashioned” (The Way Hollywood Tells It, p. 152).
I’ve been very disappointed in Scorsese’s 2000s films for many reasons, but one of them is because he’s succumbed to a fashion in Hollywood movies, and not for the better. (I remember a scene early on in Shutter Island where Scorsese and editor Thelma Schoonmaker use like nineteen close-ups to convey the very simple action of Leo’s character unwillingly handing over his gun to a guard. I would have done it as a single, wider shot.) (Actually, I wouldn’t have even put that scene in the film.)
So when I watch Inception, I don’t see Christopher Nolan doing anything innovative or original. Rather, I see him uncritically obeying the dominant directing/editing fashions of our times. And his doing so flies directly in the face of what I tend to consider art: great directors challenge conventions, and try to find new ways, or remarkable ways, of making things.
This view of art is deeply rooted in the writings of critic Viktor Shklovsky—in particular his 1925 collection Theory of Prose, which I am forever quoting from, and which I’ve written about here, here, here, and here. Shklovsky’s basic argument is that everyday life inures us to experiencing life:
If we examine the general laws of perception, we see that as it becomes habitual, it also becomes automatic. So eventually all of our skills and experiences function unconsciously—automatically. If someone were to compare the sensation of holding a pen in his hand or speaking a foreign tongue for the very first time with the sensation of performing this same operation for the ten thousandth time, then he would no doubt agree with us. [...]
And so, held accountable for nothing, life fades into nothingness. Automatization eats away at things, at clothes, at furniture, at our wives, at our fear of war.
If the complex life of many people takes place entirely on the level of the unconscious, then it’s as if this life had never been. (4–5)
He proceeds to argue that the role of art is to refresh us, to teach us how to see the world again:
And so, in order to return sensation to our limbs, in order to make us feel objects, to make a stone feel stony, man had been given the tool of art. The purpose of art, then, is to lead us to a knowledge of a thing through the organ of sight instead of recognition. By “enstranging” objects and complicating form, the device of art makes perception long and “laborious.” The perceptual process in art has a purpose all its own and ought to be extended to the fullest. Art is a means of experiencing the process of creativity. (6)
But art can’t do that if it’s as habitual and automatic as everything else that surrounds us.
1 + 1
Someone who really understands Shklovsky (although I don’t know whether he’s actually read him) is the French filmmaker and theorist Jean-Luc Godard.
A central principle behind many of Godard’s films is the idea that a film’s image track needn’t match its soundtrack in a one-to-one fashion. Thus, if a film features a character’s voice on the soundtrack, you don’t necessarily need to show her speaking (or show what she’s speaking about). Godard, acting in a rather playful French manner, has called this idea “1 + 1 = 1,” meaning roughly that 1 image + 1 (different) sound = 1 (total) cinema. (This dialectical idea is directly comparable to the Situationist International technique of détournement, and other radical collage-based techniques.)
Godard has explored this 1 + 1 idea extensively throughout his career; for a clear starting point, check out his 1968 film One Plus One, better-known as Sympathy for the Devil:
Here we see the idea in a still-nascent form: Godard films long takes of the Stones recording their (then new) song, intercutting more didactic, fictionalized skits featuring Black Power militants, celebrities, street activism, and Maoists. The scenes are obviously intended to comment upon one another in a dialectic fashion. (You watch the British Stones appropriate African-American blues music and sing about revolution, and then you watch Black Power militants read passages about revolution.)
In later films, starting in the 1970s (when he began working more on video), Godard would begin pushing his sound and image tracks further and further apart, eventually creating more immediately dialectical effects:
Admittedly, this is a pretty far-out use of cinema, but Godard has always been out there, and his work is at all times rather innovative and instructive. (Watching it really opened me up to considering more of what cinema was capable of doing; I also find it pretty beautiful in its own right.)
So, returning to Nolan’s use of sound + image: I consider him pretty over-simplistic. I’m not arguing that he has to do anything as extreme as Godard, mind you, but there’s plenty of room in between. When Cobb mentions his kids, Nolan doesn’t have to cut to a shot of them—he could just show Cobb talking. Maybe Leo could act so we’d see his yearning on his face, and the absence of the children would be poignant? (And what if we never once saw the kids, but only heard Cobb describe them? It might make him less reliable as a character; after a while, we might wonder if he even has kids….) As always, there are a million things that an artist can do—but not when he’s wedded to following rote formulas.
Incidentally—and I can write more about this later if people are interested—I’d lay much of the blame for contemporary Hollywood’s inelegance and lack of artistry at the feet of Peter Jackson. His Lord of the Rings trilogy (2001–3) is by far my least favorite contemporary Hollywood production, and might even be my least favorite set of movies of all time—because it, more than anything, promoted this formulaic method for making movies. Consider any random clip from the LotR films; they consist primarily of four things:
- Close-ups of characters looking very concerned while expositing to one another;
- Insert close-up shots of whatever it is they’re talking about (Gollum, in the clip linked to above) (usually these are push-ins on that object);
- Swooping shots through computer-generated landscapes;
- Bombastic music laid over all of it.
Inception is Jackson’s heir. (It swaps in gun play for CGI fly-throughs, but otherwise it draws from the same small toolbox.)
Why did Jackson make his films this way? Why did Nolan follow Jackson’s lead? I’m sure each director has his own reasons (in Jackson’s case, it was probably the most efficient way to shoot so much footage with so little money), but part of the reason is that they are following gladly in the modern Hollywood tradition, which openly thinks that the audience is stupid. The supreme influence here is George Lucas, who said it best:
“Popcorn pictures have always ruled. Why do people go see them? Why is the public so stupid? That’s not my fault.”
[Lawrence] Kasdan’s main criticism was that Lucas glossed over the emotional content of a scene in his hurry to get to the next one. Gary Kurtz agreed and left the story conferences feeling troubled about Empire. He thought the ending failed to resolve the characters’ conflicts and that Lucas was relying too much on the revelations contained in part three of Star Wars. Lucas’s response to these criticisms was typical: “Well, if we have enough action, nobody will notice.” If Lucas has a weakness as a filmmaker, it comes in sacrificing thematic and character development for action. Kurtz says, “He’s afraid of going too slow.” (211–2)
Directors like Lucas and Jackson and Nolan, like so many of today’s Hollywood directors, believe that if they don’t cut every three seconds, the audience will lose interest. Their view of cinema, unlike Godard’s or Wilder’s, is one of distraction (or, as the Situationists called it, spectacle). And once you make the decision to cut every three seconds, you sacrifice a great deal of cinematic storytelling technique.
Not that there still isn’t a lot that can’t be done, to be sure. Edgar Wright’s films, as I described in this post contrasting Inception and Scott Pilgrim vs. the World, feature very fast-paced cutting, but not at the expense of the direction. That is to say, Wright still actively shapes and directs his movies, creating surprising and unpredictable effects, rather than just applying a standard “fast-paced” formula (master shot, dialogue sequence done in shot-reverse shot, some nifty CGI) to every scene.
And it’s formula—habit—that has always been the enemy.