Thursday (3/1/07) 11:57am - ... wherein Peter writes at some length about "Out of Gas".
I don't expect this to be of general interest (hey, that should be the subtitle of my blog), but I've written a lengthy essay about the Firefly episode "Out of Gas".
What is this essay?
Usually, I do my homework -- and when I don't, it kind of irritates me.
Several months ago, I took a screenwriting class at ACC. It included a lecture about non-chronological narrative, an analysis of Eternal Sunshine of the Spotless Mind
, and an assignment: pick a film, and write a 200-word essay about how that film employs non-chronological storytelling. We were further instructed to show how that film still hews to a traditional screenplay structure.
I never got around to that assignment; not too long afterwards I dropped out of the class. But the essay topic kept rattling around my head, mainly because I could think of several good TV episodes that I might want to analyze in that fashion.
So I'm writing that old assignment about "Out of Gas", episode seven of Firefly
. I'm writing it more for my own edification than anything; it should be instructive to pick apart one of my favorite hours of television and try to tease out how its twisty timeline works.
The year was 2002.
Before we start talking about the script itself (which is here
), let's put things in a little historical context. This episode aired in late October of 2002, only a month after the first episode aired, and the situation was already dire. FOX, which had forced showrunner Joss Whedon and right-hand man Tim Minear to write a second pilot ("The Train Job"), now wanted to see an episode that introduced the central cast -- nine characters -- yet *again*.
But that wasn't the *big* crisis. The *big* crisis was that Gina Torres, the actress playing Zoë, was getting married that week. Somehow they had to put together a show that (1) barely included Zoë; (2) re-introduced as much of the central cast as possible; and (3) didn't suck.
Wait -- it's a clip show?
What Whedon and Minear came up with was something kind of like the lamest, least-creative style of episode there is: a "clip show". You know, that episode where the wacky stop'n'shop employees all get locked in the shop freezer and, facing their imminent death of asphyxiation and/or exposure, reminisce about their favorite zany adventures on Stoppin'n'Shoppin'
. The clip show is the episode that happens when the production budget runs out, and they throw together a 'greatest hits' reel of earlier episodes. Audiences really hate it.
And this was *kind of* what they came up with. An mechanical failure on the ship knocks out all power, including life support. It's a rather explode-y mechanical failure, and it takes Zoë out of commission for most of the episode (aha! logistical problem #1 solved). And it's eventually down to Mal, alone & running low on oxygen (aha! the episode title, it is a pun), trying to fix the ship. And, yes, he relives old memories along the way.
So if it hews to the lamest format possible, how is it one of my favorite hourlong episodes?
Well, this episode varies from that threadbare 'clip show' format in several important ways. First, none of the flashbacks are clips of earlier shows -- instead, they're flashbacks to times that precede the series. Second, it has a real plot -- unlike the average clip show, which just has the characters sitting around in a room (or freezer). Instead, the characters struggle like hell against the very grim odds. And third, they pull off some fairly sophisticated nonchronological storytelling that you'd never see in your average sitcom.
The General Shape of Things
So let's look at how the show is put together, timeline-wise.
We've got three main threads. Thread A takes us from just before the explosion to just before Mal repairs the ship. Thread B takes us from that same moment to the end of the story, when the ship is working and everybody's okay. Finally, thread C is a series of flashbacks to times before the first episode of the show.
The naming convention might seem a bit silly (or geeky), but it's a good way to tell these threads apart. Also, they correspond pretty nicely to the A-, B-, and C-stories in many dramatic episodes: the A-story is the main narrative thrust, the B-story is a subordinate storyline, and the C-story is bits and pieces that add color, but don't necessarily form any continuous story.
If you rearranged everything in chronological order (as several blinkered network executives asked them to do), it would be "all C", "all A", then "all B" -- and the A- and B-threads would stitch together seamlessly.
But it's not in chronological order at all. Joss Whedon broke down the show into a very specific non-chronological set of beats; then Tim Minear went off and wrote 3/4 of the script in a single Sunday afternoon. (Bastard.)
A quick explanation -- commercial breaks split up a TV show into contiguous chunks of (say) 8-12 minutes. Each chunk is called an "act", except for the tiny bit before the credits, which is called a "teaser". Firefly had four acts and a teaser.
So let's start off by taking a look at the teaser.
The most important fact about nonchronological storytelling we can glean from the teaser is this: nonlinear narration frightens and confuses your audience
. Audiences expect a story to take place in chronological order. And a TV audience has more conventional expectations than (say) an avante-garde black-box-theater performance-art audience -- they expect a TV show to be... well, *normal*. So if you're going to go nonlinear, you owe it to your audience to make it *absolutely clear* that your story jumps back and forth in time.
What does that mean?
First off, if you have an episode that is somehow nonlinear, you want to introduce that little quirk as soon as possible. If you go through (say) half the show in chronological order, the audience, trained by a lifetime of experience, will strongly expect the *rest* of the show to proceed in a linear fashion. If you start changing things up late in the game, the audience just assumes the writers started blathering nonsensically. Even if the audience catches what's going on, it will still jar the audience right out of the story.
So they're very careful about setting up the show's first flashback. They put it in the teaser, maybe two minutes in -- aha, we'll know to expect more of this as the show proceeds. And everybody does everything they can (short of flashing a hot-pink chiron that reads "Y HELO THAR AUDIENCE THIS IS A FLASHBACK") to ensure that the audience catches that this is a flashback. They have Mal collapse, semi-conscious, on the floor. A salesman's voice appears as apparently a voice in Mal's head. And (this is the key bit) they go to a completely different "look" for the flashback sequence -- different color balance, different film stock (called "reversal stock"
), different everything.
So by the time we're in the first sequence, we know it's Mal's memory. We know it's a flashback. But even then, Minear keeps hammering it home by showing the ship on the ground on a bright sunny day, with Mal entering from a new location. By then all we have to do is reconstruct how long ago the flashback was.
So that's how they introduce the nonlinearity -- clear enough so that somebody blind or deaf could sort out what's going on.
The other more obvious fact about nonlinear storytelling that we can glean from the teaser is this: any jump in time raises the 'where are we?' question.
For me, all dramatic writing boils down to one task: compel the audience to keep watching. One of the simplest ways to do that is to set up questions for the audience and then answer them. Why do you have conflict in a scene? For many reasons, but paramount among them: it makes the audience ask, "How will this turn out?" They wait around for the answer, and they're satisfied when they get it.
Similarly, any time you jump around in the timeline, the audience has immediate questions: "Where are we?"; "When are we?"; "What's going on?"
A lengthy digression: note that this principle -- "a jump in time creates the 'where are we?' question" -- also applies to *linear* storytelling. Stories that are told in chronological order are rarely told in real time. The story regularly jumps forward -- and these jumps give you a chance to raise the "how did we get here?" question. Even a show like 24
-- a show in real time -- still has jumps; say, when we move away from a storyline and return to it, or start a new act after a commercial break.
Thought-experiment time: say you had a TV show that had no jumps forward at all, and just followed one storyline.
Technically, the very start of the show would act like a jump. You always have that initial time-jump *into the story*, and all those concomitant questions.
So any story -- chronological or not -- gets this benefit for free. Any time we arrive at a new spot in the timeline, we get a new batch of questions we can pay off. *However*, linear storylines almost criminally underutilize this technique. They play things very straight, very clear, and are often so scared of disorienting the audience. You start somewhere clear and obvious, and proceed in a reassuring, straightforward way.
Nonlinear stories have some abilities that linear ones don't. They can *show* you what happened in the past, without having to resort to *telling* you about it. That means they can jump into a very weird scene, and then cleverly make sense of it by showing you how we got there.
Therefore, more nonlinear stories play this game of repeated disorientation and payoff in a spectacular way (see also: Memento
). Most of all, they take advantage of the opening shot: a nonlinear show will begin with the weirdest situation the writers can imagine, almost daring the audience to sort out how we could possibly end up *there*.
Most nonlinear TV episodes just limit themselves to that one trick: start with an incongruous shot, jump backwards some amount of time, and spend the episode showing how we got to that point. (Consider "Trash", which opens with Mal standing naked in the middle of the desert, musing that "That went well.")
"Out of Gas" is no exception. In prose form, "Out of Gas" would not start with the words, "It was another ordinary day on Serenity...." Far from it: everybody but Mal is gone (where?), and the ship is dead (why?), and Mal's trying to *do* something (what?), and (in the act out) we discover that Mal's been shot (by whom?).
What the hell's going on? Well, we've got a nonlinear story, so the show can do a good job of backtracking and *showing* us -- we're going to spend the rest of the show answering the questions 'how did we get here?' and 'how will it turn out?'
The start of act one illustrates a very useful principle for nonlinear storytelling. To explain it properly requires a quick diversion into the fluffy world of improv.There is an improv game called "String of Pearls". It starts with one improvisor walking up to stage far-right and saying a sentence. Then a second improvisor walks up to stage far-left and says a different sentence. Then the remaining improvisors, one at a time, fill up the space in between, each of them contributing a sentence. Finally, we hear all the sentences, in order, from stage left to stage right -- and those sentences strung together tell a complete story.
Now, there's a right way and a wrong way to play String of Pearls, and it's all down to those first two improvisors. Here's an example of the right way: person one says, "In 1845, James K. Polk was inaugurated."; person two says, "... and that's why every one in Iowa Springs had to have an appendectomy." Here's an example of the wrong way: person one says, "I decided to go to the store to buy some ice cream."; person two says, "It was the best dessert ever."
Do you see the difference? In the first case, points 1 and 2 are so far apart, the audience can't imagine how you could possibly string them together. In the second case, you foster no such curiosity.
How does this relate to telling a nonlinear story? Simple: make the audience wonder about the gaps. When we jump from one time to another, create a situation that's as different as possible and the audience will be curious to see how we could possibly get from point a to point b.
"Out of Gas" follows this precept. The first flashback in act one takes us to the exact opposite of the opening shot: everybody's sitting around the table, having dinner, telling stories, and having a grand old time. Everybody's okay. The ship is fine. How the hell did things go so wrong?
In between the laughter and storytelling, Minear slips in some key exposition: Wash tells us he's laid in a weeklong course that will steer clear of absolutely everyone. This brings us to another useful point: when the audience knows things end badly, it's easy to be foreboding.
Remember: with nonlinear storytelling, you create questions about how we got from point a to point b. That means that the audience will be watching your story like a hawk for any clue about how things link up. And that means that if 'point b' looks very dire indeed, any little indication of how you get there will create an 'oh, crap' moment for the audience.
In this case, we know that Serenity's heading out to the middle of nowhere -- and that something bad is going to happen there, beyond all help or contact.
The same goes for the grinding noise from the engine. And the flickering lights. And River's line: "Fire." And then the explosion.
For a while, we proceed linearly, as the crew scrambles to get the ship fixed and to keep Zoë from dying. Then there's a massive confrontation: Mal has to get Wash to leave Zoë's side, go to the bridge, and figure out how bad the damage is.
This brings us to another lesson: motivate your flashbacks. Ideally, a flashback will satisfy two criteria: (1) a character will have cause to remember (or re-tell) what's going on in the flashback, and (2) the audience will want to learn what the flashback has to impart. The teaser flashback (Mal first seeing Serenity) worked via method 1: even if the audience wasn't desperate to know about the time Mal first showed Zoë the ship, it was something that Mal would recall in that moment. This act-one flashback is more like method 2: the show has positioned us so we're thinking intently about the Wash/Zoë relationship, and so when we learn how that relationship first started, it's of some interest to us.
(Also note the sharp differences they create between that scene to the current timeline, with Zoë just not quite trusting Wash -- and with Wash sporting a moustache.)
At the very end of the flashback, they do a cute misdirect by introducing the 'genius engineer' and revealing it to be not Kaylee but Bester. Remember that viewers create very strong expectations as they connect the dots; you can thwart those expectations and surprise your audience. Once you've established that the story connects point a to point b, the audience will always look for the simplest storyline that connects the two points. Take these as opportunities to show that things are a little bit more complicated. Here, by introducing one little twist, we raise a question that we can pay off later. (See? It's "make the audience wonder about the gaps" all over again.)
Then we're back from the Wash flashback and into the a-storyline. Zoë needs (and gets) a shot of adrenaline. Then we're back into the b-story, with Mal self-administering a shot of the same thing.
This setup gets to the heart of why they did this a- and b- story to begin with. Minear is trying to tell this story of Mal, shot in the gut and barely holding on, trying to repair the ship before he dies. Looked at another way: Mal is wandering around his ship all alone. There is nobody to talk to. There is nobody to listen to. There is precious little Mal can do besides staggering around and trying to fit a metal thingy into the ship's engine.
What am I getting at here? If you told this story in a linear fashion at the top of the episode exposition would be damn near impossible. How the hell are you going to explain, say, that Mal is injecting himself with adrenaline so that he doesn't pass out? Sure, you could give the medicine bottle a big, clear label -- but that's kind of clunky in a telling-not-showing sort of way (and how does Mal know where to find that bottle, anyway?). And here's the thing: even if you solve the 'medicine bottle' problem, you've got a bazillion other bits of exposition you'll need to get across, and no way to do it.
"No problem," you think to yourself, "I'll just put the Mal-staggering-around bit at the end." Oh, really? So you'll spend all of act four of your show, the one that's supposed to be full of high-stakes, accelerated action, with Mal staggering around the ship very slowly? Um... no.
And *this* a good reason to split the current-day story into these a- and b-stories. By arranging the a-story parts correctly, they can plant every piece of information we need to understand the b-story. And they can tell this story without making the acts' pacing go wonky.
Are there lessons to take away from this? Well, it does hint at a basic principle: in a nonlinear story, all of the principles of linear storytelling still apply. If you want the audience to understand stuff later in the show, you'll have to put the appropriate exposition in earlier in the show. You want the show to accelerate. You want the stakes to get higher and higher. And you want this to be true even if the timeline is Cuisinarted to hell.
And also: nonlinear storytelling allows you to *dramatize* important background information. If the audience can't understand this scene without knowing what happened yesterday, you don't have to have some Basil Exposition *tell* us what happened yesterday; you can just do a time jump and *show us that scene*.
Then we're back in the a-story -- one of the most expository scenes in the show, as Kaylee explains exactly what's gone wrong with the ship. (But notice how Minear makes it a scene with conflict; Mal has to drag the information out of her.) We end the act with the reveal that the ship has only two hours of oxygen left. It elegantly raises the stakes on *both* stories at once -- in both the a- and b-stories, people are running out of oxygen.
Again: even though it's a nonlinear story, the *same rules apply*. The act out should raise the stakes and take the story in a new direction, even though you're telling things out of order.
Act TwoSo, we just heard from Kaylee about the engine. And we start act two with a good look at that catalyzer Kaylee was talking about. By this point we the audience are ready to see a flashback of Kaylee.
That is indeed where we go next: the c-story, the distant past, with Kaylee cheerfully fixing the engine in seconds.
And this sets us up beautifully to return to the b-story with Kaylee forlorn and unable to fix the engine. Sure, those two events are months (years?) apart, but they occur one after the other during the show. And even though it's nonlinear, the usual rule applies -- wherever possible, you want contrast between one scene and the next.
Again, the b-story serves as exposition: it's explaining to us exactly what's broken, and exactly what replacement part they need, and exactly where it goes. This spells out exactly what Mal needs to do in the b-story, but it does it without *feeling* like exposition -- instead it feels like Mal, a bit desperate, forcing Kaylee to keep trying.
After checking in with Simon & Inara and River & Book all facing their imminent demise, we have a scene with Mal and Wash over boosting their distress signal. (Whedon points out in the commentary that they tried to show Mal dealing with each of his crew differently; whereas Kaylee gets reassurance, Wash gets shouted at.)
Essentially this whole act has been about how they're running out of air, and how the crew faces that, and how Mal tries desperately to keep that from happening. The whole act has hammered these points home, which sets us up nicely for the b-story act-out: as the oxygen-level klaxons sound, Mal *fails*. He fumbles with a new catalyzer, and he drops it somewhere in the engine.
Note that now we might wonder where this new catalyzer came from. Again, every difference between time a and time b creates a question you can pay off. It's all about creating the gaps and filling them in.
A side note: you may notice that act two felt simpler, chronology-wise, than act one. Act one was: b-story, a-story, flashback, a-story, b-story, a-story. Act two was: b-story, flashback, a-story, b-story.
Remember what I said about "even though it's a nonlinear story, the same rules of storytelling apply"? Well, one of those rules is that 'late acts are simpler than early acts.' Why is that? Well, you want an overall sense of acceleration throughout your story in order to maintain an audience's interest. And in order for an act to feel like it's going by quickly, you want it to be straightforward: one event leads to another, then to another.
What does this mean? Later acts are often more linear than earlier acts. Those time jumps, the ones that disorient the viewers so pleasantly in act one, make things feel sluggish in (say) act four.
Act ThreeHonestly, I have little to say about act three. The a-story keeps paying off the questions raised in the b-story. (Where did the crew go? Why, they took shuttles in opposite directions. And so on.) The act out puts a nice twist in both the a-story (everybody leaves but Mal) and the b-story (a salvage ship shows up), just like it would in a linear story.
Again, they *set up* the Inara flashback nicely -- by this point, we've seen three flashbacks, and when we see Inara and Mal have possibly their last fight we're genuinely expecting a flashback to show us their first one.
One quick lesson from Inara's flashback: never pass up a chance for irony. In nonlinear storytelling, you frequently get situations where the audience knows what happens next but the characters do not. You can use this "audience superiority": show characters believing things will turn out one way when the audience knows full well it turns out another. So when Inara tells Mal, "That's the last time you get to call me a 'whore,'" well... it's funny. We know things don't turn out that way.
Act FourSomething I could have pointed out earlier, but is probably clearest in act four: a flashback has to work as a scene. Just because your flashback is providing background information doesn't mean it gets to be non-dramatic. The usual rules of drama apply: you have to concisely set up a conflict, raise the tension, resolve the conflict, and quickly end the scene.
I suspect that it was especially important for this scene because the flashback is so weakly motivated: we haven't seen Jayne in this act yet, and the only reason the flashback even shows up is (presumably) because the ambush from the S. S. Walden vaguely reminds Mal of the time he met (and hired) Jayne. But they set clear stakes (Mal and Zoë are about to get shot), a clear objective (Mal convinces Jayne to work for *him*), and run the scene until Jayne changes his mind. Done and done.
The last thing we learn here is simple: fill in all the important blanks by the end. They run the a-story right up to the exact point where the b-story started off, and then they run the b-story to the end. It stitches together perfectly.
I can't think of anything useful to say about the closing flashback. It's another case of cute misdirection: at first, we imagine Mal must be looking at Serenity, but no. I suppose the real lesson here is this: chronologically, end the story wherever you please. This is where nonlinearity gives you a great liberty to end on the perfect *thematic* moment, instead of just "what happens last". And as we've been watching a show about the ship dying, and about flashbacks to how the crew came together, why not end it with the first time Mal laid eyes on the ship?
Conclusory BitsNow we get to look back on all of these (possibly-contrived) 'lessons' and try to formulate more general advice about telling nonlinear stories on TV.
First off, even when you jumble your storyline into a totally non-chronological presentation, 'audience time' remains the same. And that means that most of the rules of drama remain the same. You still want the audience to experience the same pacing: a sense of acceleration, an awareness of rising stakes, and acts that end on cliffhangers.
The big *difference* is that nonlinearity lets you show, not tell, things in the past. This is a boon to exposition. It lets you set up audience questions about 'what happened before?' that you can answer dramatically. And it lets you set up -- and play with -- audience expectations about how various chunks of the story connect.
Still, I find myself stuck with the basic question: why? Why this structure? Sure, if you took the whole story, stripped out the flashbacks, and put it in chronological order, it wouldn't have worked dramatically. It would start with a whimper (the birthday party) and end with lots of really slow walking (the b-story).
But there were a bunch of ways Whedon & Minear could have solved those story problems. Why did they solve it with some of the craziest nonlinear storytelling I've seen in an hourlong drama?
I don't know if I'll ever answer that one. Apparently Whedon, over a steak dinner, suggested opening the episode with Mal shot in the gut -- and they just ran with it.
Creativity is mysterious, I hear.
All I can guess is that showing us the b-story -- with Mal at the absolute end of his rope -- right off the bat creates this crushing dread that permeates the rest of the episode. In the a-story, no matter how bad it gets, you *know* it'll get worse. Also, the c-story -- the flashbacks -- hammer home just how much history these characters have together and makes this sense of impending doom feel even worse. And in the b-story, which runs across the whole episode, you just have no idea how Mal will even stay standing.
So finally, after all these chronological manipulations, you get something that's better than average for a clip show.
p. s. For a few words about the writing of "Out of Gas" from Whedon and Minear, click here.
p. p. s. Note that I could be wrong about all of this. Contradiction is welcome.
 Of course, showrunners really hate it, too. Bryan Fuller to this day kvetches about how Dead Like Me was so starved for production money that it had *two* clip shows in its *first season*. (And really, can you blame him?) Also, the animated Clerks series does an excellent lampoon of the format -- its second episode is a clip show.
 Interesting note: most new hourlongs these days have *five* acts and a teaser. In fact, there's so many acts that the teaser is the same length as an act, making it effectively a six-act structure. Writers are still trying to sort out just how to deal with this.
 And remember that this was late 2002 -- two years before LOST premiered, and four years before shows like Day Break and The Nine made their short-lived forays into complicated nonlinearity -- so the audience had not been primed for these sorts of shenanigans.
 And that's not just a hyperbolic turn of phrase there -- in an audiovisual medium, if you have to impart something of vital importance to the audience, it really helps to hit both the 'audio' and 'visual' channels.
 Technically, we could say "all narration is non-chronological". You know, if we wanted to be wankers about it.
 Frequently, a mystery is all about reconstructing something that happened before the story even starts. Perhaps all mysteries are nonlinear? Nah, that sounds wanker-y, too.
 In fact, it would then be incumbent on person three to hop into the middle of the 'story' with something like, "... and *that's* why we were spotted by the ninja monkey assassins!"
 Readers will note that this is only subtly different from the previous point. Both of them are about the questions we ask when we jump from point a to point b: the first question is "where the hell *is* b?"; the second is "how did we get there from a?"
 Good to know they like decent science fiction.
 It's fairly obvious what "take the story in a new direction" means for a linear narrative -- have some twist in the plot so that the stuff that happens in act (n+1) will be clearly different from the stuff that happens in act (n) -- but what does it mean for an nonlinear narrative?
In that case, it just means that our *perception* of the story changes. We thought it was about 'x' ("oh no, our ship isn't moving"), but it's really about 'y' ("oh no, our ship is running out of oxygen"). So the subsequent act, even if it shows stuff *before* what you just saw, is still viewed in a new context.
Well, that was edifying. What next? Hmm, maybe an essay about the opening action sequence of LOST....
contemplative · Music:
Sun Kil Moon - Neverending Math Equation