Sponsored By

Finishing Each Other's Sandwiches: Arrested Development Discovers Non-Linear Storytelling

Season 4 of The Greatest Sitcom Ever attempts to incorporate interactive storytelling elements games have been using for years -- and fails. But even its failure has a lot to teach us about the narrative possibilities and pitfalls of our own medium.

Tom Abernathy, Blogger

July 7, 2013

24 Min Read

Well, it finally happened: The television equivalent of the Star Wars prequels, the return of Elvis, and the Second (or Fourth) Coming of Ann Veal’s Lord has come and – well, not gone, because it’s 2013 and we have the Internet and Netflix will make sure it’s always available to stream on demand, forever and ever – but, anyway, it’s here.  And, if you’re like me, you’ve already gulped it down in a couple of days or less.

I’m speaking, of course, of the recently-released Season 4 of the show that gets this writer’s vote for Best Live-Action One-Camera Sitcom of All Time: Mitchell Hurwitz’s Arrested Development.

AD’s original premise concerned the Bluths, a formerly rich and always badly behaved family whose real estate development fortune was built on shady business practices and shoddy workmanship.  In the first episode, paterfamilias George goes to jail for said failings, and relatively normal middle son Michael takes over the family business, to keep the family and company from falling to pieces while Pop-Pop’s in stir.  The absurdist and intoxicating result played like the love child of Monty Python’s Flying Circus and The Simpsons: each episode was packed densely with wall-to-wall jokes that built on and called back to each other, and planted gags unnoticeably and paid them off brilliantly, over and over, until you were giggling like a tickled toddler, begging through delirious tears for a commercial break just so you could catch a breath.  It was that good.

It was also, sadly, a show before its time.  In 2003, there were few DVRs and virtually no VOD streaming.  Most people couldn’t pause, rewind, or rewatch live TV shows in the way they could a VHS tape or DVD.  And no one could do what we all do now in the age of Netflix: binge-watch episodes to our piggy little hearts’ content.  Having those capabilities would have made the Arrested Development experience far richer and more rewarding: planted jokes would still have been fresh in our minds; callbacks to gags from two seasons ago would have rung bells immediately.

But we couldn’t do that then, most of us.  So most of us didn’t watch Arrested Development back then.  So Fox cancelled it.

They Made a Terrible Mistake

Fast-forward to 2012, and suddenly all those capabilities are ubiquitous.  Netflix looks at its copious viewer data and sees people binge-watching all over the place, and likewise sees a lot of people catching up on that Arrested Development thing they missed back in the day, and thus conceives of a great way to seize more ground in their new beachhead strategy of offering original programming.  It’s got a built-in audience, this show, and the very trailblazing traits that made it so inaccessible to viewers back during the Bush (II) Administration are the same traits that shows nowadays try to emulate – traits that make Arrested Development and Netflix a match made in entertainment heaven.

So this past Memorial Day weekend, we all got our IVs put in and mainlined us a massive dose of the drug we’d become addicted to over the past eight years. It was like Requiem for a Dream, only funny.

Now, I know what you’re thinking.  Dude-bro, this is Gamasutra.  What the hell does any of this have to do with games?

The answer is: More than you might think.

Because as the details of the forthcoming Season 4 began to trickle out last year, the most intriguing was this: Mitch Hurwitz had realized he was no longer tied to the longtime broadcast TV model of making discrete episodes with strict time limits, commercial interruptions, and the handed-down-from-the-mountain-on-stone-tablets-to-Pat-Weaver delivery cadence of one episode per week, from September through May, with holidays off and reruns sprinkled throughout the season to give beleaguered production staffs time to catch up while also conveniently maximizing the good stuff for Sweeps in November, February and May.  Mitch Hurwitz had realized he could toss all that stuff out the window.  Mitch Hurwitz, no dummy he, had realized he had the opportunity to make Arrested Development even more of its own thing than it had been on Fox – more purely Aristotelian in its Arrested Development-ness, if you will.

To be fair, Hurwitz was also dealing with the fact that all of his main actors, having become famous over time for their work on AD, were now highly in-demand.  He’d never be able to get them all together for more than a couple of days.  Which meant this couldn’t possibly be a season like the first three, with each episode featuring most or all of the Bluths; the logistics simply wouldn’t allow it.  Intractable circumstances required creative solutions. 

(Is this starting to sound more like game-making yet?)

Drop by drop, the juicy bits of awesome dribbled out: Freed from the shackles of linear viewing, the long-awaited fourth season would be a matrix of story, told in ten – no, twelve – no, fifteen! – interlocking episodes, each focusing on one member of the Bluth family.  Stories would bob and weave; the impact of jokes would change depending on the order in which they were encountered; repeated viewings would reward the gobsmacked (GOB-smacked?) viewer more richly than ever before.

I’ll let Hurwitz explain it in his own words, from an interview conducted by Time Magazine TV critic James Poniewozik:

I kept thinking of it as a crossword puzzle. So you just always had to be thinking, “Okay, wait, six down has to end with ‘ing.’ There’s got to be an N in here somewhere.” … So it was constantly rejiggering that story to make sense… I can try to make that A look like a B or I can find a word that starts with B that sits in here.

So it was, you know, it was – I can’t imagine I will ever have a more challenging ongoing daily experience than I have in this last over a year now, last year… I’m trying to tell the story in the most surprising and effective way so I will quickly have to decide, okay, this episode, this scene which I had originally intended for George-Michael’s episode, really emotionally it’s about the grandfather’s story. So I’ve got to put it in the grandfather’s show which means I’ve got to move that other scene, you know. It’s just a constantly changing Rubik’s cube in a way. All my metaphors have to do with puzzles because that’s all it is, is a puzzle.…

Can you talk about how the circumstances of the production, and making it for an outlet like Netflix, affected the way you could tell the story?

…So my prime motive was how to most successfully and ambitiously exploit this concept of delivery system. The fact that an audience can now be in possession of this as opposed to being fed this on someone else’s timetable – well, intellectually it was very interesting but also just practically as I got into the writing of it I could put things in with the confidence that people could go back very easily and find it. People could pause it. You know, all these things that didn’t exist when I did the first show. When I did the first show there was no guarantee we were going to be on DVD, yet I put a lot of detail in there that again, in retrospect, it was like, “Wow, what was I thinking?”

And I think it might have hurt the show in the short term but it did perhaps help it have legs in the long term. I mean, I’m trying to anticipate a format where you really can jump from moment to moment and show to show, you know. We don’t have that yet on Netflix but that’s something Ted [Sarandos] and I have talked about like how great it would be if, you know, you follow the George-Michael story and when he runs into his Uncle Buster you could press a button and click over to the Uncle Buster story.

When I was on the set… you had mentioned something about having plot points on a bulletin board in the writers’ room— you used different types or colors of string, like sequential string and causal string.

Great memory. It was causal… (T)ruthfully I think you can only keep so much in your head at one point and part of that exercise was to kind of assure myself that at one point in the history of the weeks or months we’ve been doing this, this made sense. And I could kind of do that with a piece of string. So I didn’t have to every time say, “Wait, why did Maeby go to the event if she didn’t know George-Michael was going to be there while she thought her father was at home.” And I can look up and go, “Well, okay, there’s a string going from that backwards to this other thing.”

I do think the biggest difference is I really very quickly stopped making this as shows and started making it as an eight-and-a-half hour Arrested Development. And so my recommendation would be to sort of take it as a whole…


At this point I expect the average Gamasutra reader is beginning to see where I’m going with this.

Getaway, Getaway

One of the things we who specialize in game narrative encounter often in our work life is a lack of understanding of just how difficult our job is.  Everybody talks, right?  Everybody writes emails.  Everybody uses words.  How hard can it be?

The answer is – often times, at least – very, very hard.  Hard in ways that may not be apparent to people who don’t specialize in game narrative – because, until you’ve been tossed across that event horizon, there are no antecedents for the storytelling issues you’re going to encounter… all thanks to the effects of interactivity.

Storytelling is one of the most deeply rooted universal elements of human culture.  We’ve been telling each other stories for millennia – around campfires, in religious ceremonies, as entertainment.  We told stories to teach essential skills; to maintain a cultural memory before writing or printing; to try to make sense of the world and the sky and the thunder and lightning and all the aspects of human life and nature whose meanings and workings seemed so vexingly hard to comprehend.

But through all those centuries of storytelling, the stories had one thing in common: Like time itself – a phenomenon not just subjectively but, as articulated by Isaac Newton, objectively unidirectional – they were all, without exception, linear in form.  This happened after that, and then the next thing happened.  Stories started at one point and then ended at a later point.  Causes led to effects.

And when I say “all,” I’m not being hyperbolic.  You will search in vain the works of Homer or Shakespeare for a single instance of, for example, a flashback.  Stories unfolded in an entirely linear fashion, which makes sense since that’s the way we experience life and, in the days before writing, they were easier to remember that way.  The very idea of moving around in time for narrative purposes did not yet exist.  In Western culture, at least, it was not until we started to create stories intended not to be performed live, but rather to be read, that such narrative-temporal innovations emerged.  Authors, informed by the ideas of philosophers and thinkers who were exploring such questions and writing about them, naturally incorporated those ideas into their writing.

Even so, we’re still talking about simple devices: multiple points-of-view, the occasional “memory” passage – kid stuff.  What really started the clock unwinding/melting was modernism and its influence across disciplines.  Freud and Jung, Maxwell and Curie, Stravinsky and Picasso and Joyce – you can see them all ingesting each other’s ideas and bouncing them off one another in their work.  The two biggest developments, for our purposes, were the invention of moving pictures and Einstein’s theories of relativity.  The reason should be obvious: The first stopped time and broke it up into discrete moments, giving filmmakers power over the audience’s perception of time itself.  The second… well, all those did was utterly destroy everything humans had always assumed about what time was and how it worked.

(When I think about this stuff, I’m always reminded of Marcel Duchamp’s 1912 painting Nude Descending a Staircase, which made no sense to me until I read that it was his interpretation of the way the new technology of cinema was slicing time into slivers like a neurologist slicing up a brain for slide study.  One can draw a fascinating line from there across the firmament of the 20th Century through ever more ingenious convolutions of narrative time, to the rearranged plot of Quentin Tarantino’s Pulp Fiction and the splintered alternate timelines of Tom Tykwer’s Run Lola Run, just to pick out two temporally innovative works that had a big impact on me.)


But if cinema introduced the power to break the linear arrow of time and rearrange its pieces, interactive media has shown the power to shatter it. 

Before we even get to games, let’s think about the choose-your-own-adventure book genre.  These relics from the ‘70s and ‘80s hold a special place in the hearts of game makers of a certain age, because they were our first true encounter with the idea of multiple narrative paths.  Granted, any given story would be read as linear – but if one read the book again and made different choices, the fractured, mix-and-match nature of the form became apparent.

And what was dazzling to a young reader was surely brain-boggling to the writers who crafted these multi-option narratives, because to do so (as anyone who has ever worked on a dialogue tree can attest) was to take on a task unprecedented in human history, one for which our brains’ evolution had never prepared us.  You had to track two – no, four – no, eight, etc. – branches to provide the re-readability that was the point of the novelty.  One doesn’t have to be a mathematician to understand that every time you give the reader a chance to make a choice and determine the next story step, you’re potentially doubling your workload, not to mention creating multiple plotlines that soon become impossible to track if you’re not a computer.

And, really, that was only half the job, because (if you cared about doing it well) every single variation of the story had to be dramatically satisfying.

Think about that for a moment.  Creating one story that’s satisfying is so difficult that the very best, trained, experienced, professional storytellers in any medium can’t do it every time out.  But creating two?  Eight?  Thirty-two? Such a task seems impossible, because it is impossible.

This aspect alone helps to explain why interactive storytelling is so very hard – because it ranges in non-linearity from “slightly” to “bug-freaking insane.”  And the implications of that fact are far more profound even than what we’ve seen so far.

Think for a moment about all the tools in a writer’s toolbox that are dependent on controlling timing and/or sequence – and then contemplate what happens if you take those tools away from the writer:

  • Pacing: the ability to control the speed at which the audience experiences story beats, which gives the storyteller control over the emotional dynamics of the tale.

  • Suspense: the mounting sense of anticipation or dread created by parceling out information at a specified pace, in a specified order.

  • Dramatic irony: the phenomenon produced when the audience learns things at a different time in the story than when the characters themselves learn them.

There are many more, but you get the idea: Many of the storytelling tools writers have at their disposal in every other dramatic medium are simply not options when creating stories for any but the most linear of interactive media.  (And even when the storyline is linear on a macro scale, as in the Uncharted series, there are still often microscopic decision points that vary the experience from player to player.)

This is why so many game developers have had the bright idea of “bringing in Hollywood talent,” only to be disappointed when the talent in question handed in pages and pages of linear screenplay that were unusable outside the context of cinematics.  In crucial ways, game writing is a totally different endeavor from any other kind of writing any storyteller has ever done before the first time s/he attempts it.  If you keep up with industry news you’ve probably noticed more than one “name Hollywood writer” go into a game-writing gig with a lot of public braggadocio about how s/he’s going to revolutionize this benighted ghetto of crappy storytelling, only to resurface a year or so later and acknowledge that writing a game turned out to be a lot harder than s/he had expected.

(And, before anybody starts screaming at me: I started out as a “Hollywood writer” myself, and, believe me, my trial-by-fire on Heavy Gear and Destroy All Humans! was brutal.  I know whereof I speak.) 

There’s a lot more that distinguishes interactive, non-linear storytelling from the kind found in every other medium – stuff like figuring out how to integrate organically and seamlessly with design, art, and audio, so that all departments are actually playing a part in the storytelling and the writing supports and feels inextricable from their work and choices – but let’s table that for later.

It’s time to bring this all back to Arrested Development.

Well, That Was a Freebie

Because when Mitchell Hurwitz talked about “exploit(ing)… this delivery system,” when he talked about moving story pieces around, when he talked about using wall-boards and thread to visualize the various causal-vs.-sequential connections between scenes, and especially when he talked about how he hoped people would watch the new fourth season in whatever order they chose, because different narrative paths would yield different experiences… well, we know what he was talking about:

Non-linear narrative.  Interactive narrative.  The very thing that games have been doing for years now, albeit to varying degrees of success.

Indeed, Hurwitz said he hoped his viewers would watch the new, all-at-once season multiple times in multiple orders, so that punch lines that seemed sans set-up the first time through, for example, would instead pay off gloriously once encountered via a different narrative path.  It sounded brilliant.  It sounded as if, freed from the shackles of conventional American TV production and consumption models, Mitch was serious about exploring the possibilities inherent in this new way of delivering and ingesting filmed story.  Would bingeing make it better?  Could you gobble it up from back to front?  Odd episodes first?  By character?  Would different viewing choices really result in a substantially different experience of the tale?  He didn’t know, but, darn it, he was embracing his production limitations, embracing the structure.  He was gonna find out!  Maybe even discover an entirely new paradigm for TV storytelling!

But that didn’t happen.

The reason it didn’t happen was that, once Hurwitz got into the editing room – the Hollywood maxim, by the way, is that you make your movie or show three times: once when you write it, once when you shoot it, and finally once when you edit it (and that last one may be the most important) – once he got into the editing room, he discovered his plan hadn’t come together in the way he hoped.  The payoff-before-set-up thing wasn’t much of a value-add, because the viewer didn’t know which alternative path to choose.  In a season of fifteen episodes, to ask viewers to watch the whole bunch multiple times, in different orders, just on the off-chance they might find a laugh they couldn’t have understood before, was asking too much.  There were simply too many options.  I was never a math whiz, but I believe you’d call it 15-factorial, or 15! = 1 x 2 x 3 x 4 x…x 14 x 15 = 1,307,674,368,000 possible permutations. (That’s “trillion” with a “t,” for those unwilling to count the decimal places.)

Interactive storytelling is very, very hard.  Even at thirteen orders of magnitude less than that.

Remember what we said about those “choose-your-own-adventure” books?  It’s hard for the human mind to juggle more than two or three discrete plotlines; our brains just aren’t hardwired for it.  Putting it all up on a wall can help, but even that isn’t going to get you much farther, because, as previously noted, you’d have to be able to craft stories that would work approximately equally well no matter which path you took, and that, my friend, is simply not possible.  Getting it right once is hard enough.

(Which is why in big, sandbox-style games like the Grand Theft Auto series or The Saboteur, crafters of game narrative tend to use what I think of as the tent-pole method: you figure out your huge, essential plot elements – those are the tent-poles – and then you give the player some latitude regarding how s/he gets from one to the next.  Put another way: when it comes to narrative, actual player agency is neither as essential nor as possible as the illusion of player agency – not by choice, but by necessity.  And that’s the way it’s going to be until and unless AIs advance to the point where they can crank out a trillion equally awesome plotlines in real time.  Good luck with that.)

There was one more reason Hurwitz gave for his about-face on the whole watch-your-own-AD idea, one with thorny implications for game narrative.  He said in multiple interviews that, once in the editing room, he found that a preferred order of episodes began to emerge.  (I’m willing to bet it was the same order they were written in.)

This isn’t surprising.  For one thing, old habits die hard; that’s the kind of process he’s used to – searching in post-production for the single best linear assembly of the pieces that were shot.  For another – and this is the one we need to think about – he seems to have realized that he simply liked it better if it was told in one specific way.  Changing the order might make it different – but not, to his mind, better.

This is instructional in a couple of ways.  First, the dominant paradigm in game design is to iterate until the gold disc is literally pulled from your bloody, whiskey-stained fingers, and for good reason: iteration is how you grope your way toward choices that will work best for most players.  Trouble is, for both creative and production reasons, you can’t (and shouldn’t) take the same approach when constructing your story.  Narrative is like a house of cards, and, if you keep messing with the structure of a piece, sooner or later the whole thing is going to collapse.  There has to be a point where narrative choices begin to be locked down, if only because there’s voice-over recording and localization and all sorts of other parts of the process that need to happen down the pipeline, and they require a certain amount of time to accomplish.  This creates a tension between gameplay design, which rightly wants to keep changing things to see where the evolution will take it, and narrative design, which wants to stay in step with gameplay design but also has all this other stuff it has to accomplish, and, anyway, isn’t necessarily going to be improved by its previous choices being randomized every six weeks.

(This is one reason that, when I give presentations on fundamental elements of game narrative, “plot” – what most people mean when they think of “story” – is the last and least important element on the list.  A good narrative designer needs to keep the plot loose enough to roll with changes emanating from design iteration for as long as possible.  This approach solves far more problems than it creates; and anyway the only “plot” the player is ever going to remember, for the most part, will be his or her own user story.)

The really thorny question Hurwitz’s change of heart evokes for game narrative, though, is that of authorial intent.  What he realized in the last few weeks before Memorial Day was that, having finally seen the episodes, he felt one version of the season was clearly better than the others – so he publicly withdrew his previous statements and asked everyone to watch it in the posted order.  You could say he chickened out (cue your own personal version of the Chicken Dance), but I think that would be ungenerous.  A strong creative creates and conveys strong vision, and for all but the last 30 or 40 years of recorded history, across every storytelling medium, that vision was understood to have primacy.  In poetry and fiction and live theater that role of vision-carrier is vested in the writer; in film, the director; in modern television, the writer-producer-showrunner.  It’s natural for Hurwitz to have felt the way he did.

Games, however, work differently.  Game development is easily the most collaborative medium I’ve worked in – it makes filmmaking look positively solitary – and requires a coordination of choices among professionals across wide-ranging disciplines unmatched, to my knowledge, anywhere else.  Complicating things is the fact that, for every team led by an unquestioned vision-carrier like Amy Hennig or Will Wright or Ken Levine, there are approximately ten teams with titular leaders who lack either the ability or the interest to properly conceive such a vision and communicate it effectively to the group.  Sometimes this vision deficit is filled ad-hoc by other team members; sometimes the team just lurches forward in barely regulated chaos.  Or doesn’t.  Either way, the lack of a strong central vision for the team to digest and work toward inevitably works to the detriment of the project.

But we work in interactive media, for crying out loud!  We’re supposed to be giving the player as much control as possible; that’s what “interactive” means!

But authorial control is essential for meaningful art!  (Yeah, I used the a-word.)

But there’s no room in interactive media for authorial intent!



I don’t have an easy answer to this dilemma.  (The best advice I’ve got is that bit earlier about the illusion of player agency mattering more than its reality.)  The tension it creates has been present in every game project I’ve ever been a part of.  In practice, the solution often seems to be to just abandon any pretense at trying to integrate all the disciplines involved in making games, which of course is less than optimal.

But I love what some of Hurwitz’s choices suggest for interactive storytelling in our medium.  As my colleague Qais Fulton has pointed out, for example, the way Hurwitz takes scenes we saw from one perspective in a given episode and then replays them later from another, once we have more information, is particularly evocative.  Imagine a game that made a central mechanic out of re-contextualizing given scenes, missions, or chunks of gameplay – inviting the player to experience sequences again and again from a different point of view, as a different character, or with new knowledge that helps us see and feel them in a totally different way.  Some recent games (like Crimson Shroud for the 3DS) havebegun to experiment with this sort of thing.  In the age of Cage and Telltale’s Walking Dead, it seems likely that you could make such a device the core mechanic of a narrative game that would thrill the huge market for story-based experiences like the Fallout, Mass Effect, and Elder Scrolls series.

Mostly, though, I love that Mitchell Hurwitz dove into this murky, uncharted ocean and tried to do just a little of what we game narrative types do every day.  I love that he took the risk (in the words of Stanislavski, “Without risk there can be no discovery”) and I even love that he failed – because, in failing, he confirmed what some of us already knew:

Interactive narrative is very, very hard.

But it’s also the future.

Read more about:

Featured Blogs
Daily news, dev blogs, and stories from Game Developer straight to your inbox

You May Also Like