Branching Infinity: Exploring the Many Structures of Interactive Fiction

The difference between interactive and traditional fiction isn’t a clear binary at all: Both require an approach to structure that balances openness and control, and each contains lessons for the other.

Like many people who’ve struggled to see themselves in traditional fiction, I’ve long been in search of stories that want me in them. In the past, this manifested as a desire for representation. I wanted to get outside of the western canon and read fiction about outsider experiences, particularly exploring the inner lives of queer nerds like me. What that led to, naturally, was work by folks who shared my interests—games, mathematics, and the peculiar sense of unrequited longing that characterizes digital life. Drawn to this work at the literary margins, I was delighted to discover a world of fiction where I wasn’t just represented, but actually required.

This passage is from the video game Zork I. Published in 1980, this old game demonstrates many of the techniques native to the kind of storytelling that is my obsession: interactive fiction. Interactive fiction is any narrative that collaborates with its audience, relinquishes power, and marshalls computation in service of mystery, experimentation, and delight. In Zork I, after reading a short description of your surroundings in each area, you’re presented with a command line—an input field that allows you to type in actions and thereby direct the course of the story.

Interactive fiction has an intimate relationship with video games, and it flourishes when composed using digital tools. But writers and artists have been collaborating with their audiences since long before the computer age. There’s Yoko Ono’s 1964 experimental art book Grapefruit, which offers its participant readership instructions they may perform: “Draw a map to get lost,” for example. And in the early 1940s, Jorge Luis Borges imagined a labyrinthine novel containing endless scenes and outcomes in his short story “The Garden of Forking Paths.”

I find this history reassuring. It demonstrates that experimenting with interactive fiction isn’t an attempt to “fix” literature through technological innovation. Rather, it’s a way to explore ideas that have existed in avant garde art practices for many years.

One of the best tools to begin this exploration is Twine. This free, open-source tool for telling nonlinear stories is easy to learn, and you can use it online. Though it’s not the only tool out there, Twine is beloved by interactive fiction teachers and writers, so you’ll find it’s well supported by a thriving community interested in the possibilities of literary games.

Writers working with Twine for the first time encounter a set of exciting new creative problems. Rather than writing pages, Twine tasks you with the composition of passages—units of text that may be connected using a series of hyperlinks. The Zork I passage above, for example, might look something like this in Twine:

As you can see, the choice offered in the passage, to go north or south, is visualized as two mutually exclusive passages linked to the original one. Choosing north, the reader arrives in a new “location” and reads the text there, which may offer yet another decision leading to additional passages.

In my mind, the organization of interactive fiction passages and their relationships relates with what in traditional fiction is termed “story structure.” Typically, we visualize narratives as composed of acts, beats, and scenes. And there’s Freytag’s Pyramid, the diagram of rising action, crisis, and resolution that’s so central to breakdowns of western world literature. Stories, our English education tells us, build themselves around heroes and thrive amid conflict and thwarted desire. But what happens to structure when you write a story that allows for reader choice—the option to go north or south?

This “tree,” called a fractal canopy by mathematicians, is one of the most common structures to emerge as fiction writers adapt their work to interactive media. It’s the logical endpoint of the process I began above. The reader is presented with a choice, resulting in two (or more) story outcomes. Then each of those outcomes results in more choices. Branches stack upon branches, until soon you’ve got your very own garden of forking paths.

But if you’ve read the Borges story, you know that Ts’ui Pên never quite finished his branching opus. The problem is, bringing each potential narrative to its conclusion requires generating a huge number of endings. It’s theoretically beautiful—who doesn’t love fractals?—but as an organizing mechanism for writing, the branching-tree structure becomes overwhelming quickly. Confronted with vastness, the temptation to admit defeat is intense.

Still, writers do find their way through. The interactive fiction community is expanding, and I’m filled with joy every time I encounter a video game that also positions itself as literature. And if the literary world is going to delight in more such works, I think we’ll need more writers ready to contend with the infinity at the heart of all fiction.

I offer you the following leads on achieving meaningful structure in interactive fiction, brought back from my own experiments in Twine and teaching narrative design. These three alternatives to the fractal canopy all accommodate choice and unknowability, while also offering you, the writer, a reasonable degree of narrative control. If you’re sitting down with Twine for the first time, I humbly suggest that having pictures of these structures in mind may be helpful as you work to develop story-specific versions of them on your own.

My first structural offering involves the use of what I call “key nodes.” In this approach, there are some passages—set pieces important to the core narrative—that every reader sees. But the territory between these nodes is variable, and the reader can move from one node to the next in a variety of different ways.

The narrative indie game Night in the Woods uses a structure like this. The game takes place over a series of days; as protagonist Mae, you select which of Mae’s friends to hang out with during each day. Will you choose the mischievous, leather-jacket-clad Gregg, or the moody goth Bea? Each choice forecloses on the possibility of spending time with the other friend. But some events, like the spooky Harvest Festival, always happen, characters and plot converging as evening sets in. In Twine, such a narrative structure might look something like this:

In this story, we begin with a single passage and branch out, offering choice and multiplicity in the passages the reader visits next. The playful excitement of going north or south remains intact. But rather than branching infinitely, the passages converge at important plot points, passing the reader through a series of choke points, or gates. For the reader, who can only ever see one passage of text at a time, choke point passages seem to proceed naturally from optional ones. Unclear on which nodes are peripheral and which are central, your audience will interpret every beat in the story as both self-chosen and critically important to the plot.

When using this kind of structure, you will relinquish some control over tone. Different choices in the variable parts of your fiction will result in different emotional experiences and many possible reactions to your final ending. This is a good thing, because it will heighten the value of rereading. You’ll also be free to focus on other elements of writing fiction, like building stakes and crafting rising action, because using key nodes allows those tools of craft to remain intact.

Not all interactive fiction progresses linearly from left to right, though. My second structural offering attends to stories that loop, building their narrative argument atop the algorithms that underlie computer programming. Queers in Love at the End of the World by Anna Anthropy is an excellent example. In that game you’re only given ten seconds to make a series of narrative choices before the world ends. It’s an amount of time so short, you can’t help but play and replay this queer musing on intimacy, annihilation, and what constitutes a “right” choice just as choices cease to matter. In Twine, a looping story like this might be structured as follows:

An important thing to notice here is the lack of a single cohesive ending. This style of interactive fiction often builds looping into its narrative reality, embedding cycles of death and rebirth in the writing as a motif as well as a structuring principle. I have seen looping interactive fictions that allow readers to escape the loop, achieving a “true” ending after a certain number of iterations. But I like looping fictions best that celebrate getting lost and examine the larger literary resonances of endless repetition. These works demonstrate some of the themes that interactive fiction explores particularly well, maybe even better than traditional fiction.

When using this kind of structure, many writers end up learning a bit about code and programming. It’s natural that an algorithmic story structure would lead in this direction, and I heartily recommend looping fiction as a first foray into learning the basics of Twine programming. Think of the timer in Queers in Love at the End of the World—it’s a small piece of code that exerts a powerful narrative effect on every piece of writing in the game. In this context, a little can do a lot, which means you’ll see a great artistic return on your investment of time.

Once you start exploring the documentation for your Twine game’s story format (the default is Harlowe), you’ll discover all kinds of narrative effects that are possible to achieve with code. The third and final Twine structure I’ll describe often relies on variables, for example. These handy-dandy tools allow you to store a piece of information in the computer’s memory and access that information for various purposes later down the line.

If you haven’t yet played Zork I for yourself, you should. It’s available online for free, and it’s an excellent example of an interactive fiction that’s structured spatially, organizing its various interactions and tensions according to a map. The story of the game unfolds as you navigate contiguous areas and build an understanding of the space you’re in for yourself. Here’s what a map-based structure might look like when composing in Twine:

The most game-like of these three structures, map-based interactive fiction engages the reader by offering a world to explore. Narrative momentum comes from curiosity, the impulse to understand what’s here and why. Often, exploring a map involves what are termed in game making “lock and key puzzles.” This is just what it sounds like: You must find keys in order to unlock doors leading to new areas.

In Zork I, there’s also the possibility of obtaining items, like a lantern, which allows you to illuminate the many rooms of the vast underground dungeon lurking somewhere below the tree we’ve been lingering in for so long. When a reader “picks up” a key or an item in a game like this, that information is stored in a variable, and logic is checked so as to display new or different text or hyperlinks as a result. There’s so much fun to be had in building a story that incorporates logic like “If the reader is holding a key, allow them to open a door.”

Naturally, because each passage represents a room or other physical location in this style of interactive fiction, map-based storytelling happens through interaction with the environment. Your narrative will unfold as the reader encounters characters, gathers objects, or reads textual ephemera left behind in rooms. Of the structures I’ve proposed, this one is the most immersive, and it has a strong relationship with thought patterns used by video game developers. That said, all three of the structures I’ve mentioned are used in game writing, and there is no single answer to the question of what a game “is.”

As you navigate the world of interactive fiction, I hope you’ll carry with you my offerings on structure. Comparing visuals of different Twine constructions will help you critically consider the shape of yours. None of this, of course, is a rule. Many mature interactive fictions combine these structures, using nodes, looping, and maps to create an engaging experience imbued with variety, challenge, and mystery. And some games use other structures, unique constructions of their own.

In fact, if you have the time, I’d suggest (re)playing a narrative game that inspires you and reverse engineering its structure. Regardless of whether a fiction was built in Twine, try to imagine how the narrative was designed in terms of connected passages. My bet is you’ll start to see these structures everywhere, even in works of art that claim to be linear and noninteractive. In fact, my opinion is that the difference between interactive and traditional fiction isn’t a clear binary at all: Both require an approach to structure that balances openness and control, and each form of writing contains lessons for working in the other.

If there is one hard-and-fast rule I can offer to new writers of interactive fiction, it’s that if you start writing a game, you should commit to finishing it. Ending any story is challenging, and interactive fictions can be more so. That’s why you should do it. When I teach Twine, I am very serious about this. I say writers are strictly required to complete a finished prototype, even if it’s unsatisfying and full of bugs. Without this injunction, it’s easy to get lost in the woods, lingering forever upon small clusters of branches.

Avoid the temptation to linger, and you’ll reemerge with something beautiful, imperfect, strange—even amusing. You will have created interactive fiction, the outsider art that I so long to read.