I
A few weeks ago, while reading through Khaled Hosseini’s The Kite Runner with my senior English class, I was trying, and failing, to teach a lesson on story structure. We were hung up on making predictions about the last third of the book: now that Amir had left behind the turmoil of Afghanistan and established a life in America that was comfortable and pleasant as long as he didn’t think about how he betrayed and abandoned his oldest friend, what would happen next? My students stumbled through a few half-hearted answers, trying to find the right answer. Maybe Amir would write a book about his experience and donate the money to charity? Perhaps he’d go back to Afghanistan and reopen his father’s old business? They were grasping. Finally, a student raised her hand, and guessed that Amir would go back to Afghanistan and try to make amends for his sins, facing great danger and paying a steep price. I asked her what made her think that.
“I don’t know,” she said. “That’s just the rule for stories, right?”
She was right, of course. The first two hundred pages of The Kite Runner set up a conclusion that has to happen for the story to make sense and entertain. If Amir never goes back to Afghanistan and resolves his guilt through, say, therapy and knitting, the book would feel broken, and the reader would feel cheated. These are part of the “rules” for fiction.
Lately, I’ve been thinking a lot about rules. They’re central to the philosopher C. Thi Nguyen’s theory of games and gaming, which allows him to explain everything from Dungeons & Dragons to mountain climbing under a single framework. Nguyen also has a knack for finding this logic of gaming in unexpected places like Twitter. I think his theories also go a long way towards explaining what my student meant when she said that fiction has rules, and why that seems to help explain something about storytelling.
II
Let’s start with what Nguyen actually says. He starts with an old definition of gaming by Bernard Suits: a game is a voluntary attempt to overcome an unnecessary obstacle while achieving a goal. A soccer player’s goal is to put the ball in the net. Their obstacles are the opposing players and the strict requirement to only move the ball with their feet. If you take away the obstacle, the game is boring, and without the goal, there’s no point to playing. And so, for the duration of the game, everybody playing agrees to voluntarily compete to satisfy a goal while following arbitrary rules.
We can see this with board games and party games, too: in chess, you have to place one of your pieces on the opponent’s king while they try to do the same to yours, but both of you have to take turns and follow a specific set of legal moves. In charades or pictionary, your goal is to communicate a particular word or idea, but only in a specific, non-verbal way. Suits’s goal-obstacle framework doesn’t always work for all types of games, as Nguyen admits, but it fits most games and sports, and puts the emphasis on goals and rules.
Although goals may shape the nature of a game, they aren’t the reason for playing. Basketball players don’t play their sport to satisfy a primal urge to put balls in hoops, and boxers don’t go around punching people for fun. Nguyen distinguishes between the goals that structure a game and purposes that make us want to do them. The goal of mountain climbing is to get to the top; the purpose is to have fun, or exercise, or be in nature, or win a cash prize.
I once went to a kind of Maya theme park in Mexico where performers dressed up in traditional feathers and jaguar skins to play pitz, the famous Maya ball game. They were playing the same game as their ancestors did, with the same goal of knocking a heavy rubber ball into a hoop using only their hips. But where their ancestors thought of pitz as a dramatic reenactment of cosmic forces in opposition, with the winners becoming exalted heroes and the losers ritually sacrificed, their descendants were playing a difficult, boring game to earn money and entertain gringos. Same goal, different purpose.
Obstacles, goals, and purposes all come together in Nguyen’s theory to express something about experience and the world. In other words, games are an art. Building on Thomas Dewey’s aesthetic naturalism, Nguyen defines an artwork as something that structures the stuff of common experience like sight and sound in a meaningful way. Music is the art of sound, organizing it in ways that evoke emotional responses, painting is the art of sight, sculpture the art of form, and literature the art of language. Games, for Nguyen, are the art of agency. They take the ordinary process of deciding and doing, add goals and obstacles, and shape the experience into something meaningful and fun.
When we think of games as the art of agency, we can examine games like any other artwork for what they say about the world and our experience of it. Poker is about managing risk, probability, and deception. Chess and go turn tactical decision-making into a beautiful, easily understood experience. Most sports take the basic physical operations of the human body to their peak. Video games, depending on their type, explore resource management, reflexes, mathematical reasoning, working memory, and more. All of this is only possible when the game has a clearly-defined goal and a set of rules.
III
Nguyen’s theory allows us to see any kind of rule-based, goal-oriented system as a game. This doesn’t mean that, say, Congress or Twitter are actually games, but the theory gives us useful models. If the goal of the Twitter Game is to earn the most likes and shares, and those goals are most easily met through relentless shitposting, then Twitter is the trolling game, and everything else is ancillary (Justin E.H. Smith: “ I take the expression of substantive political opinions [on Twitter] to be something like the expression of substantive political opinions on, say, Fortnite: an absurd proposition, as whatever the opinions are, they are interrupting the flow of an otherwise engaging video game.”)
What is the Fiction Game, then? Setting aside the broader purpose of making literature (instruction, cultivating a rich inner life, making money, airing your grievances, accumulating social capital), the interior goal of most fiction, at any length, is to tell an engaging story about characters facing problems. As my student suggested, there are certain rules about how a story should be told, and much of our appreciation for how well a writer follows the rules of the Fiction Game.
We can see this most clearly in genre fiction, like murder mysteries, where certain rules have been in place for nearly two centuries: the murder must be solved by the end of the story, with the protagonist playing a crucial role in the resolution, having pieced together clues that are neither too obscure nor too simple to catch a culprit who is not the most obvious suspect. These rules are so ingrained that we notice them only in the breach: the experimental Oulipo writing group once got its hands on S.S. Van Dine’s famous “Twenty Rules for Writing Detective Stories” and competed to see who could break the most rules in a single detective story. The winner was a murder mystery where the criminal only appears in the final paragraph, having hidden in an undisclosed secret passage which opens accidentally, the killing itself done unintentionally and by undetectable poisons that leave no clues behind. Naturally, the story was called “The Butler Did It.”
But even literary fiction, which usually defines itself by a self-conscious opposition to predictable genre tropes, has its rules. As with The Kite Runner, much of it is structural, with authors compelled to answer the questions they raise, or at least lead their characters to a moment of insight. Characters in a Joyce or Chekhov story rarely fix their problems, but at least understand them better by the last page. Not even the most hardcore experimental fictions can completely get around the ancient, Aristotelian need to have a beginning and an ending; a seeming copout, like the circular story of Finnegans Wake where the last sentence becomes the first, needs to have an identifiable first and last page for the idea to make any sense. A truly circular story could start and end anywhere, and Joyce was very particular about where his last book started and ended.
Just like we admire a game-player’s ability to satisfy their goal while avoiding obstacles and following rules, much of the pleasure in reading good literature comes from the author’s way of navigating these rules. Funnily enough, it’s often the most generic, predictable stories–I’m thinking of romances and mysteries, which always have the same limited set of outcomes–that have the most voracious readers. Everybody picks up an Agatha Christie novel knowing that the murder will be solved at the end; it’s seeing how she does it, while following all the rules, that makes it entertaining. Writers of literary fiction often set themselves the challenge of saying something profound and beautiful from the most meager premises, as in the novels of Woolf or Knausgaard.
IV
The stakes for the Fiction Game might be higher than you think. In the last few years, AI writing programs have surged in sophistication, ease of use, and talent. Programs like GPT-3 can compose expository essays up to 1,5000 words at the level of a college undergraduate in an instant. You can feed it the complete works of any poet and ask for a poem in that style, and depending on the style and form, the results range from pretty good to indistinguishable. One thing GPT-3 and its ilk still can’t do quite as well, Erik Hoel found out, is write fiction. For a lot of complicated reasons owing to the complexity of fiction and GPT-3’s reliance on statistical frequency analysis, the software still struggles with “long-term coherency, causality, common sense knowledge, [and] character development.” In other words, an AI program might read hundreds of thousands of novels, but it still can’t figure out the rules of fiction. Or not yet, anyway.
Alongside Nguyen’s work, I recently read Oliver Roeder’s excellent Seven Games: A Human History, describing the history and culture of the world’s seven most popular analogue games. A funny thing about the book is that it pulls double-duty as a history of computation and artificial intelligence: starting in 1952 with tic-tac-toe and continuing up through chess in 1997 and go in 2016, computers have slowly come to dominate just about every game in the world. In every case, these programs depend on a mixture of experience and clearly-defined rules. Writing programs already have the experience, reading hundreds of books per minute; how quickly they master the art of fiction may depend on just how correct my English student was when she said that novels have rules.