I’ve been struggling for several months now trying to nail down precisely what I mean by ‘interactive storytelling’. Sure, it’s storytelling that’s interactive, and because it’s interactive, I apply Crawford’s first law: When considering any software design, the very first question you must ask, and the question you must ask repeatedly during the design process, is “What does the user DO? What are the verbs?”
This led me, from the very outset, 25 years ago, to think about how to create verbs for interactive storytelling. I pushed that line of effort almost exclusively. It didn’t really work.
I’ve been meditating about these problems, and I had a realization a month ago that I’m still digesting: there is little dramatically significant interaction among characters in stories. This may seem crazy at first, but follow my reasoning.
First, let’s note that a great many stories are really just long obstacle courses. The protagonist sets out seeking some goal, and then encounters a series of obstacles that must be overcome. Among these stories are the Odyssey, most action movies these days, The Lord of the Rings, Culhwch and Olwein from the Mabinogion, the quest for the Grail in the Arthurian legends, Star Wars, the movie Gravity, and so on. The plot can be summarized as “It’s just one damned thing after another.” No sooner than the protagonist has dispatched one challenge than another leaps out at him. At the end of Gravity, as the astronaut unsteadily walks away from the lake in which she landed, I half-expected a tiger to leap out from the bushes.
These stories can certainly be fun and exciting to experience, and we’ve done an excellent job of making them interactive in the form of video games. The player must kill an endless sequence of monsters, solve an endless sequence of puzzles, and so on. Video games clearly demonstrate that this is NOT dramatic. It’s exciting, it’s fun, but it’s not storytelling, and it certainly isn’t art.
So why does it work in storytelling? It doesn’t. Obstacles like this are window dressing, fun little diversions that make us feel good but have zero artistic content. Scylla and Charybdis were certainly scary, but you can’t assign any artistic content to that. The same thing goes for Culhwch’s tasks or Frodo’s escape from the giant spider.
What’s especially revealing about these components of stories is that we all know perfectly well that the protagonist must and will overcome them. Suppose you are reading The Lord of the Rings and you come to a page that says, “So the spider killed Frodo and sucked out his blood. The End.” Would you not be aghast? THAT’S NOT A STORY!!!!! So what’s the point of all these digressions?
There are two points. First, they create vicarious entertainment. The reader feels a sense of danger as the spider approaches Frodo, a sense of panic when Frodo gets stuck in the web, and a huge rush of relief when Frodo gets away. You’d think that people would get bored with the same old formula replayed a thousand times, but I still love to eat chocolate even though I must have eaten thousands of pounds of chocolate in my life. These humans are easily amused.
Second, they generate empathy for the protagonist. After vicariously following them through all those dangers, you end up feeling close to Frodo, Neo, Luke Skywalker, and Odysseus. (But maybe not Culhwch). This lends power to the few dramatically significant decisions they actually make.
Those dramatically significant decisions usually take place at the very end of the story. Every good story has some sort of dramatic crisis during which the protagonist must make a decision with profound significance. Will Luke Skywalker trust The Force and turn off his targeting computer? Will Frodo drop the ring into the fires of Mount Doom? Will Trinity’s kiss bring Neo back to life? These are the key points in the stories; why couldn’t we just skip through all that rigamarole of getting out of the trash masher in the Death Star, sneaking through Mordor dressed up like short Orcs, or shooting up a building to rescue Morpheus? Ultimately they don’t bear on those final dramatic moments, so why don’t we just skip them and get straight to the good part?
The answer, of course, is that all that preparatory stuff was necessary to establish the empathy the audience feels for the protagonist. If the movie starts with Frodo sauntering into Mount Doom, dropping the Ring into the fires, and saying “So long, Sauron!” I don’t think many people would be entertained.
The conclusion from this reasoning is clear: interactive storytelling will consist of an obstacle course, much like a game, in which the player must solve a series of puzzles or problems, before the player finally reaches the end of the obstacle course and gets to make one big dramatic decision that ends the story.
This is a truly disheartening conclusion. After all, lots of video games do precisely this, and we all agree that they are not providing what we expect from interactive storytelling. Sure, it’s loads of fun blowing things up, figuring out how to get past the guard, figuring out the trick, and so on, but it’s not at all like the experience of a story. It has no real emotional punch beyond elation at success.
When Samwise plunged into the river to follow Frodo across the lake, very nearly drowning in the process, the moment was dramatically powerful, but how could we create something like that in an interactive environment? We wouldn’t give the player a choice between saving Samwise and letting him drown. We wouldn’t have a short video game in which Frodo must dive into the water and swim down to Samwise and pull him to the surface. So what can we do?
So we come to the conclusion that, indeed, interactive storytelling is impossible. It can’t be done. I wasted 25 years of my life chasing a chimera. Let’s all just give up and go back to making stupid videogames. Right?
The Hand Shoots Up From the Grave
Perhaps
you have already noticed the chink in the line of reasoning above: it
discusses only one kind of story, the kind of story that dominates
popular cinema. There are a great many more stories that aren’t obstacle
courses. Let’s talk about different categories of stories.
Mysteries
As
I have explained elsewhere, mysteries are a special form of
storytelling in which the author toys with the imagination of the
audience, inducing the audience to generate multiple stories while
experiencing the story. At the outset, the audience doesn’t know enough
to make any guesses, but partway into the story, enough evidence has
been adduced that the audience forms a hypothesis: ‘the butler did it’.
But then the author pulls the rug out from underneath that hypothesis:
the butler was in the garden when the murder took place. Another
hypothesis then appears: perhaps it was the nephew. Or Professor Plum…
the author tantalizes the audience with possibilities, until the final
moment when the murderer is revealed to be… the nun!
Of course, audiences have gotten better at sniffing out hypotheses, and so some authors have pushed the boundaries ever further, making the mystery harder and harder to figure out. More commonly, the author doesn’t reveal the crucial clue until the very end; nobody could have solved the mystery without that crucial clue.
There have been many attempts at mysteries on computers. Because they are essentially puzzles, most mysteries have followed the format of the adventure game, which is really a big collection of puzzles. They fail because they always hinge on some physical puzzle, rather than a social interaction. Besides, mysteries almost always turn on some tiny detail that no player could be expected to notice. I recall a Poirot mystery in which the mystery was solved because Poirot recalled that, during an early conversation, the murderer (a sculptor) mentioned to Poirot that she hated horses. So when he saw in her flat a clay figure of a horse, he seized it and ripped it apart, much to the shock and dismay of his colleagues. There, hidden inside the clay horse, was the crucial clue that solved the case. Admit it—we couldn’t put that into an interactive storyworld, could we?
Soap Operas
These stories seem
ideally suited to interactive use, because they have no plots. The
characters in soap operas meander through a limitless dramatic terrain;
they can do pretty much anything. There are no obstacles, just conflicts
among characters. But soap operas bring a problem all their own: the
social interactions are far too complex for anything we can build just
yet. Consider this imaginary fragment of a soap opera:
Colleen: “Brad and Theena went to the movies on Tuesday afternoon.”
[The organ sounds a loud crash of dissonance.]
Marianne: “But Brad told me that he spent Tuesday afternoon playing golf.”
Colleen: “Perhaps it was on a different Tuesday.”
Marianne: “I was suspicious of Brad. That’s why I asked him what he did Tuesday afternoon.”
You try to write a storyworld that can figure out what’s going on. This is standard fare, in soap operas, but all the action takes place at extremely indirect levels of social behavior. And we can get into some very deep soap operas such as the works of Jane Austen. I don’t think that we can build interactive storytelling around soap operas.
There are many other types of stories, though. Some stories are really just collections of observations on human behavior; Huckleberry Finn is my favorite example of such a story. Huck really doesn’t do much in the story; he just drifts down the Mississippi with Jim and observes the many foibles of humanity. Yes, there are a few obstacles, but they play a minor role. There isn’t much opportunity for interaction here.
The Shakespeare plays provide us with another source of analysis. Can we imagine “interactivizing” any of the Shakespeare plays? That could only be done by vastly expanding the possible developments. A player in the role of Hamlet might have other options for dealing with his uncle, but could they really be worked into the system? And how would the author design them to function?
This line of thinking does suggest that I have been pursuing entirely the wrong approach in my attempts to devise a language for general dramatic interaction. Human behavior is manifested in too many ways to be managed with a set of standard verbs.
But I have already shown that a standard language could be used for statements of perceptions. That is, we already have a language system for statements as to how much a person likes, trusts, and dominates another person. So here’s the issue: could we interleave a very detailed encounter system with a standard language system to permit a workable interactive storyworld?