Teaching this course has raised a great many new questions in my mind. Perhaps it is good that I have stepped away from actually building interactive storytelling and concentrated on trying to teach its principles; this has forced me to think through all the issues from a greater intellectual distance and perceive larger issues I had not seen before. Herewith some of those conclusions and questions.
Complexity
The greatest difficulty I faced in building interactive storytelling technology, and the one that ultimately confounded my efforts, was the incomprehensible complexity of the systems I was building. In each case, my work grew so complex that I simply couldn’t keep ahead of it. I would correct one problem, only to create two new problems. This is a standard experience in programming when the code has gotten too messy. The solution in programming is to plan more carefully and to use a higher-level language. Neither of these solutions were available to me. How can you plan something that has never been done before? How can you move to a higher-level language when you’re using a language of your own creation (Sappho)?
The problem can only be addressed at the highest levels. Here are some strategies to keep the complexity at a manageable level.
KISS!
Keep It Simple, Stupid! This is the most important advice I can offer. You will go too far, I guarantee. You think that storytelling is simple and obvious, but that is only because you have been experiencing stories for decades. Here’s a simple exercise to help you understand the difficulty. It starts with two stories:
Story #1:
Itsy bitsy spider crawled up the gutter spout.
Down came the rain, and washed the spider out.
Out came the sun, and dried up all the rain.
And the itsy bitsy spider crawled up the spout again.
Note that this is a well-formed story in just four lines. It has a protagonist, an antagonist, a crisis, a resolution, and even an edifying moral. Any five-year old child can understand this story.
Story #2:
Once upon a time there was a handsome young prince who lived in a shining castle atop a hill.
One day, he leapt onto his mighty white charger and galloped out across the courtyard, and out into the forest.
There they fell into a hole and they both died.
Any five year old child can tell you “That’s not a story!” It isn’t; it’s all screwed up.
Any five year old child can tell you which of these two stories is a well-formed story and which is not.
I challenge you to devise an algorithm that can evaluate and determine which of these two stories is well-formed.
You cannot express in algorithmic form the story-analyzing skills of a five year old, can you?
That’s how complicated stories are. Don’t EVER forget this!
Over and over you will endow your storytelling system with clever schemes that permit what you think to be essential components of storytelling. You will then be hoisted by the petard of your own cleverness, and will be unable to get your system working satisfactorily.
Remember, this is ART! Here’s a Japanese drawing of a cat:
The triumph of this drawing is that it succeeds in capturing the essence of what a cat looks like in just a few curved lines. This is brilliant art: reducing something to its essence. The Japanese artist had the option to draw a complex image, but you do not have that option. The algorithms you can “draw” on a computer are like the simple lines and curves of the cat image; you don’t have the talent in your “mental fingers” to “draw” algorithms analogous to this drawing:
So you’re just going to have to use your artistic talents to make up for your lack of mental dexterity. Deal with it.
Game-thinking
In the earliest years of the cinema, nobody knew what “cinema” was. This utterly new medium of expression left people groping for meaning. They turned to the closest thing they could think of: the theater. Cinema is just theater with the additional feature of being recorded—or so they thought. So they made movies in theaters, placing the camera in the front row center and having the actors play their parts. It took years for them to slowly divest themselves of the prejudices that theater-thinking brought to their creative process. It was inevitable that the person who did the most to break free of theater-thinking was a failed playwright, D.W. Griffith. Being a failure freed him from any emotional attachment to theater-thinking and freed him to think exclusively in terms of what could be done in cinema. He went on to create much of the language of cinema.
Nobody knows what interactive storytelling is, so everybody turns to the closest medium: games. Interactive storytelling is just games with storytelling added, right? So everybody thinks. You need to purge that thinking from your mind. Gaming prejudice will always send you down blind alleys.
The best example of gaming prejudice is the inability of gamers to get through their skulls the notion that spatial reasoning plays no role in drama. I’ve been preaching this message to games people for decades, and they just cannot free themselves from their idée fixe about space.
Consider the universe of Star Trek: it’s huge and complex. Amazon.com boasts 75 PAGES of books about the Star Trek universe. You can find detailed information about the different races and planets, timelines, technical specifications for the different spacecraft, and on and on. The one thing you can’t find is a map. There is no map of the Star Trek universe. That’s because spatial considerations have no place in storytelling. Characters jump from stage to stage. Spaceships jump from planet to planet.
Sure, there are stories about journeys. Huckleberry Finn is a good example of such stories. But consider what happens in the book: Huck and Jim travel down the Mississippi River, jumping from stage to stage. At no point do we know precisely where they are or how far they have traveled. There are no distance or time specifications. It’s a journey without distance.
Ultimately, drama is not about maneuvering in space. It’s about PEOPLE! So lose the damn spatial reasoning!
Future Directions
How can we continue the development of interactive storytelling?
The Phylogeny of Interactive Storytelling
If we consider the process by which we develop interactive storytelling in evolutionary terms, we begin by recognizing that interactive storytelling will have a phylogeny: a sequence of developments beginning with the simplest forms and advancing to more complex forms. The question then becomes, how should we shape that phylogeny? What sequence should we pursue in developing interactive storytelling?
One useful guide is Haekel’s Law: Ontogeny Recapitulates Phylogeny. The development of an individual living creature follows a sequence of steps similar to the evolutionary path that species followed. We could apply the concept of Haekel’s Law as a guide in our development of interactive storytelling. To do so, we must reverse it so that the phylogeny of interactive storytelling—that is, the evolutionary sequence it follows—recapitulates the ontogenic sequence by which a person learns about stories. In other words, we should begin with children’s stories, developing interactive story worlds that encompass the range of children’s stories. I am currently contemplating how to research children’s literature as a body of stories.
Introducing Characters
One of the more difficult problems I have experienced lies in making players aware of the personalities of the different characters. Drama works only when we have a sense for the personalities involved, and the player cannot be confident of any decisions without a firm feel for the personalities of the characters. In traditional literature, the revelation of the personality is an intrinsic part of the story, but in interactive storytelling, we face a dilemma: the first time the player encounters the storyworld, the player needs that information, but on subsequent replays, the repetition of the introductory material becomes tedious.
I see two solutions to this problem. The first is to provide a conventional introductory story to accompany the storyworld; the player must first experience the introductory story before entering the storyworld. This could be done with text or video. The player can then skip the introductory material on subsequent replays.
A second solution is to intersperse background material on the different characters through the use of Encounters. With a large enough base of Encounters, the player can continue to learn more about the characters and the storyworld with each replay.
Artificial Intelligence
After decades of impotence, AI has finally blossomed into a tool of great power. The magic ingredient that triggered this explosion was the availability of huge amounts of information on the web. That information was used to train neural networks to recognize complex patterns that were previously unfathomable.
Theoretically, much the same could be done with storytelling. We certainly have monstrously huge databases of stories; tens to hundreds of thousands of new stories are presented each year in movies, novels, and television shows. But there’s a killer problem: recognizing the structure of a story remains far out of reach of our algorithms. Recall the example above with the two simple stories; if we cannot devise algorithms for comprehending such simple stories, how will we be able to analyze the latest episode of, say, Game of Thrones?
There have been some such efforts. Vladimir Propp’s Morphology of the Folktale comes closest to our needs, but ultimately doesn’t work. Georges Polti’s The Thirty Six Dramatic Situations is too high-level to be of use. Stith Thompon’s The Folktale provides a great deal of useful information, but it is not organized enough to be of direct use. Barre Tolken’s The Dynamics of Folklore provides some interesting high-level analysis of folktales, but again, the information is not categorized well enough to be of use. At the opposite extreme is the Aarne-Thompson-Uther System of motifs of folktales. Its flaw is that it catalogs Objects, not Processes. (Interestingly, Vladimir Propp made exactly this criticism nearly a century ago.) Of greater utility is An Atlas of Interpersonal Situations, but I warn you that the model this book presents is difficult to apply without thorough study.
I’m done for now.