1001 Electronic Story Nights

The Explorer-Guide relationship comes in a variety of flavors. In the "Wheel of Life," we were concerned with creating a theatrical experience within physical space. The Explorer might not even recognize that the Guide is present and playing a role. But the principle of such a relationship can lead us into some very rich emotional situations. Consider the story told to me by a friend (who is also a new media critic). David had become very discouraged about the limited emotional range provided by interactive experiences, and frequently complained that the medium would never succeed until we can trigger aspects of surprise and delight in a viewer. However, one day a flushed and excited David related to me how he had reached a sort of interactive epiphany. He had logged onto a MUD -- does everybody know what a MUD is? Okay, I don't have to explain it? I do? In brief, a MUD (or "Multi-User Dungeon") is a text-based, networked experience which can simultaneously accommodate any number of participants. It is generally constructed using a spatial mise-en-scene: visitors can create new rooms or enter existing rooms in which other people on the network are "hanging out." Participants visit MUD's in search of semi-anonymous social interaction, where they are completely free to be themselves -- or to be someone else, playing a fantasy role. Interestingly, MUDs are often peopled by programs which simulate humans, as well as by actual humans. MUDs are slowly evolving into graphical worlds where a participant will be represented by an avatar, a sort of audiovisual surrogate for yourself which dwells within the synthetic space and acts under your remote control. Perhaps you have heard of them. What is most important about a MUD is that it provides a social venue for a community of players. Returning to the story, my friend, David, entered a MUD on the day in question. He journeyed into a bar. A group of other people were standing around the bar talking, and there was a robot in the room, standing all alone. The robot seemed sad and lonely because nobody was talking to him. David, being a sociable fellow, went over to meet this robot. Instantly, the robot seemed to cheer up and began to share his stories with David. In retelling this experience, David observed that at some point he was surprised to notice that he had placed his arm around the robot's shoulders. Shortly after this, the robot asked David to tie him up. What would you do if a robot in a MUD asked you to tie him up? David considered the proposition and concluded that he had established a trusting relationship with this character and that, in order to be true to their friendship, he would honor the robot's wishes. However, no sooner had he finished tying the robot up, then the robot turned and furiously berated him: "How could betray our trust, how could you capture me and tie me up!" (All of this was transpiring in text, of course). David was stunned. He claims that he did not know what to do next. Everyone in the bar had followed David's recent interaction with the robot. They had all witnessed the robot's temper tantrum. They sympathized with the robot, and unanimously turned on David. David felt at once betrayed and emotionally lost: he claims that the feeling stayed with him for days. I know this sounds ridiculous, but it shows that the person-to-person connectivity over the network -- and particularly, the identification with yourself as a first-person character or player -- introduces the potential for a truly intensive emotional interaction.

Now, I'd like to show you a video. We will need to turn the lights down, but first let me set the stage. What you will be looking at is the visual portion of the Ph.D. thesis of Tinsley Galyean, one of my students at the Media Laboratory. The story, "Dogmatic," runs on a SGI Reality Engine. It is a one-on-one experience in a 3D virtual world. You control your point-of-view by rotating a mouse in front of you in 3-space. Now, I need everyone in the audience to feel that you have the mouse in your hand and you are looking around the space. The reason that this part is so jiggly and uncinematic is that the active user is learning that she can control where she looks within the synthetic scene by moving the mouse. In order to have any sense of the experience, you must pretend that you are controlling the view shown in the videotape. You're looking around the environment. You hear a car; you try to keep the car in sight. The dog jumps out of the car and pads toward the cactus. If you had taken your eyes off the dog, he would have come over and barked at you to get your attention before he takes a pee. The movie continues: the dog brings a severed human arm over to you: clenched in its fist is a note, "Lucky Strikes." Later, there is a fight; you hear the car screeching out of control, coming towards you, you black out and come to. You see the world on its side.

This film noir is experienced in a virtual environment with a virtual character, the dog, who exhibits some high-level autonomous behaviors. In an interactive story, the actions of the user must have consequence. In this experience, you turn toward the sound of the car; this triggers the action which results in your catastrophic downfall. "Dogmatic" highlights a range of research issues for interactive narrative. The piece challenges the continuous flow of space and time which has become the tradition of virtual reality experiences. I think "Dogmatic" shows us that cinematic language can enhance a virtual, first-person story. In this piece, the cuts are all executed under programmatic control: the program detects where you are looking and determines when to cut away, and for how long. For some short period of time after the cut, we take away the real-time control over point of view; this allows us to normalize your new view of the world. The effect at the end of the experience startles the participant viewer. Almost everybody who experiences this work in first-person mode is absolutely wiped out when they realize, after the crash, that they no longer have physical control over the world. They cannot sit up. The emotional response to this ending is similar in effect to David's rejection in the MUD, which I mentioned earlier. Both scenarios reveal the real consequences of your actions, which in turn generate emotional investment and reaction.

The irony of the ending is nicely revealed. The fact that the viewer can look around the world gives the participant viewer the false impression that she can control the world. In reality, the viewer has very little control. The program has a fixed content which may be triggered by the viewer's actions; however, the viewer is not aware of the exact relationship between their act of looking around and story. Most viewers do not realize that they cannot walk down the street or enter one of the venues of the desert town. From the author's viewpoint, constraining the viewer's movement placed realistic constraints on the author. We did not have to juggle the implications of multiple noir endings: you can only die once.

"Dogmatic" required a musical accompaniment which could adapt to the pacing of the participant viewer's interaction. The music is written in a somewhat ditsy, plinky voice because it must be terminated or extended depending on viewer action. Dialog is totaling missing from the piece. If we are going to invite dialog, then the storytelling system needs to have some understanding of what the viewer might say and respond appropriately. This is a big research area, and a comprehensive solution is unlikely to appear in the immediate future. By pointing out certain limitations, I would like to argue that the storyteller system is not a magic box which already exists; rather, it is a system which is designed for (and around) particular interactive stories. In order to make such a magic box, there needs to be a plan, a taxonomy, a knowledge of relevant issues and past solutions.

This brings me to another point related to the language of interactivity. Can we define types of interactive engagement? If so, how might we embed these activities into a narrative experience? In short, how can we invite the viewer to participate? Jean Piaget, a Swiss researcher, studied children's game play in the 1960's and was able to identify four discrete types of activity: practice, pretend, playing by the rules, and construction. These activities are often combined. For instance, hopscotch combines practice with a rule base, Doom combines practice with a goal-driven pretend experience. Today, the designation of the "twitch" or "reaction-timer" type of computer game perjuritively refers to a particular class of practice games. These games require manual dexterity and concentration at a task level; success is unambiguous. The shoot-em-up actions which dominate practice-based computer games today can not easily be replaced by complex-force models; to the extent that there are rules, they only weakly associated with a modeled world. The main impediment to transitioning to a richer content-base is the public itself. Publishers depend on sales; shoot-em-up games are startlingly popular, particularly for males in the 12- to 18-year-old range.

Pretend scenarios allow us to develop effective interpersonal skills. Stories teach us about temporal structures, moral frameworks, language, and communication. Children often create imaginary worlds to play in. In a "good" movie, the creators sculpt an experience which invites empathetic transference by the audience. A critical attribute of computer-based virtual reality environments is the immersive nature of the experience. Actions are executed by the viewer as a first-person character. This brings us to the subject of voice and interaction; interaction is intrinsically a first-person act. Last year, we created an interactive story, "Lurker." The story is played out over the network to and by an audience of six people. Prior to the start of each story, six people sign up as Lurkers. In the act of subscribing, our society of audience accepts a pretend role of being a Lurker. Lurkers assist the hackers. The hackers fall into tense times when one of their group disappears. The Lurkers are called in to assist. As part of your assistance mission, you must practice some difficult computer skills.

Games with rules constitute a particularly rich class. Many of us grew up playing action games, such as hopscotch and hide-and-seek, as well as board games such as Monopoly and Scrabble. Sports are also rule-based games which incorporate the practice of forming and executing strategies. In sports, the goal and the basic skills are known quantities; the challenge lies in how you accomplish this goal. Certain computer activities -- most notably MUDs -- combine rule-based game play with an important element of pretend or role-playing. In a MUD, you play at being a character; the pretend element is minimized when you play yourself, but posturing and selective self-revelation are commonplace pretend elements even then. As with practice games, rule-based games are often designed around winning and losing. MUDs suggest a new direction in that they focus on socializing.

Finally, interactivity is built into construction-based activities. For example, the activity of movie editing is constructionist by nature. For many years, Seymour Papert, who worked in the 1950's with Jean Piaget, has been actively engaged in embedding the activity of construction as a mode of learning in schools around the world. He champions what he calls "hands on and heads in" learning, which includes the social aspects of learning. This combination of active doing and social interaction effectively defines the potential of new media.