In a conversation, responsibility for the quality of the experience as a whole is divided between the participants. While it is possible to have an experience that is recognizably a conversation when one of the participant’s language skills are severely limited, it is difficult to consider such a dialog as a quality example of a conversation.


In a computer game, it has become standard practice to train the player in the mechanics of game-play as they advance through the experience. As player capability scales, game difficulty is designed to keep them immersed in what has been described as a “flow state”, where they are constantly feeling as though they are succeeding at a task that they perceive as being challenging. In order to sustain this experience, players are provided with opportunities to familiarize themselves with new skills before being called upon to apply those skills.

In improvisational theater, actors train and practice the difficult task of making and accepting dramatic offers designed to keep the scene moving forward. It is one of the most difficult tasks to perform in the theater, requiring extensive effort, creativity, and experience to maintain a dialog of even mediocre quality. Learning improvisational techniques requires taking many small steps, and is often supported through the use of exterior prompts in order to assist the performer and provide structure to the scene.

Dungeons and Dragons, when done well, has elements of all of the above mentioned examples. Good role playing involves a command of sophisticated rules and game mechanics, the ability to converse at a high level, and the creative agility that arises from experienced improvisation. In a tabletop role playing game, the quality of the experience is often directly proportionate to the degree to which the players in the group have mastered these skills. When members of the group fail to perform up to a high level, the overall quality of the game suffers.

In all of these examples, there are two common factors. First, each experience is implicitly interactive, requiring the participation of several individuals. Second, in each experience, all of the participants must achieve a level of fluency with a set of complex skills before the experience can be considered quality. The axiom here is :in an interactive experience, quality and satisfaction are often constituted by the fluency of all of the participants.

Similarly, it is hard to imagine an inexperienced reader getting the most out of a work like Moby Dick, or (for the sake of not appearing completely dated) Neal Stephenson’s Cryptonomicon. Enjoyment of the novel is a learned skill, the acquisition of which allows for a deeper enjoyment of the narrative form. Similarly, watching a film or reading a comic book involves the acquisition of a set interpretive skills, the honing of which leads to a deeper appreciation experience.

The same can be said, I believe, about any narrative form. The richness of the narrative experience is partially constituted by the quality of the narrative artifact, and partially by the fluency of the viewer. This is, in some ways, a re-stating of Zimmerman’s ideas about cognitive interaction, or of Eco’s ideas of the Open Work, or of Bakhtin’s ideas from Discourse in the Novel, or even some of Barthes’ ideas about the Death of the Author. In each of these works, the reader is considered partially if not wholly responsible for creating the meaning of the media object, much in the same way that a participant is responsible for the quality of the interactive experience. I would argue that in either an interactive experience, or a narrative one, responsibility for the overall quality of the experience is shared dialectically between participants.

Broadly speaking arguments around the role of narrative in games and interactive experiences can be split into two informal camps: Ludology and Narratology.

Ludology strenuously opposes any attempts at casting games as narrative artifacts. Narrative, it is said, is in opposition to play. “Play”, as a category of actions, is a hotly contested idea in it’s own right, however it is agreed upon that actions taken in the course of computer game “play” are motivated by a desire to win, or a desire to explore, more often then they are motivated by a desire to “perform a narrative role”. Play in games is characterized by a series of victory conditions such as “get the most points,” “escape the maze”, “kill the monster”, “acquire the treasure”, or “become more powerful then the other players”. Agency, the underlying principal which is often connected to ludic behavior, emphasizes momentary gratification. There is a sense that in order to provide players with unrestricted agency they ought to be free to do whatever they can imagine, without barriers, whenever they see fit. In catering to this principal…the idea that this type of complete agency is what players want…games are designed which provide little option to do anything other then freely explore within an environment with little structure and consequence, beyond the most rudimentary of puzzles and skills challenges. Such an environment is custom tailored to deny the possibility of dramatic narrative, and is often used as the basis for claims about the unsuitability of games and narrative to each-other.

In contrast, there are arguments from narratology that claim that games and play are essentially meaningless if they don’t tell some sort of story. Games, it is argued, are a valuable medium for exploring interactive storytelling. Narrative is impossible to avoid, according to narratologists, emerging from every situation we encounter, whether or not it has been encoded into the experience intentionally or not. The promise of interactive narrative lies in the somewhat nebulous idea that if a reader is in some way explicitly participating in the creation of the narrative, that their enjoyment of the story will be increased. This conception casts the reader/player in the role of co-author. The goals of a co-author, in contrast to the goals of play discussed above, are harder to quantify. They might be “tell a dramatically interesting story”, or “perform the role of the protagonist in a consistent and creative way”. There are very few examples of systems that implement this type of goal as a victory condition.

What we end up with when these two views collide is a profoundly dissatisfying experience. Succeeding at the goals of ludic play requires players to learn a sophisticated set of skills, and games are very good at training them in how to execute these skills, and at evaluating their competence. Succeeding at the goals of co-authoring, on the other hand, is not a task that we have even really asked players to perform. Instead, there is an assumption that it is possible to create a meaningful narrative experience that can adapt to player actions as they go about the task of performing ludic goals. Drama management systems such as those found in OZ, Facade and Anchorhead implement planning algorithms that forecast every possible event configuration whenever a key plot point is reached, evaluate the potential outcomes based upon a set of quality criteria, and then select the event that is most likely going to yield a quality narrative, as understood by the system. Player actions are under constant observation, as a variable, however the responsibility for telling a good story is considered to be within the domain of the system, which adapts to player choices on the fly.

This is analogous to a conversation in which one participant is attempting to solicit feedback on a complicated philosophical idea, and the other person is reading random words and phrases from a list. The first person can attempt to incorporate the words provided by the second into his speech, and the list can be designed to have bearing on the topic at hand, but the end result will not be quality conversation experience because there is a misalignment of goals and fluency between the participants. If we cannot imagine how this type of conversation could be mutually enjoyable, why do we think that this approach to interactive narrative has the potential to yield a satisfying story experience?
The problem lies with an unwillingness to properly divide up responsibility for the quality of the experience. Any experienced Game Master in D&D knows that no matter how skilled he might be at telling a story, the ultimate responsibility for a quality game is shared with the players. There is a cultural expectation, in a quality role playing game, that the players hold up their end of the social contract, and put the effort into performing. This is the idealized version of D&D admittedly; most games fall short of this mark, because the task of co-authoring a story is both difficult, and ambiguously defined, even in situations where every participant is (presumably) of human intelligence. In computational interactive narrative, there is no expectation that the player is going to do anything more then express their desire for unfettered agency in a spontaneous and unpredictable manner.

The question is: is this expectation a result of a desire among people to express their agency through spontaneous patternless action, or is this default behavior the result of years of designing systems to cater to this perception?  Is the desire for unfettered agency what we really want out of an interactive experience, or is it just what we’ve been trying to make out of interactive experiences?  Much effort goes into facilitating agency, and than training players in how to take advantage of all of the affordances of the system. Very little work has been done in facilitating dramatically meaningful interaction, and than training players in how to perform the role of protagonist, or co-author.

If the responsibility for the quality of an interactive experience is shared amongst participants, and the responsibility for the quality of a narrative experience is shared between the author and the reader, then responsibility for the quality of an interactive narrative experience must perforce be shared across the system designer and the player. As long as we continue to design interactive narrative artifacts around the assumption that the player’s role is simply as a problematizing source of error that must be corrected for, then we are denying half of the equation. Responsibility for creating enjoyable interactive narratives starts with the designer, who needs to learn to view players as collaborators, instead of opponents, however it ends with the player, who needs to engage in the project of interactive storytelling from an authorial perspective, rather then a ludic one. Until we begin designing with this relationship in mind, however, it will remain a source of tension which must be contended with.