Group Narration: Power, Information, and Play in Role Playing Games

When I say that I'm in a role-playing game, most people picture a stereotypical Dungeons and Dragons game: a wizard, a warrior, a barbarian and a thief rampage through an underground dungeon, killing monsters and looting treasure. But role-playing games can be a complex storytelling experience. In fact, there is much theory and criticism of role-playing games as narratives, though much of this discussion takes place outside of academia. In this paper I would like to address complex questions of authorship, narrative, and performative experience in relation to role-playing games, in order to expose possible further directions for research.

Role-playing games defined

The participants in role-playing games, or RPGs, are engaged in a complex process of group narrative; they are the authors, narrators, characters, actors, readers, and audience of a text that can be both experential and product-oriented. Studying RPGs as examples of group narrative will, I hope, be helpful to gamers, game designers, and people engaged in other forms of group narrative or collaborative writing.

Lisa Padol, in "Playing Stories, Telling Games: Collaborative Storytelling in Role-Playing Games", defines a role playing game as follows: "A role-playing game is composed of one or more sessions where a group of gamers gets together in one place to play (par. 2). This "make believe with rules" is played by 2 or more players, and the rules vary from simple to sophisticated.

Most of the game's players have one character, called a player-character, or PC. There is usually one person in charge, referred to as the GM or Gamemaster; some systems use the terms Dungeonmaster, Narrator, or Storyteller. This GM is often referred to as "running the game". She plays the part of all the characters not played by the other players; these GM characters are usually called the NPCs or non-player characters, despite the fact that the GM is also a player in the game. PCs and some NPCs have character sheets - documents that describe the characters' backgrounds, personality traits, and abilities (Figure 1).

Figure 1: The Onion's parody of a RPG character sheet.

Many role-playing games are set in a variant of the fantasy genre, with magic, magical creatures, swords, and some form of feudal society predominating. Dungeons and Dragons, Ars Magica, Lord of the Rings, and Trollbabe are examples of fantasy genre games. Other game settings vary widely. There are settings that are futuristic, or historical; there are worlds based on a specific book or narrative series; there are settings based in alternate histories and realistic modern settings. There are also general or generic game systems designed to be used in any setting that the GM wishes to create.

The game system is a system of rules for character creation and interaction, published online or in a book. These game books often include background material, sample characters, examples of play, and suggestions for plots or adventures. It is used as a reference by players and GM. In this paper, Ars Magica, underlined, refers to a book. Ars Magica, not underlined, refers to the game system.

In the game, the players and GM sit and talk about what the characters are doing. They speak in character or out of character. In character, players and GM make their characters speak to each other directly like characters in a play or novel; they also can narrate their character actions in first or third person. It is similar to an improvisational performance or a musicians' jam session, and it usually lasts for several hours. I will refer to this group experience as the game session.

Besides the game system, ther references for participants in the game could include history books, novels such as Tolkien's Lord of the Rings series, and material created by the GM and players. I will refer to these references as source texts.

Trends in RPG Theory

Many lively discussions of role-playing game theory have centered around modeling and classifying game play into three styles. They involve questions of plot, character, story genre, and power balance between players and GM. Current RPG theorists tend towards looking at various styles of play as useful for accomplishing different end goals, rather than valuing one style over another. This end goal is sometimes assumed to be the making the game sessions enjoyable for all the game's participants, but other end goals, such as the literary quality of the story told, are posited.

Figure 2: Models of Styles of Play in Role-Playing Game Theory

Online forum participants such as Brian Gleichman have questioned the usefulness of these theories to game design or game play. Gleichman contends that such models provoke endless discussions and flame wars, as gamers attempt to shoehorn each other into one of the three categories, or as they passionately argue for the merits of one style over another (par. 2-4).

Defining Collaborative Writing and Authorship

Other RPG theorists such as Lisa Padol and Dariel Quiogue have focused on the game as a collaborative text or textual experience. The question of authorship is central to these discussions of collaborative writing and storytelling. Who are the authors? Who has authority over what aspects of the story and of the game? It is helpful to look beyond narrative theory to models described and proposed by collaboration theorists.

In Singular Texts/Plural Authors, Lisa Ede and Andrea Lunsford define collaborative or group writing as "any writing done in collaboration with one or more persons." Writing is defined as "any of the activies that lead to a completed written document. These activities include written and spoken brainstorming, outlining, note-taking, organizational planning, drafting, revising, and editing" (14).

Ede and Lunsford also point out that the identity of "the author" of a text is particularly problematic in collaborations. They see textual production as a group process, and discuss problems of power and authority, hierarchy and consensus, in the construction of a textual product. They interviewed various groups of collaborative writers, all of whom write non-fiction.

Chris Holstrom, in his 2002 thesis project, posits several different systems for group authorship and authority over a collaborative narrative text. Holstrom built a system for online group authorship called "The Mahoney Project." So that each author gets credit for his or her work, Holstrom set up an authoring system based on atomization. Each author's contribution is an "atom", and credit for each atom, or post, is attributed to an individual. Mahoney Project writers can edit their own posts, but not the posts of other writers (78-85). Holstrom retained editorial control over the entire project, however, in order to edit out anything he considered pornographic (67-75). His goal seemed to be to allow a large number of authors to participate in story construction with a minimum of any editorial control. Unfortunately, the story has not progressed beyond around 10 story atoms or nodes.

In his essay "Role-playing Games as Collaborative Fiction", Dariel Quiogue declares that considering RPGs as collaborative fiction allows for the design and playing of better games; he suggests that the GM see herself as the editor, not the author, of a collaborative text in which she shares authorship with the other players (sect. 4). Quiogue also points out that although many game systems position the GM as sole author of the game's plot, the players have significant influence over the plot; Quiogue thinks this player influence should be acknowledged and encouraged.

Quiogue and Padol both advocate that GMs consciously give up some measure of power and authority to players, but Padol's focus is on defining author, text, and audience. She defines the authors of a role-playing game text as the GM and players together, and perhaps also the authors of game supplement books used in the game session (Padol, par. 8-10).

A GM often claims authorship or ownership of "his game"; he regards the background ideas, plans, notes, and the overall concept of the game as his own property. The playing of the game by that GM and a group of players is referred to as the GM "running a campaign". The GM could take his game and run it again with a new group of players, and a different campaign or game session would result.

This leads me to think that there are at least two issues in attributing the title of "author" to a writer or game player: who gets credit for what, and who controls what. Within the collaborative narrative of a game, who has authority to decide what happens or what is "true"? For example, as a player in a game session, I could at any time say the words "I see a dragon, and I waste him with my crossbow." But though I say it, do I actually do it within the world of the game?

In most game systems, I as a player would have no authority to declare the existence of a dragon, unless my character has a magical ability to create or summon dragons; the encountering of a creature or NPC is in the provence of the GM. My character sheet may not include a crossbow in the list of equipment the character normally carries. The game system or setting may preclude the existence of dragons: perhaps it's a futurististic game system, set on a spaceship far from earth, where robots are more likely than flesh and blood dragons. If the dragon is allowed to be there, do I "waste" it? In most game systems, the outcome of a conflict is determined by rigorous application of the game system rules.

Players, though, have control over the actions of their characters; what their characters, the story's protagonists, decide to say and do largely controls the direction of the story. The GM's assertion of authority and control varies widely according to her philosophy, the assertiveness of the players, the game system, and the structure of the game itself.

What would be likely to happen in this example, given the dragon's existence: I attempt to waste him with my crossbow, but I depend upon the game system rules, the abilities on my character sheet, the combat abilities of the dragon, some rolls of the dice, and the GM's decision, to see if the dragon is indeed vanquished. In the meantime, other players may intervene and interfere with my declared actions. In a game that uses fate cards, I or another player might use a card to lend authority to the dragon's existence or non-existence. In short, we can see that in a role-playing game session, there is no one author for a given part of the story, whether we consider that part of the story as a speech act, an utterance, a node or atom, or an act of writing/creation; in fact, the range of authorship influences depart from human agency. "What is the author?" becomes a question as valid as "Who is the author?"

Figure 3. Authorship: who or what determines "what is true" in an RPG session?

Action type	Specific instance
Speech acts	Player declarations GM declarations
Meta-game tools	Dice rolls Whimsy cards Plot points
Source texts	Character sheets Rule systems World backgrounds

The game book authors, the world builders and rule makers, are authors writing part of a story and inviting other authors to collaborate. The existence of these fictional worlds, written without their stories, are an invitation for stories to happen. Every instance of a Dungeons and Dragons or Ars Magica game shares a fair amount of authorship with the writers of the game system, rule book, or module authors.

Rather than continue to try to define the identity of the author or multiple authors, I'd like to try to re-think authorship by looking at how power and information are distributed among players, GM, and source texts. How are authorship and authority structured?

Dialogic and Hierarchical Modes of Collaboration

Ede and Lunsford point out that most of the collaborations they studied depend on a rigidly structured hierarchy, which results in high efficiency in producing a final textual result. According to Ede and Lunsford, in collaboration that is focused on productivity and efficiency, "the realities of multiple voices and shifiting authority are seen as difficulties to be overcome or resolved." They associate this hierarchical structure in part with male gender, calling it "a masculine mode of discourse."

Ede and Lunsford assert the existence of an alternate method of collaborative writing which exemplifies Bakhtin's concept of the dialogic; in dialogic mode, the group is loosely structured, authority and goals are fluid, and the process or experience of writing and collaboration is valued over the result, end goal, or textual product. Ede and Lunsford think of this mode as predominantly feminine (133).

Many gamers describe this split in game playing philosophies as gendered. Knights of the Dinner Table's one female character, Sara, consistently emphasizes character, storytelling, and non-violence in her play. Hong Ooi, in a humorous article "Real Men Don't Play GURPS", contrasts the Real Man with his opposite, the feminine or feminized "Quiche Eater":

The easiest way to tell a Real Man from the rest of the roleplaying crowd is by the game he plays. Real Men play Dungeons and Dragons. Quiche Eaters play GURPS and Storyteller. Mark Rein*Hagen, the designer of Storyteller, was once asked, "How do you pronounce the dot in your name?" He replied, "It's unpronounceable, and symbolises how meaningless are the labels that we attach to ourselves." One can tell immediately from this comment that Mark Rein*Hagen is a Quiche Eater. Real Men don't need the abstract concepts introduced by Quiche-Eating games - like characterisation, immersiveness or realism - to get their jobs done.

Women that I've known online have been much less competitive. It's not about beating up someone else, it's about the story. The process by which a goal is reached, not the goal itself. I think we have more fun with the telling than with beating an enemy's head in. (Elizabeth B.)

Figure 4. A stereotypically gendered view of the dialogic and hierarchical modes of discourse.

Masculine	Feminine
monologic goal-oriented centripetal violent wars authority competition conservative	dialogic process-oriented centrifugal verbally diplomatic relationships consensus cooperation open to change

While I agree the dialogic mode might be associated with feminism, I am wary of assigning the dialogic to part of an essentialist concept of the feminine. Instead I would suggest that the deeply subversive qualities that Ede and Lunsford assign to the dialogic mode of collaboration make it a useful tool for any marginalized group of people; it might not be something to assign to the feminine, to women, or to feminists, but rather to any group who perceives themselves as disempowered and who wishes to be subversive. In other words, the dialogic is not a feminine mode of storytelling or play, but it can be a useful tool for feminists.

A further exploration of feminism and gender in gaming is beyond the scope of this paper, but I felt that I should at least suggest that it would be interesting to apply feminist theory to gaming. The subculture of gaming is male-dominated and sexism is an issue for many players and game designers; despite (or because of) the importance of the topic, it is very difficult to discuss in online game theory forums.

Gender aside, there is (or has been) a split in gaming theory between hierarchical and dialogic authoring philosophies. Applied to role-playing games, Ede and Lunsford's view of collaborative writing modes sheds some interesting light on differing philosophies of game play.

Hierarchically structured RPGs

Some RPG systems encourage a deep division of power and authority between GM and player. Although there are clearly several people collaborating in order to produce the game session, the GM is considered to be the author, just as in Bakhtin's concept of the monologic unitary speaker, the language of a text seems to come from a single unified source, even if it actually comes from several speakers (270-272).

Gygax's Dungeons and Dragons, which is thought of as the first role-playing game system, gives the GM final authority on all aspects of the game, and does not emphasize the game as story. Gygax refers to the GM as "the creator and ultimate authority" (7). In the first edition of the Advanced Dungeons and Dragons Dungeon Master Guide, there is very little instruction for the GM on how to handle the process of storytelling. Two hundred pages of densely packed information outlines world background and mechanics, lists of spells, and tables of numbers for determining the results of combat. There are several examples of play scattered throughout the book.

During one of its few meta-level discussions of GM-player interaction, the Dungeon Master Guide advises that the GM assign penalties to players/characters if they mix up out of character and in character knowledge. The book humorously conflates player and character, suggesting that the GM "cast a lightning bolt on the offending player's head" (Gygax, 110).

Several structural elements in Dungeons and Dragons weight the system towards a hierarchical distribution of power. Descriptions in the game book of GM and players struggling for control of the story create their own meta-story of authority, resistance to authority, and the suppression of that resistance. The book itself is intended to be read only by GMs. Devices such as the cardboard "dungeonmaster screen" explicitly divide GM from players, allowing the GM to hide dice rolls, maps, and notes behind a physical shield.

Mechanisms of rewards dispensed by the GM also contribute to the GM's power. For example, The Lord of the Rings Role-playing Game exhorts the GM to be sensitive to player interests when planning the game. Yet the GM is also told to set secret scenario goals before each game session, and awards experience points to the players, who can then use the points to give extra skills to their characters. The amount of points given is based on how close the players came to achieving goals which are only known by the GM. In this system, if the players choose for their characters not to follow the plot outline imagined by the GM, they are doing something wrong.

Dungeons and Dragons games do not have a single text as the goal or product of their game, but there is an assumption that strong GM control of rules, plot, and story improves the end product of the game, whether that end product is simulation, narration, or "a good game". Gygax describes the goal of his game system as "a superior campaign, a campaign which offers the most interesting play possibilities to the greatest number of participants for the longest period of time possible" (7).

In writing this paper, I have found it difficult not to favor dialogic discourse over monologic, but I recognize that strong authority can be useful and good in the production of stories, and agree with Joshua Macy when he contends that rules and authority have a complex function inside and outside the story:

"...this forces you to experience the fictive world in much the same way that you actually experience our own. You make decisions based on your understanding of the world, and then you wait to see what are the consequences. The world is fictive, but the decision, and the anticipation during that period of uncertainty are real. So too are the emotions generated by success or failure" (Macy).

In my experience, when player expectations and GM expectations of this end goal are similar, strong GM authority and a hierarchically structured system can result in "a good game".

Yet the archetype of the GM as beleagured, lonely authority struggling against enemy players runs deep in gaming communities. In a story in the comic book Knights of the Dinner Table, a series about a group of gamers, the game's players directly challenge the GM's authority. This challenge to authority de-rails play and destroys the game experience (Blackburn). The issue is dealt with humorously, but the comic strip game's GM often views the players in some sense as "the enemy" and attempts by players to control aspects of the story, or to communicate with each other out of character, are described as revolt, revolution, or "anarchy" (see figure 5).

Figure 5. A challenge to GM authority (Blackburn).

Based on my own experience in and out of games, I would say that reminding a person in power to "be sensitive" is not likely to result in empowerment for their subordinates. Instead I advocate the creation of methods and structures that facilitate the sharing of power and information.

Structures that encourage a dialogic mode of play

Some game systems, such as Jonathan Tweet and Mark Rein*Hagen's Ars Magica, encourage that power and authority be as evenly distributed as possible. To this end, Ars Magica holds out a vision of the leveling of hierarchies by rotating positions of power. The system provides other narrative structural devices, such as Whimsy cards, intended to share authority between GM and players .

Ars Magica encourages game participants to take turns playing in the GM or "storyguide" position. It also encourages players to create more than one character; at least one powerful magus, a non-magic using companion, who has less status than a wizard, and a grog, or henchman. Because the characters vary widely in power and ability, players experience varying degrees of power to affect the direction of the game.

In Ars Magica, Tweet and Rein*Hagen spend many pages discussing the terminology and structure of storytelling. They include narrative theory as something important to game designers and players. For example:

"Ars Magica is meant to be played by a group numbering anywhere from three to ten or more. This associate of players is called a troupe. Unlike many other roleplayiung games, in Ars Magica, each troupe member controls a number of distinct characters, not just one. In most Ars Magica troupes, different players take turns leading the story, that is, setting out the plot, the action, and playing the roles that are not handled by other playears. The person who is leading the story at any given time is called the storyguide. Usually the storyguide has created a long-term plot against which individual adventures are set. This overarching storyline is called the saga, and the shorter adventures within it are stories." (12-13)

Each player has several characters of varying degrees of power: a magus, who is a powerful magician; a companion, who is a skilled yet non-magical character; a grog, who is a lower-class henchman with no magic. Social and personal inequalities are built into the game system. This is also true of The Buffy the Vampire Slayer Role-Playing Game; in a Buffy game, the scoobies (players without magical powers) are allocated extra drama points which they can use to influence the outcome of conflict.

Some games provide rule structure or game mechanics, for non-traditional player interventions; for example, during a conflict in a Trollbabe game, a player can declare the existence of a sudden ally, a handy geographical feature, or a carried object. These items can be declared to exist in retrospect; if a dragon has breathed fire on a character, that character can attempt to "suddenly remember" that she's wearing her fireproof armor. The Trollbabe system also encourages players to embellish scenes at which their characters are not present, with editorial power in the hands of the person narrating the outcome of a scene (Edwards 16-17).

With Whimsy, tarot, or fate cards, players can affect outcomes in a way that they can't merely by controlling their own character. For example, in a game I was in recently, two of the other players played cards on a situation involving my character, Kjartan. Heather specified that her "Horrible Failure" card should foil Kjartan's attempt to placate his uncle Poul; Jim played a "Personality Clash" card to further infuriate Poul. (Kim, Session Log). Without the cards, there would have been no structured way for Heather and Jim as players to interfere with the actions of either my character or Bill's character, Poul. The resulting fist fight between Kjartan and Poul worked well in the story.

In many game systems, or under some GMs, it would have been considered out of place for these two players to interfere in a situation between two other players' characters. It is often referred to in a somewhat negative way as crossing a line between player knowledge, outside the story, and character knowledge, inside the story; in short, cheating.

With MUSHes I have encountered, there are authority structures in place to direct and edit the collaborative storytelling. As in most RPGs, in MUSHes there is no goal to create a final textual product. The experience of play is the goal, although players often log their sessions, as Cari McAskill has for her character Catharine on the Heart of Darkness MUSH (McCaskill, Catharine Russell).

World of Darkness is a game which is set in the Belgian Congo of the 1880s. The MUSH's web site describes the world, points the reader or player to source texts such as Joseph Conrad's Heart of Darkness, maps, dictionaries of African names, and histories of Africa (Heart of Darkness). McAskill writes on the nature of MUSH storytelling:

To some people it's a very ephemeral form of entertainment; it's like jazz improvisation. That's actually a very good analogy--with, say, written drama being like classical music, where there's a script to follow. MUSHING falls into the "improvised narrative" school, like a commedia dell'arte without any division between players and audience. We are the audience.
(McAskill, Interview).

The structures used by MUSHes, with wizards and committees to build areas of a game world, shifting GMs for scenes or storylines, and flexibility to new players, was used very successfully by Davenford, a large ongoing role-playing game collaboration in Los Angeles (DeSanto).

Heart of Darkness MUSH, unlike Holstrom's Mahoney Project, requires participants to apply and submit writing samples, which are reviewed by an editorial board comprised of "wizards" from the MUSH. The application guidelines, written by Mark Allen and Sean Ware, emphasize three main attributes a writer should have. The first is the importance of corrrect spelling, punctuation, and grammar; Allen and Ware write, "one of the critical ways to consistently achieve a very high level of quality role playing is to place exacting emphasis on proper English grammar, correct spelling and correct punctuation. These fundamental tasks form the building blocks of all that occurs and is perceived in a MUSH environment ". The second characteristic the editors look for is players who write about their character in a compelling, world-consistent way. To this end, players must learn a fairly large amount of information, mostly alternate history of the "World of Darkness" gaming system, and real-world history of the 1880s and colonial Africa. Another quality good MUSHing should exhibit: "internal character motivations" - what makes the character tick? (Allen and Ware).

There are RPGs that emphasize player willingness to learn background or history beyond that contained in a game system rulebook or sourcebook. Shadows in the Fog is especially noteable for its direct demand that players work hard and learn as part of the game; they must play by using their knowledge of the history and literature of 1880s London. This game emphasizes player knowledge of source texts and of mythological image systems such as tarot cards over rule systems or dice roll mechanics (Lehrich 5-6).

There are also software tools for group authoring. Version control programs that track changes and authorship are common in business and in software development, but not in the production of narratives. WikiWikis provide an interesting tool for exploring group authorship; the StoryTellingWiki is a good example of current exploration in this field.

Another useful area to investigate would be a further relation of role-playing games to MUDs, MUSHes, and interactive fiction or hypertext, using Aarseth's theories of cybertext and ergodic literature. I have only read one chapter of Aarseth's book on Cybertext, but role-playing games seem to fall into his definition of ergodic literature (Aarseth ch. 1).

In any of these forms of decentralized collaboration, the emphasis on improvisation and shared authority translates into high expectations for all participants. Just as jazz musicians must have a basic skill with their instruments, and should know music theory, group storytellers should make an effort to study and learn the basic elements of stories. Storytelling improvisation places a high demand on players to be sophisticated story creators, but also sophisticated story consumers.

Labyrinthical maps, stories, and plot structures: Unicursal and multicursal

In this improvisational style of role-playing game, GMs sometimes struggle with the issue of how to prepare for the game. "In our first GM-ing attempts, most of us made the mistake of creating an adventure with fixed "plot points" which the player characters had to reach and solve before they could advance to the next plot point, solve that, and so on until the ending point is reached" (Quiogue, Synergistic GMing par. 11).

Quiogue advises against creating a unicursal plot structure, because if players don't "choose the right path", the GM's preparation will no longer be useful. The plot won't go the way the GM expected, and she might not be ready to improvise all the details necessary to make the game go smoothly. If the other players participate in creating characters with depth to their personality, and world background that allows for improvisation, and if everyone in the game loosens up, a good story will follow. Quiogue describes his ideal GM laying out possible plots as a network of railroad tracks where plot hooks and decision points are the switches. This suggests that he would like to see GMs plotting stories as flow charts or decision trees, but that they should be ready to improvise when players leave those ready-laid tracks.

These issues of plot structure are also relevant to hypertext theory and computer game design. A MUD builds a multicursal map or world architecture, but this should not be confused with a multicursal plot; a physical garden of forking paths is a map which does not result in a story. Wandering around is not necessarily a story.

People just learning about MUDs, for example, often confuse the multicursal nature of the story space with the existence of an excitingly multicursal story. Confusing these two aspects of story architecture can lead to problems that result in reader/audience/player frustration. For example, in some Infocomm text adventure games like Zork or Plundered Hearts, the player is free to wander about in the interconnected rooms or areas of the game world, picking up objects, pulling levers and pushing buttons. Rooms could be visited in any order not limited by the interconnected map, and the player could solve various puzzles posed by the game. Yet if certain actions were not done, certain puzzles not solved in an order not at all obvious from cues internal to the game, then the player would be unable to solve the game. To a player faced with this situation after hours of game play, this can be quite frustrating. The unicursal labyrinth of plot of Zork or Plundered Hearts did not coincide with the multicursal labyrinth of the game world's map.

Authors of plot-based video games as well as hypertext authors and roleplaying game GMs would benefit from careful consideration of these issues.

Fig. 6. A territorial labyrinth and a geneological labyrinth.

Brygjafael Family Tree Tjaraholt Family Tree

Intertext

Levels of discourse

Bakhtin suggests that mixing of levels of discourse blurs the boundaries between authorial speech and the speech of others. (Bakhtin, 319-321). In a role-playing game, different levels of discourse are still relevant to the story, even if they take place outside the story.

The author of a book often has information about the book's possible future readers, or might have direct feedback from past readers; this information can influence that author's future writing. Readers, in turn, often have information about an author; their biographies, interviews, letters, or other books; this can influence the way a story is read. Information outside the text and outside the story told by the text nevertheless influences the reading and writing.

In a role playing game, the players' knowledge of each other outside of the story can contribute to the evolution of the story. For instance, in the Vinland game, I know a little bit about a player, Bill, and his real life relationship with his father. Knowing this, I can manuver my character to let the story play with concepts of father-son relationships. Similarly, the GM, John, knows about my real life ambitions and conflicts around being a poet, and is able to bring them into the story, so we explore how our Norse Vinlanders saw poets and poetry.

Part of the enjoyment of the story for some of us has been in private "meta-game" jokes. The game is set in John's hometown of Piermont on the Hudson River in New York; the Althing or Norse parliament meets in Manhattan Island; storylines have taken us to my and my sister's home in Narragansett and to Boston and MIT where Heather went to college. It is amusing to run across a familiar landmark of New York or Boston, translated subtly into 1392 Vinland.

A shared story experience is enhanced by out-of-story personal knowledge; consider the difference between watching a movie "alone" in a crowded theater full of strangers, versus watching that same movie with a good friend. The experience of the story is quite different.

Another common out of character interaction is to quote movies or television shows; often this functions as comic relief or ironic commentary. For instance Bill's character Poul might be trying to revive a fallen comrade on the battlefield, and when he fails, Bill might say "He's dead, Jim." a reference to "Star Trek" that would make us all laugh. Or when two characters seem at odds, a player might hum the "Dueling Banjos" theme from the movie "Deliverance"; a suggestion that a serious feud may be developing, a suggesting maybe of disturbing violence or homoerotic undertones to the enimity. In a game set in 1392, movie references have no place, but to the 21st century players, they are a convenient shared cultural reference.

Some gaming groups and game systems discourage the blurring of levels of discourse, and some theorists and game systems consider it an important point to address in play and in design.

John Kim, in his essay "Techniques for In-Genre Planning", asserts that it's important for GMs to be aware of the out of character personalities and capabilities of their players. For example, a group of characters might logically require a certain character to act as a leader, such as a military commander in a war; however, that character's player may be an inept leader, or may not want to be in a leadership role. Kim reminds GMs to "...remember the leadership is primarily a meta-game issue about the players making decisions, not just an issue of the characters. When deciding on a leader or diagnosing problems, look at the qualities of the players in question. If a player does not make an effective leader, but plays a leader PC, then you need to enlist the other players in supporting that" (Kim, Techniques sect. 6).

Lisa Padol further explores issues of textual boundaries and interactions between game participants. She defines the game as follows:

Anything which reaches the interface between the GM and the players is part of the text. Anything which does not reach the interface, and, therefore, does not affect both the GM and the players, is not part of the text... the text is defined by the interface between or interaction of at least two gamers. This interaction must be directly related to the game...also, information which does not reach the interface, which is known by only one of the gamers, is not part of the text. A GM will almost always have such information. This information may shape the text, but is not part of the text unless and until it reaches the interface and is transmitted to at least one other gamer" (Padol 2-3).

The game session is both a story and a performance and as such, is the ephemeral text that Padol proposes.

There are other ways that game play can cross story/real life boundaries:

• players use characters to explore aspects of their own personality
• current events or issues, such as war, become central to the story
• players' real life relationships influence events or understanding of the story
• players get to know each other better in real life through analysis of each others' characters
• players use characters who are unlike themselves to explore different identities or ethical systems
• players use a game to deepen their understanding of a particular history or culture
• players game in a fictional world such as Tolkien's, and deconstruct it

I've already turned away from the concept of the author; I propose now to look past the idea of a sharp-boundaried "text" for a moment to consider the narrative as a system. The game's boundaries are not sharp; text and story bleed into real life in varying degrees, which can be modeled as greater or lesser information flow. The text has fuzzy boundaries. If we accept this, then it also follows that the world of fiction and fantasy has a fuzzy boundary with the "real world" we live in.

Information Flow

When Padol asks, "What is the text of a role-playing session?... A supplement is not the text of the sessionl Neither is the text of the rules of the role-playing game, although one might argue that the rules constitute a meta-text". (par. 11-14)

Padol reads the experience of the role-playing game session much like she might view an individual and unique performance of a play, where the script, stage directions, and the work of rehearsals are not "the text". The players and GM both have authorial status in relation to the experential text of the game. This idea is useful for understanding the importance of improvisation and experential play, and also gives the critic a clearly defined text for close analysis; however, Padol's session-focused view of the game is very useful in limiting the scope of discussion of a game as a narrative, but this view also limits the understanding of the game as a dynamic system.

In my view, source texts, new texts created by players and GM, and game session are all part of the narrative system that makes the story. Texts, rules, dice, cards, characters, GM, and players are all agents who have varying degrees of authority to create the story.

To some degree, GM status does not determine authority. A player might gain authority in the eyes of other players by knowing arcane details of the rule system, and manipulating these rules to advantage. Alternatively, if a player knows details about the history, background, or setting of the game world, that player could work that knowledge into the story. For example, if I as a player in a game set in medieval Iceland happen to understand how inheritance works in Icelandic law, I can share this knowledge with the group or the GM, and maneuver so that my character will inherit substantial property. This level of player contribution is encouraged in game systems such as Christopher I. Lehrich's Shadows in the Fog. Background knowledge can also be rewarded more overtly; for example, a player who writes a diary entry for their character, or makes a map, or invents an entire language, could be rewarded with experience points or in-game perks such as wealth or rare items, so that the economy of knowledge is integrated more closely with the game's other economies.

Remembering the events and people of past games in a particular campaign is also important to authorial power. Notes, session logs, family trees, casts of characters, and timelines that are accessible to all a game's players provide a base of information for building future story events.

In an information system, rather than defining a "text", I would ask the question, "Who can read information, and from where? Who can write information?" Various textual sources and products then can be viewed as databases; repositories of information that have varying degrees of read/write access for various people.

Figure 7. Two possible ways to model information flow in collaborative writing.

In the first model, I consider the game as collaborative writing (Fig. 7). Human agents are represented by circles. A, B, and C are players or authors, who can both read and write a working text, drawn as a cylindrical database, that is also readable and writeable by an editor, or GM. The GM and players also have two-way communication between each other. This model can apply to role playing games, but could equally well be applied to book production in which one or more authors generate a text, which is then edited and published. Just as the degree of editorial/original authorial control varies, the balance of power between GM and players can vary widely.

The second model is a different view of games, combining Padol's ideas with my own (Fig. 7). The experience of the game - Padol's definition of the game's text - is readable and writeable by GM and players. Players can also see some of the game system information and other source texts; this degree of indeterminacy is indicated by the dotted line. The GM or editor has read/write access to all source texts. In this system, we can also add final texts generated by a player, multiple players, or by the GM.

Preserving and organizing texts generated by game play and making these texts accessible is especially important for ongoing games that are long term. Computer software and the Internet facilitate complex structures of read/write access, which could open interesting possibilities for information sharing; for example, character sheets could include evolving "point of view" or "gossip" texts, so that players could define and describe what their characters think about other characters or NPCs. Players could make this information available only to players whose characters should know it in-game.

The third diagram models the information of the game as filtered through various layers of textual access (Fig. 8).

Figure 8. Layers of information access in collaborative writing and role-playing games.

By analyzing a particular campaign with this model, it could be made clear to players and GM exactly which parts of the story players are allowed to influence, and which fall under the control of the GM. It can help a collaboration to clarify the ways in which multiple GMs and players can be allocated their own areas of the game world to research and build.

Diagramming information flow exposes some of the ways information access is important to power, authority, and authorship in a game or in other narrative systems. When we apply an information flow diagram to the process of writing a book, it helps to expose the extent to which writing is a process that involves multiple people; writing is very rarely an activity carried out in isolation. Reading, too, is not an activity with definite boundaries, but is part of larger systems and feedback loops between individuals - readers and authors, audiences, performers, and players. Through looking at role-playing games as narratives, it is possible to create theory which is applicable to narratives in general.

Works Cited

Aarseth, Espen J. Introduction. Cybertext: Perspectives on Ergodic Literature. Baltimore, MD: Johns Hopkins University Press, 1997. 7 May 2003 http://www.hf.uib.no/cybertext/Ergodic.html .

Allen, Mark, and Sean Ware. "MUSH Application Guidelines". 1998. Heart of Darkness MUSH. April 2003 http://www.erinyes.org/hod/app-guide.html.

Astinus. "A History of Role-Playing." 1998. Places to Go, People to Be. May 2003 http://ptgptb.org/0001/history1.html .

Bakhtin, Mikhail. "Discourse in the Novel." The Dialogic Imagination. Ed. Michael Holquist, trans. Michael Holquist and Caryl Emerson. Austin, TX: University of Texas Press, 1981.

"Bill Gates Grants Self 18 Dexterity, 20 Charisma." The Onion. 18 June 1997. May 2003 http://www.theonion.com/onion3121/billgates.html .

Blackburn, Jolly R. "The Great Revolt." Knights of the Dinner Table Vol. 1. 2nd ed. Kenzer & Company, June 1998.

Briggs, Amy. Plundered Hearts. Interactive fiction. Infocom, 1987.

Chokey, Jim, Heather Downs, Laura Henry, Liz Henry, Bill Humphries, and John Kim. "Vikings & Skraelings" role-playing game campaign, 30th session. Transcript. 17 Apr. 2003.

DeSanto, Mike. "Campaigning at the Con: Getting it Organized." Nov. 2002. RPGnet. March 2003 http://www.rpg.net/news+reviews/columns/campaigning28nov02.html .

Ede, Lisa and Andrea Lunsford. Singular Texts/Plural Authors. Carbondale: Southern Illinois University Press, 1990.

Edwards, Ron. "GNS and Other Matters of Role-playing Theory." 2001. The Forge. 2 May 2003 http://www.indie-rpgs.com/articles/1/ .

---. Trollbabe. Adept Press, 2002.

Elizabeth B. "Do Women Game Differently Than Men?". Post to online discussion. 24 Jun. 2001. May 2003 http://www.geocities.com/pjandjenn/hazediffer.html .

The Forge Independent Game Forums. May 2003 http://www.indie-rpgs.com .

Gleichman, Brian. "Layers of Design". RPGnet. 11 Dec. 2002. May 7 2003 http://www.rpg.net/news+reviews/columns/elements11dec02.html .

---. "Should There Be a Threefold?" Date unknown. 6 May 2003 http://home.attbi.com/~b.gleichman/Theory/Threefold/Should.htm .

Gygax, Gary. Advanced Dungeons & Dragons: Dungeon Masters Guide. Lake Geneva, WI: TSR Games, 1979.

Heart of Darkness MUSH. Home page. 5 Sep. 2002. Apr. 2003 http://www.erinyes.org/hod/.

Holstrom, Christopher Ryan.Ê "Storytelling as Collaboration and Community: Designing an Interactive Multi-Author Environment for Hypertext Fiction." Thesis. 2002. University of Washington. May 2003 http://www.cholstro.net/mahoney/table.shtml .

Kauth, Dennis, and Rob Lazzaretti. "Dungeon Catacomb or Sewer." Map-a-Week. 8 May 2003. Wizards of the Coast. 9 May 2003 http://www.wizards.com/dnd/images/mapofweek/02catswerdng_72dpi_bni35t.jpg .

Kim, John H. "Techniques for In-Genre Planning." Feb. 2003. May 2003 http://www.darkshire.net/~jhkim/rpg/styles/genre/planning.html .

---. Personal Interview. 24 Apr. 2003.

---. "The rec.games.frp.advocacy FAQ." 1997. Apr. 2003 http://www.darkshire.net/~jhkim/rpg/styles/faq_v1/faq1.art .

---. "Serpents in the Cave". Vikings and Skraelings Role-Playing Campaign Session Logs. 28 Apr. 2003. 7 May 2003 http://www.darkshire.net/~jhkim/rpg/vinland/campaign/sessions/session34.html .

Lebling, Dave, and Marc Blank. Zork I. Interactive fiction. Infocom, 1979.

Lehrich, Christopher I. Shadows in the Fog: Occult Roleplaying in Jack the Ripper's London . Jan. 2003. Advance copy. May 2003 http://auroragames.com/pdf/shadows.pdf.

Long, Stephen S. and John Rateliff, Christian Moore, and Matt Forbeck. The Lord of the Rings Role-Playing Game . Los Angeles, CA: Decipher, 2002.

Macy, Joshua G. "On Roleplaying Games." 1999. Webamused. May 2003. http://www.webamused.com/onrpgs.html .

McAskill, Cari. "Miss Catharine Russell". 20 Apr. 2003 http://www.darkshire.org/~cari/logs/catherine/ .

----. "Questions about GarouMUSH." Email to the author. 25 Apr. 2003.

Ooi, Hong. "Real Men Don't Play GURPS," 2002. 8 May 2003 http://www.zipworld.com.au/~hong/dnd/realmen.htm .

Padol, Lisa. "Playing Stories, Telling Games: Collaborative Storytelling in Role-Playing Games." RECAP: Publications, Inc.. 1996. http://www.recappub.com/games.html . Apr. 2003.

Quiogue, Dariel. "Role-Playing Games as Collaborative Fiction." 1998. AEGISweb. Apr. 2003 http://www.geocities.com/aegisweb/9-Articles/91-CollFic.html.

---. "Synergistic GMing." 2000. The Vaults of Atreon. Apr. 2003 http://www.geocities.com/dquiogue/musings/synergy.txt.

StoryTellingWiki. "WhatAreWeDoing: Editorial Questions This Project Raises". 6 Mar 2003. Collaborative online writing project. 10 May 2003 http://www.onegecko.com/wikigecko/?WhatAreWeDoing.

Tweet, Johnathan and Mark Rein*Hagen. Ars Magica . Northfield, MN: Lion Rampant, 1987.

Wikipedia. "WikiWiki". Collaborative online encyclopedia. 24 Apr 2003. 10 May 2003 http://www.wikipedia.org/wiki/WikiWiki.