4. Dramatic Interactors: Collaboration, Constraints, and Engagement

For the nonspecialist, the idea of a dramatic model may seem to have more to do with content—interesting situations and colorful characters, for instance—than with structure. As a structuralist, I have been assailed by both theatre and computer people for taking what they perceive as a rather bloodless approach. Structure is not always well understood, and even when it is, its uses are seen to be analytical rather than productive. When we see a good film or go to a good play, we are moved by things that seem to transcend structuralism—a beautiful image, dialogue and action that speak deeply and genuinely about life. There seems to be a contradiction here—if it’s all so structured, how does it get to seem so lifelike? Surely there is more to it than structure, more to it than a computer could be programmed to create. People sometimes criticize my approach by countering that a computer program can never be smart or sensitive enough to make a beautiful work of art. Yet artists use computational tools to do so, and those in turn are enabled by the artistry of designers and programmers.

These observations point to the artistry that is essential in every beautiful made thing. Artistry transcends and saturates the process. We do not know what it is that gives a person the ability to conceive of or create magnificence in art. Structure is not a wholly sufficient explanation for beauty. Human-computer interaction, like other art forms, requires artistry that can only be contributed by human imagination. Artistry is deployed within the constraints of the medium, the tools, and the formal and structural characteristics of the kind of thing that one is trying to create. Artistry and structure are interdependent; both must be present if beauty is to be the result. Perhaps more important in this stage of the evolution of computer-based media is the fact that artistic sensibility should drive the notion of desired experience, from which the design of technological components must be derived.

Human-computer interaction is like drama in the sense that the principal designer (or playwright) is not the only human source of artistry in the completed whole. In theatre, the director, actors, designers, and technicians who are involved in rendering a performance all make contributions that require artistry. In human-computer interaction, there may be a legion of programmers who have designed and architected programs on which a given kind of action depends, graphic designers who create images and animation, wordsmiths who authored text (or text-generating algorithms), and so on. A fundamental but sometimes overlooked source of human artistry is the people who actually engage in the designed interaction; that is, the interactors.

1. Noah Wardrip-Fruin, personal communication, 2013.

The authorship of the designer(s) is of a different order than the creative inputs of the player; the designer authors the world and its affordances, while the player creates a distinct path through the game world that can be said to be the player’s “plot.” This is a stronger force than the reader-response theory, but weaker than the authorship of the designer(s). As Wardrip-Fruin (2009) points out, without players there is no game.

To explicate the diagram shown in Figure 4.1, I want to walk you through it in terms of the four causes (in gray). In Chapter 2, we discussed the efficient cause as the author and her tools. In human-computer interaction, the “authorship” of the interactor’s particular experience is shared in interesting ways. Designers of interactive media are part of the equation, typically working as teams that include many specializations. Their “tools” can be described as representation, computation, and research. Tools for representation include those used for creating graphics, animation, audio, layouts, and interface affordances. Computational tools include the programming of the interactive application itself as well as the code that powers authoring tools for the design team. Another sort of tools, often overlooked, are the methods of design research—studying the intended audience, looking at comparable products, and creating and testing mock-ups and prototypes. Beta testing without the benefit of other design research methods is inadequate. Remarkable resistance to human-centered research persists in many areas—especially in the game industry, with “serious games” as a notable exception.

Interactors typically share in authorship to a lesser degree than designers in that they create under varying kinds and levels of constraints as provided by designers. Affordances for interaction are the most intimate level of collaboration between designers and interactors in the sense that they circumscribe the means, manner, and scope of the interactor’s creative contributions and provide the tools whereby interactors can influence the action.

We have said that material causality reflects the influence of materials upon how they may be formulated at any level in the hierarchy of dramatic elements. The palette of multisensory materials offered up by the designer constrains the sort of patterns or rhythms into which they can be formulated, and those patterns or rhythms constrain how the semiotics or “language” of a piece can be formulated. Thought as expressed or available by inference constrains the formulation of characters, and so on.

Recall that formal causality works in the other “direction,” where the most formal element—plot—constrains the sorts of characters, thoughts, etc. that are appropriate to the action. These two causal forces are at work simultaneously, rather like taking inductive and deductive approaches simultaneously in problem-solving. Game designers often iterate on the basis of observations of or interviews with play-testers and players. Their privileged position allows for intervening and tweaking a game over time. Will Wright famously strolled about The Sims in various forms to observe game play and provide new materials and functionality as he observed emerging play styles (Laurel 2004).

I refer again to the additional causal chains suggested by Michael Mateas (2004). He posits that the player’s intention creates a new chain of formal causality. Mateas’ formulation points to some key differences between drama and dramatic interaction in the operations of causality. But for my purposes, I see the player’s intention as part of the end cause for the player as a co-creator. Mateas also suggests that “material for action” is a separate causal chain in that material requires some sort of interactive affordances in order to be usable by the player. I agree that such affordances are essential, but I see them as being provided by the designer at the level of enactment. If we look at things this way, we may not need to introduce additional complexity to the model.

The authors are working toward similar end causes—the representation of a whole action that produces pleasure. But differences exist. As Mary Flanagan (2009) observes, many players intend to subvert the game—that is, to deny the game’s authority to set the player’s goals. She has produced many games that deliberately leverage this subversive spirit to increase activism and cultural change. Other players may intend more than “winning” or “experiencing” the whole game; they may intend to find personal meaning that transcends a game’s structure. Henry Jenkins, renowned for his work on fandom and popular culture, makes the point that, in order for people to become “fans,” they need to be able to appropriate characters, elements of plot, etc. to construct their own meanings. Jenkins points to the “slash” phenomenon in Star Trek and other cultural properties where fans construct new stories that are personally relevant by writing stories or constructing videos from pieces of the originals that have new plots. Much slash focuses on creating relationships (usually homosexual) or backstories that are not supported in the official canon (see Jenkins 1992 and 2006a). Several sources of causality outside the purview of Figure 4.1 will be discussed ahead.

Multiple interactors engaging in discourses of all kinds go back at least to the earliest BBS (Bulletin Board System). An early precursor to dial-up BBS-like systems was the Community Memory Project in Berkeley, created by Lee Felsenstein in 1973, an electronic walk-up kiosk that worked like a physical bulletin board. Usenet, established at Duke University in 1980, supported threaded discourse among distributed interactors. A person could sign up for a “news feed” on any number of topics. Readers responses were emailed in for moderation, and if they passed the test, their responses would likely show up in the feed in the next few days.²

2. I want to take a moment to honor Eugene Maia for inventing the FAQ (“Frequently Asked Question list”) in the early 1980s. His inspiration was getting sick of people asking the same questions over and over; the canonical answers were posted once a month in any given Usenet discussion list.

My traversal through some of this space began in the mid-1970s at CyberVision, when I was introduced to the Control Data PLATO system. PLATO was heralded as the first “computer-aided instruction system,” created by the University of Illinois beginning in the early 60s. PLATO introduced me to multiplayer flight and maze games as well as message boards, real-time chat, and multi-person forums, some of which had the makings of early collaborative work environments. PLATO also provided me with my first experience of flame wars, in which two or more users would go at one another with ever-escalating vehemence, often “baited” by an original message intended to be provocative. Flame wars can be dramatic, but they pose great challenges to moderators.

The role of the moderator in these early forms was liminal and dynamic. Some of the earliest BBS systems were not moderated, or the “moderator” was likely to be a systems administrator just trying to keep things running smoothly. With the increasing complexity and scope of systems, the mediator’s role tended to become more actively engaged with the community, struggling with governance, setting or enforcing policies, and censoring inappropriate comments, actions, or characters. There was also a pastoral side to the moderator in keeping the virtual community connected, vibrant, and safe.

Free speech and censorship have been abiding issues. Whether getting “toaded” on a MUD or mediated into silence on Usenet, people had things they wanted to talk about that didn’t fit into “polite societies.” Pornography was the leading topic (and probably still is), but all sorts of marginalized voices—from Furries to faeries—wanted to participate in these new forms of communication and community where their own voices can be heard. The alt.* hierarchy was created by John Gilmore and Brian Reid in 1987 in response to a reorganization of Usenet that would eventuate in greater censorship of topics. “Alt” referred to topics that were “alternative”; that is, not part of mainstream popular culture. Although sexual interests made up a fairly large percentage of alt.* topics, many were (and are) also devoted to activism, human rights, and free speech issues.

The WELL (Whole Earth ‘Lectronic Link), founded by Stewart Brand and Larry Brilliant in 1985, became a very tight community in which many of the digerati of those days found a home. The community was friendly toward the Whole Earth movement and reflected some of the distinguishing bits of Northern California culture (e.g., technology; the Grateful Dead). It was originally a dial-up BBS, morphing with technology into its current form as a user-owned virtual community on the Internet.³ The WELL required that people use true names, removing the shield of anonymity that had characterized many early systems. As the World Wide Web became popularized, BBS systems and their kin tended to morph into or be replaced by wikis, Internet forums, websites, and social media.

3. I am forced to recall a certain boss of mine, who in 1993 told a group of researchers that the Web would never be mainstream. He described it as for “. . . only a few geeks and WELL-heads like you, Laurel.” Actually, I think his point was to look past the Web to possible new models. But at that moment, we were all stunned.

Other forms of interaction among interactors happens in the domain of computer-supported collaborative (or cooperative) work (CSCW). The aim here is to facilitate collaboration on a particular problem or opportunity by people in different geographical locations. CSCW relies on any of a variety of computational tools: file-sharing, shared “whiteboards” and tailored work environments, VNC (Virtual Network Computing) as a way to share screens, specialized tools related to the task (e.g., industrial design, architecture, or any of the sciences), video- or voice-conferencing systems, blogs or email, and IRC (Internet Relay Chat), used heavily by such distributed communities as Linux programmers. The tools are varied and rich. Shared goals, the facilitation of collaboration, and working toward consensus distinguish CSCW interactors from participants in forums or social media.

This tiny history reveals the complexity and centrality of interactions among interactors in non-gaming communities. Once the architecture for a BBS or Usenet group or forum has been set up, its content (except that which is “moderated away”) is entirely user-created.⁴ Designers create formal constraints and affordances while interactors provide material all the way up to the level of plot, depending upon magnitude and shape. Interaction between or among interactors may become the primary creators of the plot—the whole action—complete with complication and resolution, discovery, surprise, and reversal.

4. That is, until the onset of the advertising invasion.

Of course, many different kinds of “interactions among interactors” are possible in such systems. People may exchange information, opinions, or goods. One may respond to a post or start a new thread hoping to begin a discussion and possibly to form a new community. One may work with distant colleagues on an invention or a problem. Or one may search anonymously for providers of illicit goods under the anonymity afforded by the alt.net or various “black market” Web sites. In social networks, relationships are power, and groups of like-minded individuals can take meaningful political action. One may enjoy the lives of one’s children by “friending” them on Facebook (ahem).

Habitat, developed by F. Randall Farmer and Chip Morningstar at Lucasfilm, was first launched as a prototype in 1986. It stands as an extremely important transitional form. More than a series of chat rooms or a community like the WELL, Habitat was a graphical virtual community that was both a descendant of the forum and an antecedent to massively multiplayer online games. They called their interactors “players” because they meant the world to be an environment for entertainment and play. Each player took on an “Avatar”—a graphical representation of a character with various signifiers—to represent them. One could also argue that Habitat foreshadowed what became “social media” in the early 2000s (avatars got married in Habitat—in-world only, of course). Randy and Chip’s vision was to make a real instance of “cyberspace,” which, they asserted, was “necessarily a multiple-participant environment” (Morningstar and Farmer 1991). Each of the thousands of “regions” in the game contained “a set of objects which define the things that an Avatar can do there.” The object-oriented approach in building the system was the key to the sort of play that was enabled.

Chip and Randy were constantly observing and tweaking the prototype precisely because it was not a game with rigid rules:

Habitat . . . was deliberately open-ended and pluralistic. The idea behind our world was precisely that it did not come with a fixed set of objectives for its inhabitants, but rather provided a broad palette of possible activities from which the players could choose, driven by their own internal inclinations.

The unexpected actions of players kept Chip and Randy busy, both writing new code and intervening in-world as Avatars. They, like Pavel Curtis, were working at the transformation point of the role of “moderator” from sys-admin to dynamic designer of a community. The success of the prototype and its influence on future forms demonstrate how robustly interactions among participants can shape the dramatic action.

Of course, “non-game” interaction did not end with Habitat’s excursion into an entertaining, graphical, social world. But I see Habitat as a pivotal precursor to later online communities—the world of wikis, Web sites, and blogs—as well as graphical multiplayer games and even “social media.”⁵ The spirit behind it was fundamentally experimental, even though the external driving force was to create a “product” for Quantum Link.

5. Social media is fundamentally narrative, to be discussed later in this chapter in the context of Character.

MUDs (Multiple-User-Dungeons, originally based on Dungeons and Dragons gameplay, later revised to the more generic Multiple-User Domains) arrived on the scene in the late 1970s in the form of Adventure (1975) and Zork (1977). These were text-based multiplayer adventure-type games, and I personally loved playing them (age check). The PLATO system also hosted progenitors for MUDs and MOOs (MUD Object-Oriented) during this time period.⁶ Massively Multiplayer Online Games (MMOGs) showed up in the late 1980s. An explosion of games in the genre followed, while the genre itself branched out to include great new acronyms like MMORTS (Massively Multiplayer Online Real-Time Strategy games) and MMFPS (first-person shooters). Doom is an example of the latter; later examples of the genre include Halo and Call of Duty. In 1991, Neverwinter Nights, published by America Online, was the first graphical online role-playing game (MMORPG). The MMORPG genre was popularized on the Internet by Ultima Online (1997) and Everquest (1999). MMORPG games dominate the landscape today, although the MMFPS and MMORTS forms continue vigorously as well.

6. This and more information on PLATO games is available an the Universal Videogame List, www.uvlist.net/platforms/games-list/181 (©1998, retrieved 04/23/13).

The elephant in the marketplace is World of Warcraft, originally introduced in 1994 and going strong with over 10 million subscribers in 2013—the largest MMORPG in history. Various types of interactions are enabled by the various “realms” of the game, each with distinctive play properties to suit the palate of the player (for example, how much fighting they want to do). Non-player characters (NPCs)—often with fairly sophisticated AI structures—serve as enemies, friends, wizards, familiars, monsters, and other sorts of forces on the level of character to shape dramatic action. Through devices like deeds, quests, and guilds, WoW as well as many other games of its ilk, provides affordances—often necessities—for significant interaction among players, to work together for common goals or against common enemies.

Lord of the Rings Online (LotRO, launched in 2007) employs similar structures that necessitate collaboration. Player-characters have vocations and talents, and most quests cannot be completed by a solo player because they don’t have the requisite talents. For example, the “vocation” of the player-character consists of two talents that go together and one that does not. A Tinkerer, for example, can find ore, make jewelry, and collect wood. She can’t make anything out of wood, but someone who can will trade her for it. “Everybody gets good equipment out of exchanges,” says regular LotRO player Lisa McDonald. Trade and commerce—the internal economy of the game—are extremely important to gameplay.⁷

7. Interview with LotRO player Lisa McDonald, April 2013.

Beginning in the early 2000s, Voice over Internet Protocol (VoIP) has grown as part of multiplayer game experience. Players use a voice channel to shorten communication time, for example. They also use it in-game for social chat and networking. This channel of communication is human-to-human, not to be confused with speech recognition. Voice can enrich the game experience in many ways, from direct impact on the strategies and actions of a player or group to emotional depth and social interaction as well as opportunities for shared criticality.

Like other kinds of properties, multiplayer games engender enthusiastic fan activities outside of the game world, including fan art, conventions, Cosplay, and intertextual fan activities (Jenkins 2006a). These activities form economies of attention as well as legitimate commerce. They provide ways for fans to extend their personal constructions of meaning. Purple Moon provides an earlier example, but one near to my heart. Near the end of the company (and its eventual acquisition), we discovered multiple fan sites where “scarce” gifts and objects from the Web site were being traded by girls. In 2012, the mother of one of the original players alerted me to a Facebook Community called “I Miss Purple Moon.” Sweet!

Of course, interaction among players has its dark sides. Sexual harassment continues to be an issue. Cheating in various forms continues as a thriving parasite industry. The ready availability of “legitimate” cheat books and websites suggests that the game industry has had to give up on controlling many forms of cheating and find ways to embrace them. Some forms of cheating may be “blessed” as “subversive play,” but not all. In many cases, the player simply wants to “get ahead” without breaking a sweat—and that’s not subversive, just lazy in a human sort of way.

“Black market” activities are a constant plague for players and companies alike. The sale of virtual gold for real-world money in WoW has been a flashpoint; Blizzard (publisher of WoW) and Antonio Hernandez, a WoW player, have both filed suit against companies for such practices. Hernandez’ suit, filed in 2007, was meant to be a class action against Internet Games Entertainment (IGE). Patentarcade.com, a website devoted to IP protection and the gaming industry, reported that:

The amended complaint in the Hernandez suit alleged that “IGE’s calculated decision to reap substantial profits by knowingly interfering with and substantially impairing the intended use and enjoyment” of WoW through its gold-farming, camping spawns, and spamming chat . . . led to lost time, competitive disadvantage, and diminished experience for honest game subscribers (Patentarcade Staff 2009).

Both of these suits were settled, but such practices continue to pop up in ephemeral companies that form fluid but irrepressible parasitical industries, including the sale of accounts with highly valuable characters—a practice forbidden but not snuffed out by most publishers of multiplayer games.⁸ Hey, sounds like a good game premise to me. I’m sure somebody’s done it.

8. Like the “ask” for donations to political candidates and even to parties trying to advance legislation, it seems that gaming—like democracy—comes with corrupting influences that must be borne (for the time being) by players and citizens. Of course, opinions may vary.

In conclusion, the previous two sections are intended to illustrate many of the ways in which interactors or players exert causal influences through their interactions with one another that are outside of the direct control of designers. By providing affordances for discourse and discussion as well as affordances that encourage or require group action within multiplayer games, designers create conditions for an efflorescence of possibilities for action and experience. At the same time, designers rely on the social, strategic, and artistic actions of individuals to enhance the dramatic shape of incidents and whole actions. Both designers and players can fall prey to parasitic forces that intend to subvert the intended experience. In many cases, designers have had to “embrace and enfold” such forces because of their power (e.g., sale of in-world materials for real-world money) or popularity, as in the case of “cheats,” acknowledging to varying degrees that they have become normative. Both designers and interactors are constantly called upon to deal with the various dark economies that plague (and tempt) them. It is up to the designer (or publisher) as well as the virtual community of interactors to safeguard the experience.

Constraints—limitations on people’s actions—may be expressed as anything from gentle suggestions to stringent rules, or they may only be subconsciously sensed as intrinsic aspects of the thing that one is trying to do or be or create. People are always operating under some set of constraints: the physical limitations of survival (air to breathe, food, and water); the constraints of language on verbal expression; the limitations of social acceptability in public situations (e.g., wearing clothes, usually). The ability to act without any such constraints is the stuff of dreams—the power of flight, for instance, or the appeal of immortality. Yet even such fantasy powers can be lost by the failure to comply with other, albeit mythical, constraints (witness Prometheus). It is difficult to imagine life, even a fantasy life, in the absence of any constraints at all. Good designers are more likely to argue for than against constraints on their own work; constraints give us things to push against and may call up our highest creativity.

The design of human-computer interaction should be informed by an analysis of constraints to determine what kinds of constraints are most appropriate. That analysis begins with understanding the various reasons why constraints are necessary.

The platform-related reasons for constraints are fairly straightforward. They will also change, depending upon the elaborateness, completeness, and cost of various implementations of the system. For example, pointing devices that can be used to enable gestural input may have a limited range, constraining people to stand within range of a receiver. The Wii and Kinect are examples of systems that extend human physical involvement; the constraints surrounding their use must be manifested effectively. For example, Nintendo tried to control the all-too-common accident of throwing the controller at the TV by adding a strap, but this did not control people who ignored the strap or got sweaty palms. Physical acts like running or manipulating objects in a Virtual Reality world require conventions whereby the desire to perform such actions can be expressed. Such conventions, mandated by the technical limitations of systems, are a form of constraints.

Constraints are necessary to contain the action within the mimetic world—a design problem. For example, in an interactive fantasy version of a Sherlock Holmes mystery, it would be important to constrain people to the customs and technology of Arthur Conan Doyle’s 19th-century London (e.g., no computational spyware). Any human-computer system, no matter how elaborate, cannot be expected to comprehend all possible worlds simultaneously. Constraining how or whether people can introduce new potential into a dramatic interaction is essential in the creation and maintenance of dramatic probability.

9. From someone who did just that: “Cyberspace. It sounded like it meant something, or it might mean something, but as I stared at it in red Sharpie on a yellow legal pad, my whole delight was that I knew that it meant absolutely nothing.”—William Gibson, on his invention of Cyberspace, from his talk at the New York Public Library on April 19, 2013.

Many thinkers have explored the relationship between creativity and limitations in some depth. In general, the literature continues to argue for the value of constraints in encouraging creativity, in the arts, business, and life. Patricia Stokes (2005) says:

I like to think of constraints for creativity as barriers that lead to breakthroughs. One constraint precludes (or limits search among) low-variability, tried-and-true responses. It acts as a barrier which allows the other constraints to promote (or direct search among) high-variability, novel responses that could prove to be breakthroughs.

In his classic book The Courage to Create (1975), psychologist Rollo May asserted the need for limitations in creative activities:

Creativity arises out of the tension between spontaneity and limitations, the latter (like river banks) forcing the spontaneity into the various forms that are essential to the work of art. . . . The significance of limits in art is seen most clearly when we consider the question of form. Form provides the essential boundaries and structure for the creative act.

A system in which people are encouraged to do whatever they want will probably not produce pleasant experiences. When a person is asked to “be creative” with no direction or constraints whatsoever, the result is, according to May, often a sense of powerlessness or even complete paralysis of the imagination. Limitations—constraints that focus creative efforts—paradoxically increase one’s imaginative power by reducing the number of open possibilities. Limitations, May says, provide the security net that enables a person to take imaginative leaps:

Imagination is casting off mooring ropes, taking one’s chances that there will be new mooring posts in the vastness ahead. . . . How far can we let our imagination loose? . . . Will we lose the boundaries that enable us to orient ourselves to what we call reality? This again is the problem of form, or stated differently, the awareness of limits.

The nature of a mimetic world provides a similar security net. Generally speaking, people know that things work better when they respect the limits of a mimetic world as indicated by its structure and affordances as well as the model of it that people are building through experience. In exchange for this complicity, people experience increased potential for effective agency, in worlds in which the causal relations among events are not obscured by the randomness and noise characteristics of open systems (like “real life”). People may likely push on the edges of a mimetic world as part of exploration or even in an effort to hack it.¹⁰ Designers need to be flexible and to apply new constraints when they observe actions that disturb the desired structure of experience.

10. Since I first wrote this book, subversive game-play has become quite a bit more popular. Initially, there were scandals over black-market sites for illegitimately acquiring more powerful characters. The designers of “America’s Army” were taken by surprise when players hacked themselves unforeseen superpowers. Since those days, subversive gaming has been (subversively) legitimized as a genre in its own right.

Constraints can be either explicit or implicit. Explicit constraints, as in the case of menus or command languages, are undisguised and directly available. In terms of what Wardrip-Fruin (2009) calls the “surface”—the appearance, affordances and behaviors of the delivery system, peripherals, and controllers—explicit constraints may be straightforwardly expressed (although one may still have to consult a manual). In game and non-game environments, explicit constraints may be expressed during the “setup” phase or in the exposition. Exposition in a game can be as simple as the descriptive text on a cover or Web site. Implicit constraints, on the other hand, must be inferred by interactors and players from the behavior of the software system. Implicit constraints exist in the presence or absence of affordances for making certain kinds of choices or performing certain kinds of actions. For example, in most combat-based action games (FPS), it is not possible to negotiate with the enemy.

Explicit constraints can be used without damage to engagement if they are presented before the action begins. Arguments about rules can seriously disrupt the flow of both work and play. Both activities feel better when the “rules” or “operational principles” of an activity are clearly articulated as a frame for action. An interesting exception is the ongoing process of rule-making and enforcement that is sometimes an element in children’s play—a sort of meta-game that provides its own distinct pleasures. In my research on play, I’ve learned that kids generally do not experience disruption as they shift from play to meta-play.¹¹ A similar meta-game occurs in the theatre when stagehands and “real people” wander in and out of the action, as in some of the plays of Christopher Durang and Thornton Wilder, or in certain productions of Brecht. Seen in this way, the meta-game is also mimetic, and the actors are merely performing the roles of “real people” as well as portraying other dramatic characters. Because it is mimetic, this is a “false” context shift, much like a play within a play, or a dream in which one has false awakenings. Such meta-games or meta-plays do not necessarily violate engagement, but may enhance it through the same means as the mimetic “core” activity.

11. For example, I once observed the following with a couple of six-year-olds: “Oh I am the princess in the tower and you must save me.” “Okay, I am coming on my horse to save you.” [some action ensues] “Okay, now we are married and we have a baby!” “Give me the baby. I want to hold it.” There was no evidence of disruption in this sequence.

Constraints may also be characterized as extrinsic or intrinsic to the mimetic action. Extrinsic constraints have to do, not with the mimetic context, but with the context of the person as an interactor. Constraints should not be left entirely to the interface affordances of the hardware. Avoiding the “reset” and “escape” keys during play of a game has nothing to do with the game world and everything to do with the behavior of the computer. Playing a theatrical scene without the use of language (as an exercise) is an extrinsic constraint designed to improve the actors’ acuity in physical expressiveness—a different context than the mimetic one. Extrinsic constraints have been used successfully in a variety of sports and other disciplines to distract the part of consciousness that can interfere with performance.¹² The technique should be used cautiously in human-computer interaction, however, because it has the potential to set up a secondary context that demands part of a person’s attention, disrupting “flow.”

12. See, for example, one of the first best-selling books in this genre: W. Timothy Gallwey, Inner Tennis: Playing the Game (New York: Random House, 1976).

Extrinsic constraints can be made to appear intrinsic when they are expressed in terms of the mimetic context. If the “escape” key in a game is identified as a self-destruct mechanism, for instance, the constraint against pressing it in the course of flying one’s mimetic spaceship is intrinsic to the action. A person need not shift gears to consider the effect of the key upon the computer or the game. Expressing constraints this way may preserve the contextual aspect of engagement.

Designer Emily Short (2013b) describes the extrinsic and intrinsic qualities of an interactive storytelling system called Versu in different terms. She refers to “extrinsic” representation as “information about the extrinsic narrative part of the story,” and intrinsic representation as “character files that contain the intrinsic content.” In Versu, interactors can create stories using these two forms of representation, in which the story is dynamic within the larger narrative frame by getting people “to remix aspects of the story.”

Ideally, intrinsic constraints should not shrink people’s perceived range of freedom of action, but rather enhance them: intrinsic constraints should limit, not what people can do, but what they are likely to think of doing. Intrinsic constraints, when successful, reduce the need for explicit limitations on people’s behavior. Context is the most effective medium for establishing implicit constraints. The ability to recognize and comply with intrinsic, context-based constraints is a common human skill, exercised automatically in most situations, and not requiring concentrated effort or explicit attention. It is the same skill that a person uses to determine what to say and how to act when he interacts with a group of unfamiliar people—at a party, for instance. The limitations on behavior are not likely to be explicitly known or consciously mulled over; they arise naturally from one’s growing knowledge of the context.

The situational aspects of the current context and the way in which they have evolved over the course of the action establish dramatic probability that influences a person’s actions and expectations. In summary, then, constraints that are implicit and intrinsic to the mimetic context are least destructive of engagement and flow, although explicit and extrinsic constraints can be successfully employed if they frame rather than intrude upon the action.

We can look for guidance in the development of constraints to other dramatic forms: theatrical performance and improvisation. In the theatre, the actor is constrained in the performance of his character primarily by the script and secondarily by the director, the accoutrements of the theatre (including scenic elements, properties, and costumes), and the performances of his fellow actors. The actor must work within exacting constraints, which dictate the character’s every word, choice, and action. In spite of these narrow limits, the actor still has ample latitude for individual creativity. In the words of legendary acting teacher Michael Chekhov (1953):

. . . every role offers an actor the opportunity to improvise, to collaborate and truly co-create with the author and director. This suggestion, of course, does not imply improvising new lines or substituting business for that outlined by the director. On the contrary. The given lines and the business are the firm bases upon which the actor must and can develop his improvisations. How he speaks the lines and how he fulfills the business are the open gates to a vast field of improvisation. The “hows” of his lines and business are the ways in which he can express himself freely.

The value of limitations in focusing creative activity is recognized in the theory and practice of theatrical improvisation. Constraints on the choices and actions of actors improvising characters are probably most explicit in the tradition of Commedia dell’arte. Stock characters and standard scenarios provide formal constraints on the action, in that they affect the actor’s choices through formal causality. Conventionalized costumes for each character, a collection of scenic elements and properties, and a repertoire of lazzi (standard bits of business) provide material constraints on the action.

For instance, a person interacting with a simulation of a space station might be trying to redesign it, trying to learn how to operate its controls, or perhaps to experience the environment under various conditions. There is the beginning of a “plot” implicit in each of these goals; a well-designed system assists in bringing that plot to life. When an interactor’s objective has been established with a high degree of confidence, the system might kick off a specific scenario by presenting a tailored exposition or inciting incident.

Mateas’ thesis regarding new lines of causation (Mateas 2004) submits that an interactor’s intentions form a new vector of formal causality, and I asserted earlier in this chapter that the interactor’s intention may be understood as part of the end cause—that is, what the interactor desires at the end of the day. But there are subtleties in the domain of intention and motivation that make their influence on causality even more slippery to pin down. The overlapping nature of the interactor as a person and also an agent (character) in a mimetic context creates complexity.

Let’s work through a few examples. As a person, I want to create a budget plan for my household. As an agent (or character), I wish to use the affordances of my application(s) to do so in a clean and effective way. As a person, I begin to discover dependencies and categorical subtleties that I have not foreseen. Perhaps I have categorized both household products and food together as groceries, or perhaps I have remembered that sales tax has an impact on my income tax, but I have not recorded where I bought certain items and which sales taxes they have been subject to. As an agent, I recognize that I can’t meet my goal with mushy categories and incomplete data (change in thought). As a person, my goal changes to create a more precise set of categories and to figure out how to do better accounting of sales tax. At this point, I revise my motivation from planning a budget to creating a better record and understanding of what I am spending now (change in end cause). As an agent, instead of a planner, I am now a researcher and record keeper (change in character). I find myself dealing with different affordances to take different actions for different goals.

Several things are going on here. On the face of it, we have a relatively simple state machine. The subtlety is how a stalled state on the part of the agent (thought) causes the person to change their end cause as an interactor. A change in the purpose of the activity will change its plot (the whole action). That exerts a formal force on the “character” of the agent—its traits and predispositions—in order to produce appropriate actions.

Character serves as a constraint system in rather a different way in a multiplayer online game. If, as in LotRO, a character with a particular vocation needs to find folks whose talents and possessions lead to fruitful exchanges and group actions, then the character’s needs constrain the player to behave in certain ways—to find things to trade, to become more visible to potential partners (reputation), to show oneself to be trustworthy, and so on. In WoW, the PvP (Player versus Player) realm assumes combat as the norm; players must fight and can be attacked at nearly any time. Those characters constrain their players to develop combat skills, create alliances, and acquire battle gear and powers. The PvE realm (Player versus Environment) allows players to choose whether or not their characters battle with other characters. WoW is much more complicated and subtle than this, but the mode of play as well as affiliations and commitments among characters constrain players’ choices and actions in both simple and complex ways. It is not surprising that many players run multiple characters in different realms to experience different kinds of play.

13. One of the hardest personas to shake is the one a parent takes on when children are young. As the children grow up, there is a strong pull to change the persona to one that is more equal or genuine, complete with bad behavior. A fine line is walked by parents who are also tugged at to protect and advise their grown children.

“Character” has two meanings that problematize things further. The first meaning is one’s authentic moral or ethical nature, as in “he has a good character.” The second meaning is drawn from drama and narrative—the way we have used “character” in this book—to mean representation that is made up of the material of thought and performs actions that contribute to the plot, or whole action. I have tried to make the case that an interactor in a non-game environment is performing “character” in essentially the same way. Persona comes about in a different manner. Persona creation consists of acts of collage in social media. Even in multiplayer games, a player may have character (in both senses of the word) as well as a persona. For example, I may perform a very bad-assed, tough-guy character (dramatic sense) in a multiplayer game, but because I don’t cheat or camp or behave in otherwise dubious ways, other players may judge that I as a player am a person of good character. In voice communication or on boards related to the game, I may present a persona to other players—e.g., expert player, mentor, activist, etc.

The construction of personae in social networks that are not games is a tricky business. Let’s use photography as an example. We all see the enormous changes that have been wrought by point-and-shoot cameras that can enable unskilled folks to make high-quality (if not high-art) images. The coupling of cameras with smart phones makes the point-and-shoot practice even more tempting because it enables near-instantaneous sharing of one’s “here and now.” The ability to distribute images through various social networks has accelerated the proliferation of photography as a way to play and communicate as well as a way to create one’s persona.

In terms of technology, one may be an expert in using a sophisticated camera and editing suite, or one may choose to point and shoot, then apply any of a plethora of “effects” applications to create interesting-looking images with much less investment in time or expertise. Both can be engaging activities when the tools and their affordances are well designed. They do differ, however, on the level of character. A “professional” photographer intends to express complex ideas with greater depth—to create “fine art.”¹⁴ A casual photographer intends, usually, to express the here-and-now with less planning and attention to expression. As a persona, “expert photographer” will likely not be a good match for the casual photographer who uses quick effects. Although the lucky casual photographer with an excellent eye may be able to get away with it, the casual photographer is more likely to earn a persona identified with a sort of visual gregariousness. In this hypothetical example, we can see how character can constrain persona: A person’s “character” (choices and actions) may constrain the sort of persona they may credibly create. Since social networks are so much about personae, I see them as more narrative than dramatic in structure. It may be, however, that the action of constructing a persona has its own dramatic arc.

14. Hilary Hulteen (professional photographer and my grown daughter), personal communication.

In this section, we can see that Character operates as a lynchpin in holding the structure of the experience together. Character is where the interactor’s and the designer’s intents meet. So, in addition to the forces of the four causes shown in Figure 4.1, we can view the level of character as an important locus of constraints.

Engagement, as I use the concept in this book, is similar in many ways to the theatrical notion of the “willing suspension of disbelief,” a concept introduced by early 19th-century critic and poet Samuel Taylor Coleridge.¹⁵ It is the state of mind that one must attain in order to enjoy a representation of an action. Coleridge believed that any idiot could see that a play on a stage was not real life (Plato would have disagreed with him, as do those in whom fear is induced by any new representational medium, but that is another story). He noticed that, in order to enjoy a play, one must temporarily suspend (or attenuate) one’s knowledge that it is “pretend.” One does this “willingly” in order to experience other emotional responses as a result of viewing the action. When the heroine is threatened, we feel a kind of fear for and with her that is recognizable as fear, but different from the fear we would feel if we were tied to the railroad tracks ourselves. Pretending that the action is real affords us the thrill of fear; knowing that the action is pretend saves us from the pain of fear. Furthermore, our fear is flavored by the delicious expectation that the young lady will be saved in a heroic manner—an emotional response that derives from knowledge about the form of melodrama.

15. For an analysis and thorough bibliography of Coleridge’s criticism, see Literary Criticism: Pope to Croce, pp. 221–239.

The phenomenon that Coleridge described can be seen to occur almost identically in drama and computer games, where we feel for and with the characters (including ourselves as characters) in very similar ways. Yes, someone might cry, but manuscripts and spreadsheets aren’t pretend! Here we must separate the activity from its artifacts. The representation of a manuscript or spreadsheet as we manipulate it on the screen is in fact pretend, as compared to physical artifacts like data files (in memory or on a storage medium) and hard copy. The artifacts are real (as are actors, lighting instruments, and scenery in a play), but the rules involved in working with the representations of dramatic actions or interactions are distinct from the artifacts. Why? First, the fact that they are representations is the key to understanding what we can do with them. Second, their special status as representations affects our emotions about them, enabling experiences that are, in the main, much more pleasurable than those we regularly feel in real life. The distinguishing characteristic of the emotions we feel in a representational context is that there is no threat of pain or harm in the real world.¹⁶

16. This principle suggests that activities like running a nuclear reactor or launching a spacecraft—things with real potential in the real world—should be taken off the table when we talk about dramatic interaction. For example, the control system on a nuclear reactor involves many, many representations of the state and operations of various system components, but in the context of real-world consequence, these representational affordances are much more about human factors and tele-operations than they are about the pleasure of interaction.

Further, engagement entails a kind of playfulness: the ability to fool around, to spin out “what if” scenarios. Such “playful” behavior is easy to see in the way that people use photo editing suites and document creation software. The key quality that a system must possess in order to foster this kind of engagement is reversibility; that is, the ability to take something back. In the age of the Internet, taking something back once it is published is nearly impossible. We and our children need to understand that; fooling around is playful, but publishing is forever.

Engagement is what happens when one is able to give oneself over to a representational action, comfortably and unambiguously. It involves a kind of complicity. We agree to think and feel in terms of both the content and conventions of a mimetic context. In return, we gain a plethora of new possibilities for action and a kind of emotional guarantee. One reason why people are amenable to constraints is the desire to gain these benefits.

Engagement is only possible when one can rely on the system to maintain the representational context. A person should not be forced to interact with the system qua system; indeed, any awareness of the system as a distinct, “real” entity would explode the mimetic illusion, just as a clear view of the stage manager calling cues would disrupt the “willing suspension of disbelief” for the audience of a traditional play. Engagement means that a person can experience a mimetic world directly, without mediation or distraction. Harking back to the slogan “the representation is all there is,” we can see that designers are often engaged in the wrong activity: that is, representing what the computer is doing. The proper object of interaction design is what the interactor is doing and experiencing—the action. Thinking about things this way automatically avoids the trap doors into meta-level transactions with “the system.”

Agency is a key component of first-person experience. Mateas and Stern (2005) provide an excellent description in relation to the development of their experimental game, Façade:

Like contemporary games, Façade is set in a simulated world with real-time 3D animation and sound, and offers the player a first-person, continuous, direct-interaction interface, with unconstrained navigation and the ability to pick up and use objects. More importantly, as in successful games, the player is intended to have a high degree of agency. A player has agency when she can form intentions with respect to the experience, take action with respect to those intentions, and interpret responses in terms of the action and intentions; i.e., when she has actual, perceptible effects on the virtual world.

Although one may describe experiences in which one is not an agent using first-person pronouns (I saw this, I smelled that), the ability to do something sooner or later emerges as a criterion. On the one hand, doing very simple things can be an expression of agency: looking around, for instance, or reaching out and touching something. Such simple types of agency are often responsible for the “breakthrough” experiences reported by many people who have used virtual-reality systems.¹⁷ On the other hand, doing something relatively complex in an indirect or mediated way may not have a first-person feel. In the early days of computing, a programmer would submit a program and data on punched cards and come back to pick up the results a day or two later. Although they were telling the computer what to do quite exactly, during the hours of waiting for the computer to “crunch” those numbers, programmers were not experiencing a feeling of agency. Today, imploring a system to do something in highly constrained, formal language can engender a similar feeling that somebody (or something) else is in control.

17. Rob Tow formulates what he calls “The Principle of Action” in terms of sensation and action in virtual reality (Laurel, Strickland, Tow 1994).

First-person sensory qualities are as important as the sense of agency in creating satisfying human-computer experiences. Quite simply, the experience of first-person participation tends to be related to the number, variety, and integration of sensory modalities involved in the representation. The underlying principle here is mimetic; that is, a human-computer experience is more nearly “first-person” when the activity it represents unfolds in the appropriate sensory modalities. Trends of technical evolution in the output of simulators and games—toward higher resolution graphics and faster animation, greater sound capabilities, and motion platforms, for example—seem to confirm this notion. Likewise, mimetic input devices like force-feedback controllers, controllers that enable computers to detect motion in 3-space, and affordances for recognizing speech, gestures, and faces provide a greater sensory palette (and greater “directness”) for the interactor.¹⁸

18. An interesting exception is the “big-pixel” look currently popular in Indie games like Sissyfight. In this genre, it is likely that lower visual definition and imitation of the look of early games act as signifiers for the Indie Games movement.

Sensory first-personness, then, is clearly not limited to the system’s “output”; it includes the modalities that people can employ when they take action in mimetic worlds. The desire for symmetry between “input” and “output” modalities is strong. Engagement may be disrupted when an application talks to me (especially if it asks me a question) and I can’t talk back, at least until conventions of communication have been successfully (and hopefully painlessly) communicated. Further, the real-world relationships among modalities affect our expectations in representational worlds that include them; for instance, greater force applied to the throwing of an object should make it appear to go farther, surfaces that look bumpy should feel bumpy, and balloons should make noise when they pop.

When we contemplate the complexity involved in creating first-person experiences, we are tempted to see them as a luxury and not a necessity. But we mustn’t fall prey to the notion that more is always better, or that our task is the seemingly impossible one of emulating the sensory and experiential bandwidth of the real world. Artistic selectivity is the countervailing force—capturing what is essential in the most effective and economic way. A good line-drawn animation can sometimes do a better job of capturing the movements of a cat than a motion picture, and no photograph will ever capture the essence of light in quite the same way as the paintings of Monet. The point is that first-person sensory and cognitive elements are essential to human-computer activity. There is a huge difference between an elegant, selective, multi-sensory representation and a representation that squashes sensory variety into a dense and overheated glob (see McLuhan 1964).

Multi-sensory experience offers advantages that go beyond engagement, as media theorist Tom Bender (1976) describes:

The kinds of information we receive from our surroundings are quite varied, and have different effects upon us. We obtain raw, direct information in the process of interacting with the situations we encounter. Rarely intensive, direct experience has the advantage of coming through the totality of our internal processes—conscious, unconscious, visceral and mental—and is most completely tested and evaluated by our nature. Processed, digested, abstracted second-hand knowledge is often more generalized and concentrated but usually affects us only intellectually—lacking the balance and completeness of experienced situations. . . . Information communicated as facts loses all its contexts and relationships, while information communicated as art or as experience maintains and nourishes its connections.

Bender’s observations have been supported quite persuasively in computer-based educational activities. Educational simulations excel in that they present experience as opposed to information. Learning through direct experience has, in many contexts, been demonstrated to be more effective and enjoyable than learning through “information communicated as facts.” Direct, multi-sensory representations have the capacity to engage people intellectually as well as emotionally, to enhance the contextual aspects of information, and to encourage integrated, holistic responses. This broad view of information subsumes artistic applications, as well as traditional knowledge representation. What Bender calls “direct experience,” plus the experience of personal agency, are key elements of human-computer interaction.

19. Many years later, we went to the Borg Invasion, a motion-platform ride that was part of the now-defunct Star Trek Experience at the Hilton in Las Vegas. At one point, live actors impersonating Borg appeared through a ceiling panel and grabbed a hapless girl (ringer) and pulled her away. My younger daughter was REALLY scared—for a few moments.

Even in task-oriented applications, there is more to the experience than getting something done in the real world, and this is the heart of the dramatic theory of human-computer interaction. Our focus is not primarily on how to accomplish real-world objectives, but rather how to accomplish them in a way that is both pleasing and amenable to artistic formulation; that is, in a way in which the designer may shape a person’s experience so that it is enjoyable, invigorating, and whole.

When we participate as agents, the shape of the whole action becomes available to us in new ways. We experience it, not only as observers or critics, but also as co-makers and participants. Systems that incorporate this sensibility into their basic structure, open up to us a cornucopia of dramatic pleasures. This is the stuff of dream and desire; of life going right. It is the vision that fuels our love affair with art, computers, and any other means that can enhance and transform our experience.

The experience of pleasure in a whole action is also influenced by how that action is defined or bounded. In the domain of document creation, for instance, my pleasure and satisfaction has been enormously increased by developments in word processing, document design, and printing technology that allow me to engage in more of the whole action, from inception to final result. In the days of typewriters (age check), one created documents that would be happily transformed in appearance through the process of publication. Through the addition of document design to the application of word processing, and with the assistance of a good printer, I can now influence the final appearance of a publication through my own (design and formatting) actions, and I can bask in the sense that the thing is really done by seeing it in something that closely approximates its published form.

The most complex and rewarding result of dramatic action is catharsis, defined by Aristotle as the pleasurable release of emotion. That’s not to say that all emotions aroused by a play are necessarily pleasant ones. Pity, fear, and terror are mainstays of non-comic forms. It is not the emotion itself, but its release that is deemed “pleasurable.” Further, emotions aroused by a play differ in context and expectation from those experienced in real life. When one is viewing a play or film or even riding a roller coaster, one expects emotions to be aroused and to have the opportunity to release them. Aristotle’s point is that emotional arousal and release is intrinsically pleasurable in the special context of representations; indeed, that is one of their primary values to us.

In Chapter 1 we discussed a Brechtian view of catharsis that suggests that emotional closure necessarily takes place beyond the temporal “ending” of a play. Brecht’s hypothesis was based on a view that requires the integration of the experience of a play into one’s ongoing life. Brecht’s ideas have been interpreted primarily in a political and social light. Many contemporary “serious games” use a Brechtian approach to catharsis. For example, Inside the Haiti Earthquake (PTV Productions 2010), the companion to a documentary film, “challenges assumptions about relief work in disaster situations.” The goal of EVOKE (McGonigal 2010) is “to help empower people all over the world to come up with creative solutions to our most urgent social problems.” These games are part of the “Games for Change” movement. Catharsis could be defined as actions a player may take that influence things outside of the in-game experience.

Catharsis depends upon the way that probability and causality have been orchestrated in the construction of the whole, as well as upon our uninterrupted experience of engagement with the representation. More than that, it is the pleasure that results from the completion of a form. The final form of a thing may be suspected from the beginning or unforeseen until the very end; it may undergo many or few transformations. It may be happy or sad, because the “success” of the outcome in terms of the representational content is not nearly so potent as the feeling of completion that is implicit in the final apprehension of the shape of a whole of which one has been a co-creator. The theory of catharsis dictates that, no matter how monumental or trivial, concrete or abstract, the representation affords the occasion for the complete expression of those emotions that have been aroused in the course of the action. In plain terms, it means that we must design clear and graceful ways for things to end.

Of all forms of human-computer activity, computer games are both the worst and best at providing catharsis. They are the best when a player experiences completion (by “winning,” finishing a journey, or other means), and they are the worst when the action is truncated because it could not continue.²⁰ In task-oriented environments, the trick is to define the “whole” activity as something that can provide satisfaction and closure when it is achieved. This depends in part on being able to determine what a person is trying to do and striving to enable them to do all of it, even when they opt to do it in definable chunks. In simulation-based activities, the need for catharsis strongly implies that what goes on be structured as a whole action with a dramatic “shape.” If I am flying a simulated jet fighter, then either I will land successfully or be blown out of the sky, hopefully after some action of a duration that is sufficient to provide pleasure has had a chance to unfold. Flight simulators shouldn’t stop in the middle, unless the training goal is simply to help a pilot learn to accomplish some mid-flight task. Catharsis can be accomplished, as we have seen, through a proper understanding of the nature of the whole action and the deployment of dramatic probability. If the end of an activity is the result of a causally related and well-crafted series of events, then the experience of catharsis is the natural result of the moment at which probability becomes necessity.

20. Here again, it seems that the designers at Lucasfilm were in the forefront. Ron Gilbert (1989) counseled game designers to avoid situations in which a player must “die in order to learn what not to do next time.” In a presentation at SIGGRAPH 1990, LucasArts Entertainment’s research director Doug Crockford showed a re-edited version of Star Wars in which Luke Skywalker was killed in his first battle with Darth Vader. The story was over inside of 30 seconds.

In this chapter we have analyzed various ways in which dramatic ideas and techniques can be employed to influence the way human-computer activities feel to people who take part in them. Hopefully, it has illustrated some of the benefits of a dramatic approach in terms of engagement and emotion. The chapter has emphasized the need to delineate and represent human-computer activities as organic wholes with dramatic structural characteristics. It has also suggested means whereby people experience agency and involvement naturally and effortlessly. The next chapter explores structural techniques more deeply, returning to Aristotle’s six elements, and suggesting principles and rules of thumb for designing each of them in the computer domain.