What about interactive movies? We regret to inform you that they’re a myth. A movie, like theatre, is storytelling. Regardless of the point of view, a movie is a screenwriter’s story told through action and dialogue in a linear sequence that the author intended.
Multiple branches and endings, cute gimmicks that they are, rarely enhance a story, and are poor excuses for interactivity. Screenwriters entering the interactive arena beware this booby trap: it’s not just a movie with buttons.
– Robert Gelman and Kenneth Melville, in “On the Trail of the Interactive Grail”
Video Games as Interactive Action Films
You’re sitting in the cockpit of a sophisticated single person spacecraft. Your mission is to fly down a small corridor, walls on either side of you, until you get to a small hole in the surface of the planet, in which you will deposit a bomb. The mission is extremely dangerous: you have to dodge laser fire emanating from the surface of the planet and enemy spaceships are hard on your trail. One small miscalculation and you will go careening into one of the walls on either side of you. But the stakes are high: the freedom of all life forms in the galaxy depends on you.
This could be a description of a video arcade game, but astute movie fans will recognize it as the climax of George Lucas’ film Star Wars (1977). Although interactive narrative is new to most people, it has been around for at least a decade in the form of video arcade and home computer games, and these have been heavily influenced by traditional film. In fact, most video games are simply action films with the addition of interactivity.
The qualities which both action films and video games share include:
- one-dimensional characters, usually defined in terms of good and evil, who have simple motivations;
- simple, straightforward conflict;
- rising action involving increasing stakes (in film, referred to as the plot points between the acts; in video games, known as different levels).
The similarities between video games and action films makes it easier to understand why stories move so easily between the two media. Video games are now made into films (Streetfighter, 1994, and even Super Mario Brothers, 1995), while films are often reworked to be released as video games (series such as Star Trek, begun in 1979, and Batman, which started in 1989, as well as individual films from Blown Away, 1995, to Aladdin, 1992). In fact, some action films are shot at the same time as their video counterparts, as Pimentel and Teixeira relate: “For Jurassic Park (1993), Sega has a team of writers, artists and composers combining actual film footage with animation and sound to produce the game… Sega of America has built its own multimedia production studio so they’re able to release the game at the same time the movie debuts. [my italics]” [110]
So, with a little tweaking, imagine Lucas’ scene recast for an interactive medium: you’re sitting in the cockpit of a sophisticated single person craft. Your mission is to drive down a small corridor, walls on either side of you, until you get to the end of the line, where you will catch a spaceship which will take you off the moon. The mission is extremely dangerous: police ships are hard on your trail. One small miscalculation and you will go careening off the rails into one of the walls on either side of you. But the stakes are high: although you aren’t yet aware of it, the fate of the solar system depends on you.
And just like that you have the climactic scene from Rocket Science’s LOADSTAR: The Legend of Tully Bodine (1995). Not surprisingly, this interactive work of fiction was originally written by Ron Cobb as a feature film.
This comparison is somewhat simplistic, and should not be overblown: “Stories are not games. The experience unique to narrative is not the same thing as the experience rendered by Mortal Combat or roller coaster rides. Story writers have a lot more to do than simply string together a series of obstacles which pose a hurdle for a character seeking an objective.” [111] Comparisons of action films to video games give us some useful general ideas, but how can we apply these to more sophisticated narrative forms, to romantic comedies and Greek tragedies? In this chapter, I will look at theories about various aspects of film and see what lessons they hold for writers of interactive fictions that are primarily visual rather than text-based. In so doing, I hope to come across guidelines which will help writers interested in creating more complex characters and stories than those found in action/adventure videos.
But as I do, let’s not become snobbish about video games. As David Sheff points out, by the early 1990s, Nintendo, a single game producer, “netted as much as all the American movie studios combined and profited more than any of them, and more than the three networks combined.” [112]
The Evocations of Montage
Although there is a small number of works of interactive prose, most examples of interactive narrative occur in visual media, largely video and computer games. To get a sense of how interactive stories can best be told, it is important, therefore, to understand the specific properties of visual narrative. Since there is a rich body of theory on film, that will be the focus of this section.
The basic building block of film is the shot. A shot can be considered a single image, although it is more technically correct to say that a shot is everything in a single, uninterrupted camera set-up. A man sitting on a park bench. A couple sitting at a table sharing breakfast. A toe. The White House. A sled being held tightly to a child’s chest. Whatever other information the aural content of a shot might convey (and it may be considerable), every shot in a film should contain a wealth of visual information. However, such information does not, of itself, begin to create meaning in the mind of the person looking at it; the context for understanding a shot is in the shots which preceded and come after it.
As early Russian director and film theorist Vsevolod Pudovkin explained it, the basic significance of editing individual shots together is to guide
the attention of the spectator now to one, now to the other separate element. The lens of the camera replaces the eye of the observer, and the changes of angle of the camera – directed now on one person, now on another, now on one detail, now on another – must be subject to the same conditions as those of the eyes of the observer. The film technician, in order to secure the greatest clarity, emphasis, and vividness, shoots the scene in separate pieces and, joining them and showing them, directs the attention of the spectator, compelling him to see as the attentive observer saw. [113]
Of course, this is a simplification of the way we see things; the eye sometimes wanders; it’s constantly shifting focus; it often turns on a subject too late to see the event which interests the person. Like all art, film is an ordered, selective representation of chaotic natural processes. Nonetheless, Pudovkin’s argument is very useful: shots should combine to give the viewer a sense of the unfolding scene, as if the viewer were there to
see it for him or herself. For Pudovkin, this is how an audience creates meaning out of individual images on celluloid.
Russian director Sergei Eisenstein very much opposed Pudovkin and his theory, saying that it reduced filmmaking to “brick by brick…” construction. [114] Eisenstein argued that the meaning of a series of shots was not simply the sum of the information conveyed within each of the individual shots, but that the interplay between shots carried additional meaning. “Everyone who has had in his hands a piece of film to be edited knows by experience how neutral it remains,” he wrote, “even though a part of a planned sequence, until it is joined with another piece, when it suddenly acquires and conveys a sharper and quite different meaning than that planned for it at the time of filming.” [115] Eisenstein believed that two concrete images, when juxtaposed, created additional intellectual [116] and emotional [117] meanings not contained in the original images.
This is the basis for what has come to be known as “montage” theory. Evidence for this theory was not long in coming:
In his most famous experiment…[Russian filmmaker Lev] Kuleshov took unedited footage of a completely expressionless face (that of the pre-revolutionary matinee idol Ivan Mozhukhin, who had emigrated to Paris after the revolution) and intercut it with shots of three highly motivated objects: a bowl of hot soup, a dead woman lying on a coffin, and a little girl playing with a teddy bear. When the film strips where shown to randomly selected audiences, they invariably responded as though the actor’s face had accurately portrayed the emotion appropriate to the intercut object. [118]
That is, the audience found meaning in the combined shots which it would not have found if it had just seen the individual shots. Kuleshov concluded from this experiment that a shot contains “two distinct values: 1) that which it possesses in itself as a photographic image of reality, and 2) that which it acquires when placed in relationship to other shots.” [119] Kuleshov went on to experiment with a more complex arrangement of shots where he “cut together a shot of a smiling actor with a close-up of a revolver and a second shot of the same actor looking frightened. Audiences naturally interpreted the sequence as portraying cowardice, but when Kuleshov reversed the position of the two shots of the actor within the sequence, the opposite interpretation was made.” [120]
This has serious consequences for anybody working within a visual interactive medium. At the simplest level, when you give your participant a choice of paths, the first shot in each of the paths must combine with the last shot before the choice to convey the meaning you want to create for the participant. In a typical video game, the choice may be to shoot an alien, which leads to: a) a dead alien if you choose to shoot; b) your death if you choose not to shoot, or; c) a stand-off if you choose to run away. Each choice creates a momentary emotional response in the participant, whether it is exultation, dread or the vague uneasiness of a confrontation postponed.
Interactive drama offers the potential for the creation of more complex meanings between shots than a shoot-em-up, of course. These may include: various kinds of visual, aural and emotional contrast; parallel actions; symbolic combinations; simultaneous (though not parallel) events, and; leit-motif (reiterations of a visual theme). [121] Other methods of creating meaning will likely be found as creators experiment more with interactive media.
How best to determine which shots should follow a branching choice? A useful analogy here would be that individual shots are the equivalent of words and that sequences are the cinematic equivalent of sentences. If you were writing an interactive sentence, you would be sure to link the beginning to choices which were of the same part of speech. So, for instance, if you wanted to give the reader of the sentence “The boy jumped over the bright red moon” a choice of protagonists, the word “boy” would be linked to other nouns (ie: cow, girl, lounge singer, etc.). If the choice were of actions, the word “jumped” would be linked to other verbs (climbed, searched, spit).
Putting sequences of scenes together in order to create a longer narrative follows a similar logic: “A consecutive sequence will appear upon the screen only if the attention of the spectator be transferred correctly from scene to scene.” [122] In traditional linear films, the creator has complete control of the order in which shots and scenes are experienced, and puts them together in order to create a specific emotional effect and meaning. When writing an interactive story, on the other hand, it is important to understand how the meaning created by choosing paths through different scenes will affect the participant’s experience. Moving between scenes with radically different emotional content (for instance, from a comic scene, to a scene of graphic horror, to a scene of tender emotion) is likely to confuse the participant, leading to an unsatisfactory esthetic experience.
The sequences which make up a work must be planned, therefore, so that meaning and emotional content evolve out of the choices given the participant. This will vary depending upon its structure. A work which contains threads leading to multiple endings can assign different meanings to each thread (as in a mystery which leads to different solutions), or even different emotional qualities (it is possible to create a drama or a comedy from the same scenes which establish the premise of a story). To do this, the scenes which are shared by all threads should remain relatively neutral, while the scenes unique to each thread should carry the bulk of the work’s emotional/intellectual meaning. (Remember: neutral scenes become coloured by the emotional and/or intellectual weight of scenes which follow, so, if used effectively, they are remembered as part of a coherent whole.)
A work which contains a single ending is more complicated. Occasional scenes which go against the grain of the larger work (for example, scenes of comic relief in dramas) are a tradition in film and theatre; there is no reason to believe they wouldn’t be effective in interactive media. However, if they were accessed from multiple points in the story, or if they led out to multiple points in the story, the writer would have to ensure that each link prepared the participant for the shift in tone. If you are using a Cul de Sac Structure, the entire loop off the main story may have a different tone or meaning than the story, as long as the departure point from and entry point to the main story are carefully constructed. An Open Structure would be a difficult one in which to include a shot or scene which goes against the emotional/intellectual grain of the work since the writer has the least control of the paths which will lead to and away from it, so it has the potential to be the most jarring to the participant. (As well, the creator will have to deal with the fact that, the more scenes which go against the predominant mood or meaning of a work which are contained therein, the more that mood or meaning is undermined. If you keep adding comic scenes to a drama, it eventually becomes a comedy.)
One additional problem which creators of interactive work will face which creators of traditional works don’t is that of repeated shots and scenes. In many interactive stories, the participant can retrace his or her path, in theory continuing to come across the same scene over and over again, and often in a different order, something which isn’t possible in non-interactive stories, where the director determines the rigid order of the scenes. As Pudovkin’s second experiment showed, the order of shots creates meaning; coming upon shots in a different order can create meanings which the creator of a work hadn’t intended. A small number of films (such as Catch 22, 1970, or All That Jazz, 1979) effectively use repeated scenes to emphasize their emotional impact; in Catch 22, in particular, fragments of a single scene are returned to throughout the narrative, building to a horrific climax. This suggests that repetition can have great esthetic value.
On the other hand, many complaints about video games revolve around the fact that scenes without much emotional impact or more information than can easily be grasped in a single viewing are repeatedly experienced; this is a structural flaw which lessens the esthetic value of the work. A creator must ensure that a scene which is
repeated contains enough emotional weight or intellectual content that viewing it repeatedly will not soon become boring.
The most obvious way of dealing with repeated scenes is to structure your narrative so that no paths turn back on themselves. This is relatively easy in a Cul de Sac Structure, where every scene propels the narrative forward and the branches which run off the main narrative cannot be accessed once they have been experienced.
Another method of dealing with repeated scenes is to carefully structure a work, ensuring that the participant only experiences scenes in an order which makes sense. Carefully structuring the work can also ensure that only the scenes which offer a wealth of visual, emotional and/or story-oriented information are repeated, while scenes with a single, simple meaning are accessed by fewer branches.
For me, however, the most interesting possibility is to use the potential ambiguities and contradictions inherent in multiply traced paths within the narrative. Say, we open with a scene which seems straightforward enough, but subsequent scenes offer contradictory information; a branch which took us back to the initial scene would allow us to see it through new eyes. Using this model of later scenes recasting what we thought we saw in earlier scenes throughout a work could lead to a complex, involving experience for a participant.
Character
Traditional Views
One of the most important aspects of narrative dramas is character.
The best and strongest plots…are those which do evolve naturally, even inevitably, out of characterization. These stories have more impact, because they are more believable. There is a good reason for this. In the lives of most of us, very few important things happen for totally external reasons; what happens to us is often the result of what we do – and what we do is often a result of who we are. That is true of you and me and your potential viewers. If it is also true of your characters in what you make happen to them, they will be perceived not as concoctions, but as living characters with a dimension of depth and reality.” [123]
For a story to be involving, the characters have to be believable.
Fictional believability is not, however, reality. Consider a typical man. He likely acts in different ways depending upon the role which he is called on to play: at work, he may be ruthless, hard, even deceptive or dishonest; at home, he may be gentle, kind to his wife and children and scrupulously honest. He may find that he has conflicting goals: he’s taken a job at an advertising agency to be able to support his family, but he would really prefer to be blowing sax in a jazz band because that satisfies him emotionally. He (or she, for the same general observation applies to women) is complicated, constantly acting upon conflicting desires and goals, often displaying radically different emotional states.
With rare exceptions, such complex characters are not created in film. All art is a reduction of reality, of course; the amount of the reduction, notwithstanding the esthetic sensibilities of individual artists, is determined by the amount of information which can be conveyed by the medium. A novel, for instance, contains a large amount of information, which allows for the exploration of characters in great depth. A feature film, generally lasting between 90 and 120 minutes, does not have as much information and cannot treat characters in as great depth. But even the novel, with its much greater amount of information, can only hint at the development a character has undergone over the course of a lifetime.
In the dramatic narrative arts, many ideas of character development have arisen to deal with the limitations of various media. While these lead to the creation of characters with varying degrees of complexity, they all tend to follow Aristotle’s dictum that “In the characters too, exactly as in the structure of the incidents, [the poet] ought always to seek what is either necessary or probable, so that it is either necessary or probable that a person of such-and-such a sort say or do things of the same sort, and it is either necessary or probable that this [incident] happen after that one.” [124]
At its most basic, character is seen as a consistent set of traits which define a person. A hero, for instance, will always be strong, intelligent, loyal to his cause, etc. etc. This concept of character is tied to the belief that human nature is simple and definitively set, ideas which have been contradicted by modern psychiatry. Nonetheless, simple characters have their uses (such as in adventure stories where character is secondary to the action). Interactive narratives pose no difficulty to the creation of such characters, since they will exhibit the same traits regardless of the order in which events take place.
These character traits are revealed through what the hero does: “The essence of character is action. Actions are the windows into the character’s mind, heart, and soul. The character’s change/growth is precipitated by actions and revealed through actions.” [125] Once the character has been established, his or her growth through the forward motion of a work of art is one of its most engaging aspects. However, change is difficult to effect in characters with a single dimension. Most often, it isn’t even attempted, leaving the audience with a character with whom it is difficult to sympathize, let alone identify. When it is, such change is often tacked onto the end of a work in an unconvincing way.
Another way of looking at this subject is to give a character a small number of clearly defined but conflicting traits. “It’s this inner clash, coupled with the outer, that gives real depth and interest to a story. And you get it by choosing a central character who has climax potential – that is, a character who, though driven well-nigh compulsively by desire, still can feel the pull of another, contradictory emotion.” [126]
There are different ways of developing such contradictions. Screenwriting teacher Margaret Mehring suggests that characters be developed with two levels of goals to achieve: plot goals, which are what people physically work towards in the forward motion of the story, and; personal goals, what psychologically drives the character. Often, these two levels of goals can conflict, as when a pacifist (personal goal) must defend his family against violent criminals (plot goal). To be effective, “The goals have to be important and the protagonist must be compelled, no matter what the obstacles, to pursue those goals.” [127]
Swain and Swain write that character is made up of three elements: point of view (“a character’s subjective outlook on a given topic”), dominant attitude (“their total approach to life”) and interests. [128] These elements may openly contradict each other, or may interact in subtle ways in response to the actions of other characters in the working out of the story.
Each of the conflicting traits within a character must be properly developed in order to make the him or her credible. If this is achieved, they will satisfy Aristotle’s dictum that every aspect of a character be necessary. As I like to put it, a good writer creates believably consistent characters; a great writer creates believably inconsistent characters. Moreover, this can lead to a much more satisfying form of character growth: in the end, the character’s inner conflict is resolved when he or she allows a recessive trait to become dominant (as when a coward finds an inner reserve of strength and saves the day).
Because developing more complicated characters often involves precisely timing revelations of their traits, they may seem unsuited to interactive narratives, where nothing can be precisely timed. In practice, this may not be true. A general model would suggest that the conflicting character traits be established early in a work, with the conflict developed over the course of the subsequent action; this is certainly possible within most of the interactive structures we have looked at so far. The order in which the traits are established, for instance, is immaterial, as long as their legitimacy – that is, their importance to the character – and the conflict between them is clear. Different paths through a work may give different relative weights to traits, making a resolution between them less clear.
Since interactive narrative art has been mostly confined, to date, to action/adventure games, such issues have not been in the forefront of artists’ thinking. If, however, interactive media are to move towards more adult-oriented work, how to develop believable characters in interactive settings will be a crucial issue. As consultant Anne Hart says: “Interactive multimedia needs character-driven stories and scripts to drive the technical stuff.” [129]
Voice
Me. You. He, she or it. Anybody who has gone through grade school English will recognize the three kinds of voice: first, second and third person. Traditional dramatic arts, with a small number of invariably unsuccessful exceptions, are told in the third person: the audience passively watches as the events happen to somebody else. Interactive media allows the participant a much broader range of points of view:
In interactive storytelling, the user can be involved in three basic ways: 1. Observer. The user doesn’t control characters, but may select one plot path or another at preset story branches. As in Where in the world is Carmen Sandiego? [1992], the user can sample different viewpoints in a single story thread. 2. Director. The user controls the decisions, speech, or behaviour of one or more characters. This is the style of King’s Quest [1989]. 3. Actor. The user becomes a character (usually the central character). Myst [1993] functions in this way. [130]
The first and second examples are different ways for participants to experience an interactive story in the third person; the third is an example of a first person point of view.
First person narratives dominate current interactive works. There are some strong arguments for why this is so. “The way [a] situation develops will depend entirely on the personality of the characters,” Charles Platt explains.
This is how fiction works: the writer creates characters, and the characters create the story. Now suppose the story becomes interactive. In other words, the reader is allowed to control the story. This means that the characters have been robbed of their most important function. They have become pawns who do whatever they are told. How can the reader respect them or believe in them anymore? There’s an obvious answer – take [a character] out of the picture and give the reader that central role instead. Now the story is about you so it makes sense for you to decide what to do…” [131]
Rachel Reese goes further, claiming that not only is first person narration in interactive fiction more logical for the characters, but more satisfying for the participant(s). Referring to Brenda Laurel’s paper “Interface as Mimesis,” Reese states that “adventure games are more fulfilling when the user is led to feel like s/he is experiencing the adventure first person, rather than giving order to a system which then tells him/her the consequences of his/her orders on the action behind the scenes. And as far as allowing the user to feel more caught up in the world of the game, her notion of ‘first-personness’ seems to work.” [132] Events which happen to you, and over which you have some control, are more involving than events which happen to others, even if you have some control over them.
First person interactive narratives have drawbacks, as well. “The only way this setup can remain convincing to a wide audience,” Platt writes, “is if the central character (namely, ‘you’) becomes simple and generic. This way, the role will fit as many different readers as possible. Characters…who have complex emotional problems are seldom seen in interactive entertainment. Instead, we find simplified heroes with very basic motives: to find a reward and avoid getting killed.” [133] This may be inherent in the medium, as Platt claims, but it may also be because the medium has been employed primarily for adventure games to this point. As we shall see below, filmgoers and users of other media can identify with complex characters who are not like them; the question is, is becoming a character in interactive media so different from identifying with a character in traditional media that the writer must sacrifice depth in creating him or her?
The experience of role playing games suggests that people want to become other characters. As screenwriter/interactive media producer Michael Backes says, “The real future for some of this stuff lies in having people trying on different faces.” [134] This point of view is supported by the work of Sherry Turkle, who argues that computers allow people (in the case of her research, children and adolescents) to “try on” various personality traits and identities. [135] Ultimately, creators of interactive works might have to create first person characters with a large variety of levels of complexity which participants could utilize to differing degrees depending upon their sensitivity and imagination, their ability to become another person.
Another problem with first person narratives is that you are restricted to giving the participant information that his or her character could reasonably be expected to know. If your character is busy stopping a bank robbery, he or she cannot know that the villain is actually kidnapping somebody at the same time; rather than experience the kidnapping, he or she will have to be told about it. In this way, first person narration can often limit what a person directly experiences in a story (or force the creator to resort to unconvincing contrivances to ensure that the main character is where he or she has to be at the precise moment key elements of the narrative are being played out).
Third person storytelling does not have this limitation; its objective veneer is actually a glorious advantage.
The development of a type of narrator specific to fiction – the omniscient, impersonal narrator – has freed fictional discourse from the constraints of real world and pragmatically credible human communication. The disembodied consciousness of the impersonal narrator can apprehend the fictional world from any perspective (external observer point of view or character point of view), adopt any member of the fictional world as focalizer, select any spatial location as post of observation, narrate in any temporal direction (retrospectively, simultaneously, even prospectively), and switch back and forth between these various points of view. Fiction, like VR, allows an experience of its reference world that would be impossible if this reference world were an objectively existing, material reality.” [136]
Third person interactive narratives have the potential to give participants the opportunity to go anywhere, experience anything from any point of view and tweak situations to see how characters would react to slightly changed circumstances – who would want to pass up this opportunity to play god?
Finally, using the first person in interactive narratives creates the problem of how and when to make the participant aware of his or her character. In traditional media such as film, character traits are slowly revealed through the character’s actions; we learn who people are by what they do. In interactive media, however, a participant cannot decide how his or her character will act without some idea of who he or she is. In role playing games, characters are usually determined before the action actually begins; this would suggest that interactive media should provide some guide to characters, probably text-based, either in a booklet accompanying the work or on-screen before the work actually begins. I am extremely leery, however, of making something extrinsic to the narrative necessary for a participant’s enjoyment of it, and would want to limit this sort of introductory information to the barest minimum possible.
A more difficult, although I believe ultimately more satisfactory method of dealing with this problem would be to reveal just enough information about the participant’s character to allow him or her to make a preliminary decision, revealing more as his or her decisions became more complex. By carefully crafting the plot choices based on the amount of knowledge the participant has of his or her character at that moment in the work, the creator can give the participant enough information to learn about his or her character as he or she acts it out. As Reese puts it, an author should be subtle in revealing a ‘user character’s’ thoughts in order to “shape but not dictate the user’s experience.” [137]
Good arguments can be made for using both first and third person in interactive narratives. Knowing the pitfalls of both will make it easier for writers to employ either effectively; in addition, experimentation with different voices will give us a better idea of which are more effective for what kind of story. Ultimately, as I have stressed throughout this thesis, the decision about which voice to use should be based primarily on the nature of the story being told.
Identification
Unless the work is a sequel to or based upon a previous work, when we begin to watch a play or film the characters are strangers to us. Although we likely wouldn’t listen to a stranger who tried to tell us his or her life story on the bus, we do want to explore the life stories of characters in works of art. The skill of the storyteller undoubtedly has something to do with this. Part of this skill is giving us characters we can believe in, for, as writing teacher Stewart Bronfeld observes, there is a “simple but fundamental fact that people are primarily interested in people…” [138]
More than simply leading interesting lives, it helps to create characters who have personality traits or are experiencing conflicts which reflect the traits or conflicts of audience members (or traits which audience members would like to have or conflicts which they fear experiencing). “Good characters,” Margaret Mehring writes, “have universal qualities and worthwhile goals. They personify the qualities and goals of our culture. Who they are and what they want is common to all of us. Their needs are our needs; their fears we share; their joys we’ve known.” [139] This leads to more than mere curiosity on the part of audience members: it helps us identify with the character. In this way, the character’s quest becomes, for the duration of the work, our quest, his or her emotions become our emotions, his or her success or failure becomes our success or failure. It is through identification with a character that Aristotle’s idea of catharsis, the purgation of emotion through vicarious experience, becomes possible.
Is identification possible in interactive narratives? Certainly, third person narratives, where the participant does not take on the persona of a character, are, like traditional drama, more effective when the participant identifies with one of the characters he or she is following. “One great strength” of interactive stories according to designer Jim Gasperini, “clearly, is the medium’s ability to present the world from another person’s perspective.” [140] This would allow participants to experience a wide variety of different lives: “As time goes on, we may begin to see more personal uses of the interactive medium. One game might explore what it is like to be a woman at the edge of madness; another what it is like to be a man in the midst of a mid-life crisis; a third simply ‘what it is like to be me.’ We will see propagandistic uses: interactive Leni Reifinstahls [sic] creating strongly biased simulated worlds.” [141] As long as each contained enough general characteristics, it would be possible for a general audience member to identify with any of these specific interactive characters, and many, many more.
With some forms of interactive structure, especially, as we have already seen, parallel structures, it is possible to experience multiple perspectives on an action. This would give a participant a choice of characters with which to identify or, if constructed carefully enough, would allow a participant to identify with more than one character. In fact, by carefully crafting the possible paths a participant could choose, it would be possible for an interactive artist to manipulate a participant’s allegiance, having their identification shift from character to character depending upon the scene which is currently being enacted; this would be a useful technique to employ in, for example, psychological dramas.
With first person interactive narratives, where the participant “plays” one or more characters, the issue of identification is more complicated. “Allowing the phenomenon of character identification to occur in an interactive context seems to require the developer to create a ‘user-character’ and invite the user to interact within the framework of story through the eyes, hands and lips of that character,” claims theorist Rachelle Reese. “A writer of interactive fiction must develop a ‘user-character’ as conscientiously as a writer of paper-based fiction develops a protagonist, and the user should be filled in on how the ‘user-character’ thinks and feels about the things happening around him/her gradually, as the feelings of a protagonist might be revealed in a novel.” [142]
Is this really identification? It seems to me that taking on a new personality to play a character in an interactive work of fiction goes beyond the vicariousness of identifying with a fictional character and moves into the realm of direct, albeit mediated, experience. There is a fundamental difference between watching Gasperini’s various characters act, even if we have choices of what we are going to see (the third person model) and actually “being” those characters and having to respond to situations as we imagine they would (the first person model). Identification with a character may be an initial stage in the interactive process: we may not be likely to take on a character whose personality and goals are totally foreign to our own. However, once we have taken on the character, something more than identification holds our attention.
Artificial Characters
Computer generated characters will be necessary for most forms of interactive entertainment. This is obviously true of current interactive games, where a single player expects to interact with a world full of other people. With home systems where four or more people can play together on a local network, artificial characters are likely to be needed for stories which require more characters than there are players. (A different way of dealing with this question – having players take on more than one character – might be employed effectively. But the characters would have to be very clearly differentiated and separated in time or participants might find it difficult to jump from one to another.)
Even fully networked interactive experiences are likely to require artificial characters. Some characters which are considered necessary in order to tell a coherent narrative may be unattractive to participants, forcing the creator to take on those characters him or herself or create them artificially. In addition, creators may want to periodically introduce new characters into their narrative in order to add plot complications. Since they cannot be guaranteed that a new participant will express an interest in joining the work at the same time as they want to introduce the character, they will, again, be forced to take on the character or create it artificially.
Current practice, based on available technology, is to have the responses of characters not animated by participants prescripted and inserted at appropriate points in the narrative. If a participant decides to shoot a gun at an artificial character, it would trigger a previously decided upon response (either the character would duck out of the way or be shot, depending upon the needs of the story at that point as predetermined by the writer). This approach has a serious drawback: the number of actions which can be thought out beforehand is limited by the imagination of the artist, and the number of predetermined actions which can be stored in current delivery systems is even more limited, which seriously curtails the amount of choice you can offer a participant. You cannot allow a story to develop along certain lines if you haven’t created an artificial character response to them.
One possible solution to this problem lies in ongoing research in the field of Artificial Intelligence. According to pioneering researcher Marvin Minsky, “AI is the science of making machines do things that would require intelligence if done by men.” [143] The machines which may make AI possible are, of course, computers. For narrative arts, if computer-generated characters could be programmed to respond to a participant’s input directly rather than from a predetermined script, the dramatic potential of interactive narratives would be greatly increased.
There are currently two directions in Artificial Intelligence research. The first is the “top-down” approach, where the computer is given a goal and, employing a preset list of behavioural rules, must figure out how to complete it successfully. If the goal of a robot guided by AI is to move a block of wood from a corner of a room to the table on the opposite side, for instance, it would have to have rules telling it when to grasp the block of wood, when to release it, how to avoid bumping into walls, knocking over the table or otherwise successfully navigating the room, and so on.
This model of behaviour could have useful applications in the narrative arts: artificial characters have goals, after all, and behaviours. In drama, characters have a heirarchical series of goals which motivate their actions on a moment-to-moment basis as well as providing them with an overarching motivation for the entire work. For instance, a villain’s overall goal may be to take over the world; to do so, he must steal plutonium, which would be his sub-goal for an act; to steal the plutonium, he must find out where it is kept, what security there is, make plans and actually carry them out, a third level of goals which works itself out at the level of individual scenes.
Behavioral rules would determine how the character worked towards his or her goals. Giving a character traits such as intelligence and a sense of humor, for instance, would mean that he or she would be likely to respond to a threat with a joke; a cowardly disposition would mean the character would be likely to respond by running away; a vicious tendency would mean the character would fight; and so on. In addition, such rules would govern emotional states (a villain could be programmed to get increasingly frustrated, and correspondingly increasingly desperate, each time his or her plans were thwarted by the hero) and physical conditions (a character could become fatigued if physically pushed, which would affect its judgment).
An effective Artificial Intelligence system, to be useful as a tool for artists, would have to convince a participant that the artificial character was sufficiently complex to be an autonomous actor in the narrative.
At the center of any Artificial Personality system is an emulation of human emotions. Besides providing new motivation for believable behavior, emotions give the characters a new domain for discourse. They may interact on the levels of physical state, information state, and emotional state. In designing an interactive story, the designer must keep in mind the interlocking dimensions of physical state, emotions, character beliefs, behavior, and communication. One must also keep sight of the vision: characters displaying believable original behavior and engaging in interesting, dramatic interaction.” [144]
“Using this model, each character’s emotional state and current goals drive the selection of a specific behavior from a large set of possible behaviors. The intensity of the appropriate emotion values is then used to determine the intensity of the expression of the behavior.” [145] This set of possible behaviors would be much larger than is possible with predetermined character actions, which, in turn, would allow the participant in a work employing such characters a wider range of possible actions. It is worth noting, however, that creating believable artificial characters does not require the same criterion as the Turing Test, which was proposed to determine when artificial intelligences behaved in ways as complex as human beings. As we have seen, fictional characters are not as complex as real people; in fact, characters with a single psychological dimension are often sufficient for fictional works. Thus, an artificial character in a work of fiction needs only be convincing in its very limited context.
The need to choose from a preset list of possible behaviors places a less restricting, but still real kind of limitation on what an artificial characters can do. The other possible approach to Artificial Intelligence, “bottom-up,” is even freer in this regard. In this approach, a simple set of behavioral rules is believed to lead to increasingly complex behaviors. Rodney Brooks of the Massachusetts Institute of Technology (MIT) AI Lab claims that, “You can get intelligence…simply by stacking up reflexes until a system of behaviour begins to emerge…” [146] So, a villain would be programmed with villainous impulses (pleasure in other people’s suffering, will to power, et al); when confronted with a situation, the various individual parts of the program would interact to arrive at an appropriate action. Marvin Minsky, the original theorist of this form of AI called the negotiations between different behavioral rules the “Society of the Mind.” [147]
With this approach, an artificial character can have conflicting behavioral impulses, such as the need to kill an enemy and a belief in the sanctity of life. With a simple set of rules, such a conflict would end in deadlock – the artificial character would become paralyzed, unable to decide how to choose between actions of equal value. It is necessary, therefore, to assign each of the character’s traits a relative value; when two traits come into conflict, the one with the higher value would prevail (not dissimilar, in fact, to the way human beings make decisions).
To further complicate artificial characters, it is possible to set the values of character traits not as absolutes, but as numbers which fall within a range of values which change depending upon the input of the human participants. Using what is known as fuzzy logic, it is possible to have artificial characters change with experience. [148] This has the potential to make their behavior stunningly complex and highly unpredictable – in short, not much different from other human participants in a work.
Both approaches to artificial character generation, in their purest form, have drawbacks: the need for a list of behaviors puts a limit on the possible complexity of characters using a top-down system, while the competing traits of the bottom-up system does not guarantee an artificial character will give an appropriate, let alone dramatically meaningful response in any given situation. Some form of goal-oriented program which uses fuzzy logic in determining how to achieve those goals is likely to be the best means of applying Artificial Intelligence to dramatic narrative.
The ultimate goal is to mix characters generated by computer with characters who act as avatars, graphical representations of human participants, in a way that the human beings cannot tell the difference between the two. We are far from that goal at the time of this writing, but the theoretical groundwork offers much promise for the future.
“We’re not in Kansas any more:” Computational Drama in Oz
There is a specific kind of top down Artificial Intelligence which has potential bearing on interactive storytelling: it is generally known as an Expert System. These systems are loaded with rules by which they narrow a series of facts into a conclusion. Research on Expert Systems has concentrated on their use to the medical profession: if a patient has symptoms A, B, C, D and E, the system will tell you that he or she has illness a.
Earlier, it was suggested that the most effective interactive narrative structure would be one where the computer stored small story “bits” which would be put together in real time – an Expert System would be the ideal program to achieve this. As David Graves points out, “in order for characters to display behaviors that appear reasonable and believable, they must have their own motivations. These motivations help stimulate the generation of plot. However, without guidance for plot, chaos is the likely result. To ensure that the generated plot is interesting, the system could have some concept of drama and apply it to the unfolding story.” [149]
In the beginning, the world of possible actions (recall the Flying Wedge) would be the list of all of the narrative bits stored within the Expert System. There could be different programs corresponding to different genres (although they might share certain plot similarities, the list of bits for action/adventure films would be different than that for serious drama, which would be different again from light comedy). As the characters (both human and computer-generated) interacted, moving the story forward, the Expert System would eliminate possible narrative bits from future consideration, using rules of dramatic structure to continue developing a coherent narrative. As the story progressed, the system would continue to narrow the possibilities until only one outcome remained inevitable, at which point the story would, hopefully, be satisfactorily concluded.
An experiment along these lines has been conducted for the past several years; it is called the Oz project. According to Joseph Bates, one of the creators of Oz, “Our work applies existing artificial intelligence technology to the problem of building dramatic worlds. These worlds are composed of a (simulated) physical environment, intelligent/emotional agents which live in the world, a user interface and theory of presentation to let one or more humans interact with the world and its agents, and a computational theory of drama which plans and controls the overall flow of events in the world.” [150] Although text-based, the Oz project has much to teach creators of works of more graphically oriented interactive fiction.
According to Bates, making Oz worlds interesting meant “giving people the feelings that come with good stories, feelings that arise in part from the structure of plot, such as complication, climax, and resolution.” [151] Recognizing that this could not be done in traditional ways if people experiencing an Oz world were to be given real choices of action, he realized that “the Oz system must dynamically, and subtly, adjust the behavior of the world and its characters to provide experiences with the desired dramatic structures.” [152] As author Kevin Kelly put it, “In traditional drama, the narrative dictates both characters and environment. In Oz, however, the control is inverted somewhat; characters and environment influence the narrative.” [153]
Bates says a computational theory of drama is necessary. He offers two. The simplest comes from “having the author express a partial order on the significant events in a story (an ‘abstract plot graph’), explicitly representing the partial order in the system, and using it to drive the character goals and narrative decisions in the rest of the system. [154] Then he goes on to suggest that, “A richer approach is to develop a library of abstract plot units and then, as interaction proceeds, rapidly search abstract plot space for controllable paths that have the desired dramatic structure.” [155] This is the approach taken in Oz.
Kelly claims that the research being conducted on Oz worlds aims to answer three questions: “How do you organize a narrative to allow deviations yet keep it centered on its intended destination? How do you construct an environment that can generate surprise events? How do you create creatures that have autonomy, but not too much?” [156] Bates believes he has found the answer: “One of the claims of the Oz project is that the mental architectures and real world knowledge bases that have been developed in AI over the last 15 years, while perhaps still too weak for real robots, are well suited to the demands of interactive fiction.” [157]
Dialogue
“Where were you last night?”
“I think I’m catching a cold.”
“What…the – I mean, do you know how much…?”
“My nose – running. Could be allergies.”
“Dammit -“
“Yeah. Definitely.”
Most people do not speak the way they write. When we talk, we frequently use sentence fragments, our thoughts being broken and disjointed, and sometimes say empty phrases which don’t convey any real information. Furthermore, in conversation we often do not respond directly to what the person we are talking to has just said. No playwrights or screenwriters duplicate speech exactly as it is spoken; if you analyze their dialogue closely, you will find that even “naturalistic” writers use an idealized form of speech to animate their characters.
The most important thing for a writer to remember is that, “The primary function of dialogue is to convey information to the audience.” [158] This may be information about the story (although it is usually better to illustrate the story through action, conflict between characters can be very satisfying when played out through dialogue), about the characters (both in what they choose to reveal about themselves and what they choose to conceal, to whom they are willing to reveal themselves and how they express themselves) and about the broader issues or themes of the work. Remember: all works of art reduce reality – because you only have 90 to 120 minutes to develop complex characters or plot, you cannot afford to waste time with empty words or phrases (unless they are indicative of a trait of your character, such as evasiveness).
In order to pack the most information in small amounts of time, the idea of the “dialogue hook” was invented. Hooking requires that each new line of dialogue used by a character makes reference to the line of dialogue that was spoken before it before adding new information. As Swain and Swain claim, the dialogue hook “ensures that each speech acknowledges the one preceding it.” [159] In this way, dialogue in a scene is a constant flow of interlocked statements.
Reworked using hooks, the dialogue above might go something like:
“Where were you last night?”
“I think I’m catching a cold.”
“Don’t try to change the subject – I mean, do you know how much…?”
“But, my nose – it’s running. Think it could be allergies?”
“No, dammit, don’t you dare -!”
“Yeah. Definitely.”
Nobody’s going to mistake this dialogue for great art. Still, it illustrates the point: dialogue is easier to follow and seems to make more sense when hooks are employed to connect successive statements by two (or more) characters.
When creating interactive works, dialogue hooking is a useful technique to employ, especially in the transitions from one choice to another. Suppose you place a branch in the middle of a scene and offer the participant three options of where to take the scene. To help smooth over the break, each of the three segments should acknowledge the line preceding the choice using hooked dialogue.
Another use for dialogue hooking would be in the way you structure the responses of artificial characters to what human characters are telling them. In building a model for how artificial characters speak, a good place to start would be to allow for two elements: a response to what has gone before and new information. [160] Either of these elements could be as little as single word or as long as a monologue, but both have to exist in artificial speech to some degree.
In fact, there is a computer program which, in a simplistic way, shows how dialogue hooks can be employed to create meaningful conversation: ELIZA. In the mid-1960s, MIT professor Joseph Weizenbaum created a program which mimicked the responses of a Rogerian psychiatrist (the kind who repeats a statement made by the patient as a question in order to get the patient to go deeper into what is being said). A typical session with ELIZA might look something like this:
Men are all alike.
IN WHAT WAY
They’re always bugging us about something or other.
CAN YOU THINK OF A SPECIFIC EXAMPLE
Well, my boyfriend made me come here.
YOUR BOYFRIEND MADE YOU COME HERE
He says I’m depressed most of the time.
I AM SORRY TO HEAR YOU ARE DEPRESSED
It’s true. I am unhappy.
DO YOU THINK COMING HERE WILL HELP YOU NOT BE UNHAPPY
I need some help, that much seems certain.
WHAT WOULD IT MEAN TO YOU IF YOU GOT SOME HELP
Perhaps I could learn to get along with my mother.
TELL ME MORE ABOUT YOUR FAMILY
My mother takes care of me.
WHO ELSE IN YOUR FAMILY TAKES CARE OF YOU
My father.
YOUR FATHER
You are like my father in some ways.
WHAT RESEMBLENCE DO YOU SEE
You are not very aggressive but I think you don’t want me to notice that.
WHAT MAKES YOU THINK I AM NOT VERY AGGRESSIVE
You don’t argue with me.
WHY DO YOU THINK I DON’T ARGUE WITH YOU
You are afraid of me
DOES IT PLEASE YOU TO BELIEVE I AM AFRAID OF YOU [161]
And so on.
ELIZA does not understand what the human being is typing into the computer; it seems to be holding up its end of the conversation by using a few simple linguistic tricks, the most basic of which is that it repeats part of a statement back to the participant in the form of a question, sometimes inverting the word order, other times adding interrogatory words or phrases, yet other times picking a key word out of the participant’s previous statement and forming a question around it. Other features made it seem more sophisticated: sometimes, ELIZA will ask a question about the participant’s family if the subject hasn’t come up after a certain number of statements; if the participant makes a certain number of statements about one parent, but not the other, ELIZA will ask about the parent which has been ignored; if the participant makes three successive statements using words that the program recognizes as “negative,” it will ask why the person is being that way; and so on.
Because the rules are both simple and fairly rigid, it is easy to lead ELIZA into talking nonsense. The remarkable thing is, though, that such simple rules can lead to a conversation which makes any sense at all.
In fact, Weizenbaum created ELIZA as a joke on the community of artificial intelligence researchers: “The idea was to create the powerful illusion that the computer was intelligent. I went to considerable trouble…to explain that there wasn’t much behind the scenes, that the machine wasn’t thinking,” he wrote of a scientific paper which led to the creation of ELIZA. [162] The joke was on him, however, because many people were taken in by the program, calling it up for real advice, and for many years it was highly regarded by AI researchers.
For our purposes, ELIZA fails because it was not programmed to inject additional information into a conversation; it can only respond to what has come before it, it cannot initiate new lines of conversation. However, ELIZA is a powerful example of how dialogue hooks can, when used cleverly, create a powerful illusion of real conversation.
Willing Suspension of Disbelief
Art. Artifice. Artificial. The basis of any work of art is a very formalized unreality, yet, at its best, this unreal artifact can move us, can make us think, can cause us to reevaluate the way we view the world. When you consider the power inherent in the abstract symbols of a book or the frames of a film or the dots of a television screen, you can’t help but be astonished.
The artificiality of our mediated experience is mitigated by what has come to be known as the “willing suspension of disbelief:” “Non-actual possible worlds can only be regarded as actual through Coleridge’s much quoted ‘willing suspension of disbelief.’ The reader of a fiction knows that the world displayed by the text is virtual, a product of the author’s imagination, but he pretends that there is an independently existing reality serving as referent to the narrator’s declaration.” [163] When we sit in a theatre, we stop thinking about how uncomfortable the chair we’re sitting in is, or how we are surrounded by strangers, or any of the other details of the actual experience we are having; instead, we choose to believe that what is happening on the stage or screen is actually happening before us (this is known informally as “getting caught up in the action”).
Although not many artists appreciate it, the willingness of the audience to suspend its disbelief gives them a terrific advantage at the beginning of a work: the audience is ready to be entertained, the artist has to provide a truly uninvolving work to disappoint it. If Aristotle is correct about people seeking emotional catharsis through art, audience members have a powerful incentive to make believe.
In some ways, this is true of interactive media. “The question isn’t whether the created virtual world is as real as the physical world, but whether the created world is real enough for you to suspend your disbelief for a period of time. This is the same mental shift that happens when you get wrapped up in a good novel or become absorbed in playing a computer game. You stop considering the quality of the interface media and accept the computer-generated world as a viable one…” [164]
Anything which tends to remind a participant of the artificiality of a work will destroy the suspension of disbelief, undermining the participant’s esthetic appreciation of the work. Interactive media are more prone to this destruction of suspension of disbelief than other media: after all, what could be more artificial than periodically asking a participant to click a mouse or type in a command? The more choice a participant in a work is given, the more the person is required to input commands into the computer; each time a command is input, whether it is typed in or affected by a mouse or other device, the participant risks having his or her suspension of disbelief broken.
One possible way of dealing with this problem, paradoxically, is to have the branching choices come one after another so quickly that the participant literally does not have time to become consciously aware of the artifice; this is the preferred method of video games. It is, however, difficult to imagine this approach succeeding with an interactive film adaptation of, say, a Jane Austen novel; clearly, it is limited.
The example of video games does, however, suggest a general principle which creators of interactive art would do well to keep in mind: the more involved the participant is in the forward motion of the story (whether it be how a relationship develops or how best to beat up a kung fu opponent), the less intrusive the mechanical means of making decisions will seem. As designer Phil Saunders explains, drama can be created, and the suspension of disbelief maintained, in a work of interactive fiction by creating in the mind of the participant “The anticipation of what’s through the next door.” [165}
A possible secondary factor in maintaining the suspension of disbelief is the immersive quality of the medium; the greater the sensory input, the easier it is to believe in the artificial experience. This is a complicated area: early experience with virtual reality suggests that participants are willing to forgive the crudeness of graphic representations of a space if the interactions in the environment are, themselves, engaging. However, all other things being equal, it is fair to say that a more immersive environment makes suspension of disbelief easier. As Robyn Miller, designer of Myst, explains: “Our feeling was if a player feels like he or she’s playing a computer game, then we’ve lost something. They’ve lost that illusion that they’re lost in this world. I mean we pulled them back out of the world and we’ve placed them in front of a computer.” [166]
With the interactive art form known as virtual reality (which is described in more detail below), vicarious experience is replaced by direct experience. The pleasure we may some day get from VR will be less like the catharsis of art than the joy of discovering something for ourselves in real life. However, this new esthetic pleasure will still require a hearty suspension of disbelief; a bug in the system which causes the virtual world to black out every 30 seconds is likely to destroy a person’s belief in a virtual world very quickly.
Whatever the medium, one of the artist’s most powerful aids is the participant’s willingness to believe that a work is real.
Time
Time in a medium such as film is not the same as time the way we normally experience it. True, cinematic time can be used to describe the length of the film (this is commonly known as “running time”). However, within the 90 to 120 minutes of real time that the film is projected onto the screen, all manner of tricks can be played with the concept of time
.
“Filmic time,” screenwriting teacher Margaret Mehring writes,
is fragments of incidents selected to create an artificial period of time. Filmic time can condense events that would take two days or two decades of natural time in two minutes or two hours of actual viewing time. It can expand an event that would naturally take two minutes to occur into five or twenty-five minutes of viewing time. It can achieve a sense of experiencing different events simultaneously. It can mix time realities. Future events can precede present events and the past can follow the present. It can speed up, slow down, reverse, and freeze the natural motions of an event. [167]
Condensing time is the most common form of cinematic manipulation. Start with a shot of a person at the bottom of a flight of stairs, cut to a close-up of the person’s face, then end with a shot of the person arriving at the top of the stairs. What may have taken a minute or more in real life can be economically accomplished in a few seconds on film. This condensation of time is crucial because film stories typically take place over several days, leaving the filmmaker with no option but to find only the most important scenes in that period to show (but doing so in a way that they seem causally connected).
The Odessa Steps Sequence of Sergei Eisenstein’s Potemkin (1925) is often cited as an example of time being elongated in film. The main conflict in the scene is that a phalanx of soldiers walks down a staircase towards a group of unarmed civilians, who walk halfway up to meet it; the soldiers fire upon the civilians, causing great anguish and confusion. By cutting away to individual characters and cutting back to the two groups moving towards each other (until the shots are fired, at which point the civilians scatter), Eisenstein makes a confrontation which would only have lasted 30 seconds take up minutes of screen time. By stretching out the event, Eisenstein underlines both its horror and its importance as a catalyst for the Communist rebellion on the Potemkin.
Anybody who has ever seen a suspense film knows how events can be portrayed simultaneously: through a technique known as parallel editing. The hero rushes to stop the villain, who is enacting his plans for world domination; tension is created by cutting back and forth between the two scenes. Quentin Tarantino’s film Pulp Fiction (1994) is a good example of how time lines can be mixed up: the third segment actually takes place before the second, while there is a framing device in a diner which further complicates the placement of the segments in time.
In addition, single shots are timed for effect (long shots often force the viewer to attend to certain kinds of info, short shots often have heightened emotional impact). The rhythm between shots and the relationship of their lengths is often crucial to the development of a sequence.
All of these concepts should be part of cinematic interactive narratives. Such narratives complicate the issue of narrative time, however. Participants in interactive fiction can arrest the action at any point in time, either by pausing (where that is allowed by the program), making a choice which moves the action to a different place in the narrative or exiting the program entirely. Because of this, there will be occasions when the creator of an interactive narrative cannot control how time is experienced in his or her work.
This can sometimes be mitigated by creating shots or sequences which cannot be interrupted; the cursor, which the participant uses to initiate choices, disappears for the duration, leaving the participant with the choice of watching the shot or sequence or shutting down his or her computer. If the participant comes upon this shot or sequence several times during a work, though, he or she will quickly become frustrated by not being able to move beyond information which he or she has already seen (perhaps many times).
The best defense against a participant skipping sections, screwing up the timing of shots or sequences, is, of course, to make them so interesting that the participant will want to experience them in the order and for the duration that the artist meant for them to be experienced.
Interactive narrative also adds a new wrinkle to the experience of time in cinematic narratives: “dwell time.” [168] Dwell time is a useful way of understanding how we experience interactive narratives with worlds structures: instead of having a series of events which drive the narrative forward, we are placed in an environment which we are encouraged to explore. Unlike linear narratives, where the artist determines how much time will be devoted to each shot and sequence, in interactive narratives employing dwell time, the participant controls how long he or she will spend looking at a specific scene, controlling the length of the sequence and the interrelationship between the lengths of its parts. In this way, some control of the narrative is taken out of the hands of the creator and placed in the hands of the participant.
As the popularity of Myst, an interactive work which is structured as a series of explorations of environments, shows, dwell time can be very entertaining. However, the pleasure cannot be cathartic in the Aristotelian sense because it does not follow his causal chain of story events, nor does it require identification with the emotions experienced by a character in the work (in extreme cases, there may be no characters with which to identify). The pleasure of exploring a world in dwell time seems to me more akin to exploring a new environment in the real world: the surprise and delight of personal discovery. In this way, as we have seen, interactive narratives replace vicarious experience with direct experience.
Dwell time can also be used effectively in interactive narratives with more traditional structures. An environment to be explored at the leisure of the participant can be offered between levels of a game or acts of a story to give the participant time to assimilate what has gone before and prepare for what is to come (in the place of the dip at the end of acts in traditional three act narratives). In a narrative with a lot of action sequences, dwell time can be a useful way of revealing important information at a less hurried pace than in the ongoing narrative. I’m sure that, as artists experiment with dwell time, new and exciting artistic possibilities will emerge from it.
Openings and Closure
You know that feeling of satisfaction you get when you close the cover of a book or watch the closing credits start to crawl up the screen in a movie theatre? It’s a sense that everything you have just experienced has been part of a coherent whole, that all the pieces worked and led to a clear conclusion. Much of this feeling arises out of what is known as “closure.” At its most basic, closure is the sense that a work has properly ended, that it has neither lasted longer than the material warranted, nor ended with the main conflict or important issues brought up in the course of the narrative unresolved. This does not mean, as some people think, that an ending cannot leave questions unanswered; in fact, many narratives end with some form of ambiguity. Consider the case of Dr. Strangelove (1964). At the end, we cannot know if the human race will survive, or what form human society after a nuclear holocaust might take. Yet most viewers will feel that the film has ended in the proper place because the basic conflict – whether or not an American plane would successfully drop a bomb on Russia, triggering a nuclear device which would enshroud the world with a lethal cloud – was resolved. Going back to the idea of the flying wedge, closure occurs when we have reached the point in the story when what occurs has become a necessary outgrowth of what has come before, when the main conflict in a story has been resolved.
For an ending to be successful implies that the scenes which preceded it were. Paul Ricoeur has noted that successful closure “gives the story an ‘end point,’ which, in turn, furnishes the point of view from which the story can be perceived as forming a whole.” [169] In this way, closure is an important part of the appreciation of an entire work of art, for, as Barbara Herrnstein Smith writes in Poetic Closure, “a structure appears ‘closed’ when it is experienced as integral: coherent, complete, and stable.” [170]
For interactive works which attempt to maintain traditional three act structures, a single ending is necessary, greatly simplifying the question of closure. The only question is of the integrity of the work: when the participant arrives at the final scene or screen, will he or she be satisfied that he or she has experienced a complete work? This may not be much of a problem. In a cul de sac structured interactive work, for example, since all of the important narrative elements are on the spine, which every participant has to experience to reach the conclusion, the participant will receive a complete and coherent narrative, no matter how actively he or she pursues the offered interactive choices.
A reader moving straight through the first level to the final screen of a story with a parallel structure can feel that the story is complete, as can one who has bounced between levels before reaching the final screen. There is always the possibility that the former reader might feel dissatisfaction from the knowledge that different choices might have lead to a longer, richer reading experience; however, such a reader would likely have been motivated to make different choices in the first place. Every person can experience an interactive work in as active a way as he or she feels is worthwhile. In addition, any person who feels there is more to an interactive work can go back and, by making different choices, experience more of it. However, closure may be said to have failed to the extent that a participant feels the need to go back and reexperience a work because he or she feels the first experience was not, in itself, complete.
Structures which lead to different endings, such as simple branching, can also provide the participant with a sense of closure if, as we have seen, each of the threads has its own three act narrative arc and flying wedge. In this case, there would be several different points of closure, each corresponding to one (or possibly more) of the threads.
Because of their nature, web structures are the most difficult to imbue with a sense of closure. By definition, there is no end point to a web, just a series of connections between different points; because of this, there is no pre-defined point at which a participant can say, “I have experienced everything there is to experience and feel that the work is now complete.” As a result, most participants will continue to explore a story until they have become bored with it or frustrated by the fact that they seem to be going over the same material repeatedly; in either case, the experience will not be pleasurable. As theorist Stuart Moulthrop points out, although hypertext story space is “theoretically infinite…the lack of closure may be a theoretical strength but a practical weakness.” [172]
In non-fictional hypertext, the solution to this problem is to have a map of the material which the user can access at any time, giving him or her a clear idea of where in the information space the person is. This is awkward in fictional narratives, as the need to consult a map can destroy the participant’s suspension of disbelief. Occasionally, a map can be directly worked into the narrative (as is the case with the island setting of Myst), but this will only be possible for a specific set of narratives.
One innovative solution to this problem is to create areas in a work which offer “partial closure followed by continuation.” [173] A scene in an interactive drama, for instance, could seem to resolve the main conflict, while the following scene or scenes could renew the conflict, bring the conflict back in a new form, begin a new conflict or otherwise continue the story. This would give the participant the option of stopping if satisfied with the story or continuing if he or she wanted to see more.
There is precedent for this in serial novels dating back to the 19th century, of which Charles Dickens is perhaps the most well known writer. Each part of a serial was self-contained, with its own conclusion, and could, for this reason, be enjoyed on its own. Each successive chapter would continue from where the preceding one left off, and could be finished with the same sense of closure applied to the sum of the chapters up to that point. Using this technique, such screens or scenes would have to be carefully placed: following too close one upon another, they might seem obvious or contrived. The number and placement of closure/continuation points will vary depending upon the story being told and the temperament of the teller.
The author of such an interactive work would have to be careful to ensure that the continuation was a logical outgrowth of what came before; too radical a shift in the plot would make the narrative episodic, a series of events without a strong narrative through line. Aristotle took a dim view of such stories: “Among simple plots and actions, episodic [tragedies] are the worst. By ‘episodic’ I mean a plot in which there is neither probability nor necessity that the episodes follow one another.” [174] For creators aware of this problem, partial closure with continuation is one promising solution to the question of creating satisfying closure in web structures.
More experimentation with interactive narratives will hopefully uncover others.
Simulations have perhaps the simplest way of dealing with closure: when you have achieved the goal (whether it is landing an airplane or vanquishing a foe), you know the experience is over. Many simulations have different levels, each with its own goal and point of closure; these are a good example of the form of closure with continuation.
Worlds can be similar to simulations if a goal is established for the participant early in the work. The participant can be given a reason for exploring the environment(s), such as finding a treasure or solving a mystery, which means that the exploration will effectively (and hopefully satisfactorily) end when the goal is achieved. Where a narrative starts to emerge from the participant’s exploration (as happens in Myst), finding enough information to understand what is happening becomes the goal, with similar results. Of course, exploration can be a satisfying goal in itself, but without some other goal, works in which this is true tend not to provide satisfying closure.
One other method of closure is to set a time limit on the experience, which gives it a definite conclusion. (This has the added advantage of increasing tension.) Sports simulations are often timed when appropriate to mimic the sport simulated. Timed experiences will likely not work with simple structures because they may be perceived as an arbitrary limitation on the participant’s experience of the work (and, in any case, there are more effective methods of creating closure for these structures); they may be better for webs where there is no other clear ending (especially if the participant can determine exactly how much time to devote to the work, based on parameters suggested by the author). Whether this will prove effective with narrative interactive works has yet to be determined; I suspect it will only be helpful in a small number of very specific instances.
As you start, so shall you end. The beginning of an interactive work is an important aspect of a participant’s esthetic appreciation of the whole. Current practice in the interactive CD-ROM world is to begin a work with a non-interactive filmed or animated introduction. One survey of interactive CD-ROMs showed that each of the 12 reviewed began with a non-interactive introductory sequence. [175] As Landow writes: “Thus far, most of the hypertext fictions I have read or heard described…take an essentially cautious approach to the problems of beginnings by offering the reader a lexia labelled something like ‘start here’ that combines the functions of title page, introduction, and opening paragraph.” [176] Such beginnings are deemed necessary to orient the participant to the world of the work and to set up the conflict which will propel the action.
In most cases, non-interactive introductory passages may be unavoidable. Suppose you want to start an interactive work by giving the participant several choices of opening – you would have to have an initial screen from which to offer your choices. The one exception is web structures; not only do they have no clear end point, but they have no logical beginning, either. A program created for a web structured story could open it randomly, with a different starting point every time the person opened it.
Ultimately, we may simply not have had enough experience with interactive narratives to feel comfortable with such a possibility. As Landow points out, non-interactive introductions to interactive stories may occur as a result of “some writers’ obvious reluctance to disorient readers upon their initial contact with a narrative, and [because] some writers also believe that hypertextual fiction should necessarily change our experience of the middle but not the beginnings of narrative fiction.” [177]
It should be noted that some hypertext authors don’t see closure as an important problem. Michael Joyce, for instance, claims that: “Closure is, as in any fiction, a suspect quality, although here it is made manifest. When the story no longer progresses, or when it cycles, or when you tire of the paths, the experience of reading it ends.” [178] But Joyce reduces the definition of closure. It isn’t merely the end of participation in a work, as he claims; it is the esthetically pleasing end of a work. None of Joyce’s conditions seem to provide for such pleasure.
Constraining Choice
The cornerstone of interactive media is choice. However, complete freedom of choice is not possible, and may not even be desirable.
In the 1990s, the primary method of delivering interactive experiences was by CD-ROM. Compact Discs – Read Only Memory are plastic discs on which information is printed in the form of pits which can be read by a laser. Because the pits are embedded in the disc, CD-ROMs can be read by any player, but, unlike video or audio cassettes, they could not be written on by any but the most specialized equipment until the end of the century, when the technology became cheap enough for some individuals to buy it.
CD-ROMs hold approximately 650 megabytes of information. This translates into roughly 10,000 pages of plain text – enough for a substantial interactive adventure. As anybody who owns one knows, the upper limit of how much sound can be stored on an audio CD is about 70 minutes. If you want to store video on a CD-ROM, you can only get about 20 minutes on a disc (and the quality, in most cases, isn’t up to that of a television, let alone film); you can get slightly more animation on a disc because it has somewhat less detail than video. Research into ways of overcoming the physical limitation of CD-ROMs (for example, by compressing sound and images so that you can fit more onto the disc) is ongoing, but it seems fair to say that the medium will always have a serious limit.
This has repercussions for artists working in the medium. Some works repeat images and/or scenes in order to save on space (think of a simulation of a boxing match with a small set of punches, counterpunches and defensive moves which are repeated in a variety of different sequences). However, while some repetition of certain scenes in a dramatic work might be effective in highlighting their importance in the overall narrative, too much repetition would soon grow tiresome, so this technique may not be of much value in telling more sophisticated stories.
Another method of dealing with space limitations is to publish adventures on more than one CD-ROM. Again, while this may be effective with simulations (where breaks to change CDs can coincide with different levels of the simulation), you run up against problems when you try to employ this technique with dramatic material. Having to stop the action for mechanical reasons is a sure way of destroying the participant’s suspension of disbelief. One method of dealing with this problem is to put a hook at the end of each disc, similar to the hooks which come before commercial breaks in television shows, which will heighten the participant’s interest in learning what happens next. However, the criticism of the use of hooks in television is just as valid for CD-ROMs: they are an artificial device which are only intrinsic to a specific set of narrative genres. Thus, certain kinds of stories do not get told, or are told badly because of an inappropriate, though technically necessary intrusion in the flow of their narrative.
I believe that CDs were a transitional medium, and that the future of interactive storytelling lies in a computer network, whether it is the Internet or whatever succeeds it. In the Chapter Seven, we will see how computer networks make open narratives (which, you will recall, are stories in which the participant can actually contribute ideas rather than simply make choices from a pre-determined list) possible. For now, it is important to note that networks overcome the major limitation CD-ROMs have in creating closed narratives: storage space.
The storage space on your computer may not seem like much, but when hooked up to a network of computers all over the world, you actually have access to the space (not to mention computational power) of every computer to which you are connected. Networks, then, have the potential to deliver lengthy, complex closed narratives to interested participants.
They also solve the problem some people have with the solitary nature of CD-ROM experiences. As Rahmat and Giambruno wrote: “The hottest topic in computer gaming has to be anything with the word online in it, or multiplayer for that matter. If you’re going to play games, it’s better to play against other humans than against a computer. The best way to do that is either through an online service, modem-to-modem connection, or on the Internet. It just doesn’t seem feasible to invite a bunch of people to crowd around your computer screen.” [179] Or, as Kevin Kelly put it: “This is the call of the Net. Keep adding players.” [180]
However, even the much larger space contained by a network cannot solve a fundamental problem of interactivity: no creator or team of creators can anticipate every single choice which a hundred, a thousand or hundreds of thousands of individual participants may want to make in a given situation. [181] It is necessary, then, to place limits on what choices a participant in an interactive work can make. The logic of the story being told will place some constraint on the general direction of the narrative, as you might expect from an examination of the flying wedge: “The coherence of flight-simulation programs stems for instance from the fact that they exclude any choice of activity unrelated to flying.” [182] This means that, if you are in a flight simulator, you are constrained from solving a murder mystery or enacting a family psycho-drama. It also means that if you began a serious flight simulator, the plane cannot suddenly turn into a giant plate of asparagus; you are constrained by the logic which has been set up for the interactive world.
At the level of moment to moment choices, other constraints are necessary. Interactive adventure game designer Robyn Miller described the problem, and the solution he and the other creators of Myst came up with to deal with it, thus: “The islands idea came out of trying to find a way of having natural boundaries put around a person. If you put somebody into a city in a CD-ROM game, they want to go walk down any street that they see. If they can’t walk down that street, they feel gypped. Where if you put somebody next to an ocean, well that’s okay. So islands create a good natural boundary.” [183] In an interactive work based on the exploration of environments, you must anticipate that the participant will want to explore as widely as possible; in order to create such a work within the limitations you have, it is necessary to choose carefully which choices to offer and which to forestall.
“A major challenge in designing an interactive game,” said Phil Saunders, designer of Journeyman Project 2: Buried in Time (1995), “is figuring out how to block off access to areas you don’t want the player to go. Doing this in a way that doesn’t leave the player feeling frustrated can be really tough. Buried in Time‘s premise is that altering history is a no-no; therefore, coming in contact with other humans is to be avoided at all costs…a courtyard of soldiers prevents players from wandering around an entire medieval castle, a possibility that would have required gigantic amounts of memory.” [184] Disobeying this non-interference imperative brings the story to a definite, albeit entertaining conclusion, as producer/director Michael Kripalani related: “There’s a great example of this in the Mayan time zone. If you’re on top of the pyramid, you’re supposed to solve a puzzle to figure out how to get down into the catacombs. If you decide to just walk down, you hear a crowd sort of suck in it’s [sic] collective breath in awe, and it cuts over to a death scene of a stone Biosuit and a bunch of Mayans bowing to it as if you came from the heavens.” [185]
With both Myst and Buried in Time, the constraints on the participant’s choices were worked into the narrative, a terrific general rule to follow. This is a common occurrence which most artists, especially those in the performing arts, will recognize: making a virtue out of a necessity. It’s no different in essence from a first time filmmaker who writes a screenplay involving a small number of characters and settings knowing that he or she cannot afford more.
Another way in which Buried in Time limits the choices participants can make is by limiting interaction between characters, as Kripalani explained: “In the very beginning of the game, the first scene, you wake up, get out of bed, and walk into your living room. And boom, a TSA agent walks right up to you. It’s Agent 5, you, and he says, ‘Yes, I am you. I’ve come from the future. I’m going to send this suit over to you, just stand still and watch. You’re going to automatically cloak, don’t do anything.’ While you stand there, a second TSA agent comes in, has a full conversation with your future self, and they both time travel out.” [186] Although the temptation to limit interactions between characters is great because it is highly complex to do realistically, I think this form of constraint may lead to dramatically unengaging narratives. Kripalani admitted that, using this technique, “You don’t get to be an active participant.” [187] This seems to defeat the purpose of creating an interactive work.
There will always be tension between the need to give a participant in an interactive work of fiction meaningful choices (in order to create a proper sense of agency) and the need to work within the limitations of interactive media. It will be interesting to see how creatively artists will resolve this tension.