The core of my proposed project is to build a "story engine" that uses story fragments, created by game designers, and assembles them into more complex arcs, woven into the game setting and reacting to the players' actions. To this end, it is not only important to look at what work has been done in this area, but also to consider the question of "what is game narrative?" and perhaps ponder the difference between "good" and "bad" narrative.
Search Terms Used
The academic literature on dynamic narrative is very thin this is hardly surprising, since academia has hardly even touched on the fundamentals of game design and dynamic narrative is a rather specialized nook. This is not to say that it's a particularly obscure topic Sony is currently investing heavily in Storybricks a VC startup devoted to generating emergent narrative from AI interaction (or something, it's kind of hard to pin it down) for the next generation EverQuest game in development.
That said, I think Storybricks are fundamentally wrong-headed (and I'm hardly alone in this Richard Garriot has expressed a similar view, having tried something similar himself). I don't think the goal is impossible, but I do think that simulating one's way to it is ridiculous, since it involves implementing successful AIs as a pre-requisite, and if you can do that, you have bigger fish to fry.
This is a case where practitioners of the art (i.e. game designers) are way ahead of theorists (i.e. academics discussing game design). As such, while I have found a lot of articles that touch on game design and narrative, none of them seem just based on their titles terribly on point. I think that looking at actual games past and present is more fruitful.
Aristotle's Poetics (translation by S. H. Butcher) is frequently alluded to in several of the papers I've read (below). It's refreshing to read discourse on literature that immediately tackles subjects, such as "how to do it" and "what makes it good" which have been set aside in favor of increasingly incomprehensible waves of literary theory, including relativism, modernism, and post-modernism.
Some choice quotations:
Tragedy is the imitation of an action; and an action implies personal agents, who necessarily possess certain distinctive qualities both of character and thought; for it is by these that we qualify actions themselves, and these thought and character are the two natural causes from which actions spring, and on actions again all success or failure depends. Hence, the Plot is the imitation of the action for by plot I here mean the arrangement of the incidents. By Character I mean that in virtue of which we ascribe certain qualities to the agents. Thought is required wherever a statement is proved, or, it may be, a general truth enunciated. Every Tragedy, therefore, must have six parts, which parts determine its quality namely, Plot, Character, Diction, Thought, Spectacle, Song. Two of the parts constitute the medium of imitation, one the manner, and three the objects of imitation.
The components Aristotle lists map very well onto computer games (setting song somewhat aside). I would argue that computer games these days are heavy on spectacle, pay some attention to plot, diction (e.g. the GTA games are well-written), and occasionally character but thought by the player has been all but eliminated. Whereas in the 1990s a game might give the player a goal, but little guidance on how to complete it, today the player is either railroaded between events or, in a sandbox game, given clear directions to each subgoal.
These are papers I have been able to obtain in full either via direct download or via the UT Dallas library system. I've noted the file name I archived each under, but I am not placing them here because I don't want to violate the authors' rights.
Hunicke, Robin, Marc LeBlanc, and Robert Zubek. "MDA: A formal approach to game design and game research." Proceedings of the AAAI Workshop on Challenges in Game AI. 2004.
MDA stands for "mechanics", "dynamics", and "aesthetics" the pillars of a proposed formal structure for game design and research. MDA is analogized to Rules -> System -> "Fun". They propose that the designer sees the game as from the "mechanics" side (and thus the dyanamics and aesthetics sides more poorly). Similarly, they propose that the player sees the game as from the "aesthetics" side (and thus see the dyamics and mechanics more poorly).
Interesting because they propose something like a "theory of fun". Not really sure if this theory is much more useful than, say, Valve's philosophy when designing Half-Life figure out what's fun and put more of that in the game; figure out what's not fun and tear that out.
Their breakdown of a game design notably ignores content as a significant component, which makes me doubt the value of their model altogether. E.g. if one were to say movies comprise dramatic structures (mechanics), camera shots, technical capabilities such as editing, compositing, and special effects (dynamics), and aesthetics (the experience of watching a movie) but neglect to mention, say, script or plot, your model would seem pretty poor. It seems to me that content tends to be the first and last consideration of a game design, with the stuff they talk about merely being the means of realizing the goal.
There are, without question, games whose chief original component is a novel mechanic/dynamic (to use their terminology) such as Bejewelled or Tetris, but by far the majority of games largely or entirely recycle not only existing mechanics/dynamics, but even entire "engines" comprising fixed sets of mechanics/dynamics that are little different from entire families of games (first person shooters, real-time strategy games, sliding tiles, flight sims, etc.)
Ryan, Marie-Laure. "From narrative games to playable stories: Toward a poetics of interactive narrative." Storyworlds: A Journal of Narrative Studies 1.1 (2009): 43-59.
This paper is almost the dual of the MDA piece. It discusses content separately from the nitty gritty of game implementation (but does so explicitly; i.e. not failing to notice that the latter exists). By poetics she means literary theory.
Discusses the gamut of "digital narrative":
stories generated by artificial intelligence (AI) systems, which have yet to produce narratives that people would want to read for the sake of entertainment
interactive narratives produced through a collaboration between the machine and the user-- or, to be more precise, through a manipulation by the machine of human-produced data in response to the user’s input
This latter could cover the "dynamic story engine" I propose to build, but it seems more directed at traditional "static" or hand-made content (i.e. fixed story structures that the user interacts with). The key difference might be to think of the system as having multiple users rather than one, i.e. the person making story "atoms" versus the "player".
Also, why "digital narrative"? Later she refers to "interactive" which is correct and would have saved quite a bit of misguided and unnecessary categorization.
She describes the "holodeck":
The user is invited to step into this world, to impersonate a character, and to interact through language and gestures with synthetic (i.e., computer-created) agents. No matter what the user says or does, the synthetic agents respond coherently and integrate the user’s input into a narrative arc that sustains interest.
Then quotes Brenda Laurel:
Brenda Laurel calls interactive narrative, “an elusive unicorn we can imagine but have yet to capture”
It's this unicorn, or perhaps something suggestive of it, that I hope to capture.
a creativity far beyond the imagination of the best novelists and playwrights to be able to integrate this input into a well-formed plot
I don't think a "well-formed plot" is the goal, in much the same way that a successful role-playing campaign need not resemble a finely crafted novel.
If there is one contribution that digital technology has made to the design of games, it is their narrativization.
She hasn't played the right pre-digital games!
in an abstract game the goals of players are only made desirable by the rules of the game
Aside: what she calls an "abstract game" in essence comes down to what the MDA people have actually described.
story is mostly a lure into the game world
Again, I think this is wrong. Story is itself the attraction, and the "game world" is usually merely the minimal facade necessary to support the story.
She makes the distinction between ludus (Latin meaning "play" or "game", but which she seems to intend to mean "a game with rules") as opposed to paidia (unstructured play). She is principly interested in narrative paidia rather than narrative ludus, but the reasons for this are unclear and the categorization is not satisfying.
in a playable story there is no winning or losing: the purpose of the player is not to beat the game, but to observe the evolution of the storyworld.
I would argue that there is no clear divide between these things. Purely "abstract" games like Chess or Soccer can be enjoyed as an "experience" versus merely a competitive activity. Similarly, unstructured games can have more-or-less satisfying outcomes (e.g. role-playing games do not have a "winner" but there are definitely successes and failures).
E.g. the classic Holodeck scenario is something like Data pretending to be Sherlock Holmes, or Dax practicing Klingon martial arts. These activities can only be satisfying if the possibility of losing (e.g. not solving a mystery, or losing a bout) exists (or is at least believed to exist).
Similarly, there are clearly major competitive components to some "playable stories" role-playing games are rife with all kinds of weird competitive aspects.
Playable stories induce a much more aesthetic pleasure than narrative games because the player is not narrowly focused on goals
Has she heard people talk about outstanding athletic performances? Or discussing chess games?
In short, while it seems reasonable to discuss the narrative qualities of games (stuctured or not, competitive or not), to restrict oneself to the unstructured, non-competitive quadrant seems like a pointless exercise, especially when role-playing games (which are a major portion of the subject she wishes to examine) are in fact pretty structured.
Ryan places a lot of emphasis on "natural" interaction. And she devotes some attention to discussion of natural language recognition (failing to realize that we can't even do speech synthesis with emotional inflection) and body-language parsing. She places a lot of emphasis on dialog, suggesting that literature almost always is driven by dialog (with the exception of Robinson Crusoe before Friday shows up):
A car chase by itself may be visually stunning, but it only becomes narratively meaningful if the chaser and the chasee have reasons to behave the way they do
Indeed. For a narrative to be satisfying it must be bound to the world by cause and effect: why does the chase occur? What difference does it make how the chase turns out? A good point, but of course it's just bald assertion! And while books without dialog are very uncommon, video games without dialog are quite common, and often have strong narrative content Doom, System Shock, Myst, and on and on.
It's also important to understand that we may simply accept that a reason exists a Macguffin is a classic plot concept: something that's vitally important for reasons we don't need to know. Similarly, one can imagine a movie that comprises a single desperate chase where we never know who is chasing the protagonist (or who the protagonist is chasing) but simply get wrapped up in the desperation.
Myst comes up later:
In Myst, for instance, the player needs to pull levers, turn dials, find keys, and guess secret codes to be admitted to the next space, where she will find another page of the book that tells the past story of the fictional world. The game designer Chris Crawford calls this situation a “constipated story” (2004: 130), and the game critic Steven Poole wittily describes it as follows: “It is as if you were reading a novel and being forced by some jocund imp at the end of each chapter to go and win a game of table tennis before being allowed to get back to the story”
First I must note that all of this misses the point the "narrative" (I'd actually call it "back story") of Myst was almost completely irrelevant to enjoyment of the game. (Or to be more unkind, most of us prefer a "constipated story" with a good game, to a constipated game, which is how I would describe every Chris Crawford game I've played.)
In Myst the puzzles and presentation fantastic graphics, atmospheric sound and music were the game, the narrative merely one of several excuses to solve the puzzles. I remember playing Myst very well; I can't remember the story at all and that doesn't bother me. To me, as the player, the "story" might go something like "I wandered around in a maze of wooden boardwalks through a flooded forest, trying to figure out how to move on; eventually I realized that by paying attention to what seemed like background sounds, I could correctly channel water through pipes to unlock the maze".
By the same token, the more recently released game Monument Valley has some kind of back story which again does not matter. Insofar as many games have a story, it isn't the story the game designer wrote.
I think the concept Ryan is trying to grasp (and hints at earlier in the paper) is that a game should create its own reality where the actions available to the player seem unrestricted.
Unless the user’s choices are severely restricted, it is highly unlikely that they will produce a sequence that respects narrative logic
The key is for the user's choices to seem unrestricted; the user must tacitly accept the restrictions that do exist to be "the laws of the universe". Do you feel restricted in your inability to teleport? Fly? Grow money on house plants? The games restrictions must feel like the "laws of nature" to the player. When you play Asteroids, the lack of ability to travel three dimensionally doesn't seem like a restriction.
On the other hand, the fact that certain "quest givers" in a typical MMORPG are unkillable or exhibit highly strange behavior (e.g. stopping in fights to give rewards to other players, or running off in mid-conversation to kill a wandering monster) is jarring for many players. Likewise, the idea of a heroic fantasy adventurer who can slay dragons single-handed being stopped short by a locked wooden door or a small chest is ridiculous.
Dynamic Creation of the Story
This would appear to directly bear on my proposal. Ryan talks about top-down vs. bottom-up:
The bottom-up approach is illustrated by The Sims. The program creates a world full of things and characters. Each of these objects is linked to a set of possible behaviors, listed on a menu that comes to the screen when the user decides to play with this particular object. When a behavior is selected, it brings another state of the fictional world, and another set of behaviors becomes available. As the world passes from one state to another, a story is created.
In other words the game exhibits behavior. The behavior is the story. This is only a small part of what I am proposing. (If you pick a lock you might succeed or fail that's the game's behavior and story might be affected by the outcome.)
the top-down approach is typical of narrative games, such as shooters and adventure games. In this approach there is no event generation on the fly. The player’s progression is a journey along a path that is already traced and that leads to a fixed destination, or to several destinations when the system offers branching points.
This is what I am proposing to replace with a dynamic story engine that creates stories on-the-fly in response to player actions and the context of those actions.
The top-down and the bottom-up approaches are not mutually exclusive: scripted elements can be used in bottom-up systems to give proper narrative form to the output, while top-down systems, as already noted, would not be interactive if they did not find a way to integrate the bottom-up input of the user in their narrative arc.
This somewhat suggests what I propose as being a hybrid of top-down and bottoom-up, but I would argue that the bottom-up concept is simply intrinsic to any activity, including life itself. Insofar as any game has behavior it generates stories ("I ran past the defender and tried to shoot at goal with my left foot, but slipped and the shot went wide"). There is no pure top-down approach; "bottom-up" is omnipresent. So rather than creating a "hybrid" approach, we're discussing extending or enhancing the top-down approach.
Any future solution to the paradox of interactive narrativity will lie in a novel combination of top-down and bottom-up design.
Originally I reacted badly to this statement, but upon reflection I think that it is correct but poorly expressed. Since bottom-up narrative is a given, it's more of a matter of the top-down design being able to respond to player actions and discovered intentions (i.e. organic bottom-up narrative), and the design of game mechanics that facilitate such interactions (e.g. make it easier for the top-down dynamic narrative to notice actions, discover intentions, and so on).
Pleasures of Interactive Narrative
Bears on my goals to make games more enjoyable and replayable, but not by making them merely addictive.
All of her "pleasures" end with the word "immersion". Perhaps it can be omitted!
- Spatial I think she simply means enjoying the visceral quality of the game, spatial or otherwise
- Epistemic finding out the solution to the mystery, or what happens next
- Temporal turns out she means suspense and surprise (rather than, um, time e.g. losing track of time, which frequently happens when playing games in my experience at least)
People experience suspense when they can foresee two or more possible developments and are dying to find out which one of these paths the story will actualize. But when players can determine the path through their choice of action, the uncertainty is lost,
This seems like poor analysis. I'd suggest that suspense is most effective when people feel that desireable outcomes are impossible or unlikely, but undesireable outcomes are likely or inevitable, and is most satisfactorily resolved by an unforseen but plausible solution. (Similarly, resolution via an implausible solution, highly predictable solution, or deus ex machina are all generally unsatisfactory.)
- Emotional being affected by the impact of events on oneself and others in the game
This does at least suggest some things one might measure in attempting to measure enjoyment. I think the obvious meaning of "temporal" might actually be more useful than the weird definition she ascribes to it. E.g. if you tell players to play a game for "about five minutes" and then measure how long they actually play, this might be a useful data point. Simply measuring how much time players spend doing different things also seems potentially enlightening.
the emotions we experience while playing games—excitement, triumph, dejection, relief, frustration, relaxation, curiosity, and amusement (Lazzaro n.d.)—are overwhelmingly self-directed ones, because they reflect our success and interest in playing the game
Again, she's played the wrong games.
With the epic quest structure of most video games, interactive media have mastered what could be the oldest form of narrative
What remains to be conquered is the dramatic narrative, the type of plot that knots together several destinies into a dynamic network of human relations and then disentangles them to let characters go their own way.
This doesn't seem to me to be a terribly convincing description of a dramatic narrative, and it also doesn't seem to be the only thing remaining to be conquered, or even necessarily the most important thing.
the involvement of the player remains peripheral: with film clips he relinquishes agency while the plot is being knotted; with existing dialogue Ryan: Narrative Games to Playable Stories 57 systems he participates in a conversation rather than in a plot, or if there is a plot, he is confined to an observer role; with a simulation he holds the strings of the characters like a puppet master, without personally playing a role in the story.
This is a concrete and very good objection to the state of the art of game narrative. Why did we need to wait for the "conclusion" for this?
She quotes Chris Crawford way too much! And the mountain metaphor: ick!
Hendrikx, Mark, et al. "Procedural content generation for games: A survey." ACM Transactions on Multimedia Computing, Communications, and Applications (TOMCCAP) 9.1 (2013): 1.
This paper lays out a thorough overview of the scope and nature of content including scenarios and maps in computer games and procedural generation techniques. It also attempts to show how various kinds of content have been procedurally generated in various games, but suffers from spotty knowledge of games pre-2000 and careless categorization.
Zimmerman, Eric. "Narrative, interactivity, play, and games: Four naughty concepts in need of discipline." First person: New media as story, performance, and game (2004): 154-164.
In presenting these four terms (games, play, narrative, and interactivity), I am not creating a typology. The four terms are not mutually exclusive…
Paraphrasing his paraphrased definition of narrative: a state transition composed of a series of events leading to insight, expressed through a medium, constituted by patterning and repetition.
Not really sure most of that is necessary, but fine. He points out that the definition is satisfied by pretty much anything (e.g. a game of chess) and suggests that it's more useful to ask the question 'in what ways might we consider this thing (such as a game) a "narrative thing"?' Agreed.
Interactivity he teases out the difference between inevitable interactivity (reading a book is interactive) and "explicit interactivity" i.e. being forced to make a choice.
Play "the free space of movement within a more rigid structure. Play exists both because of and also despite the more rigid structures of a system." Meh.
Game "a voluntary interactive activity, in which one or more players follow rules that constrain their behavior, enacting an artificial conflict that ends in a quantifiable outcome." This definition is so terrible as to not be worth criticizing.
Overall, the article is simply not useful, despite a promising start.
Responses: Chris Crawford thinks the definitions are useful but do not, perhaps for political reasons, go far enough. Jesper Juul paraphrases Chris Crawford as having blamed the cinematic cut scene for radically raising expectations for production value, leading to the death of experiment. These both seem like reasonable observations.
Adams, Ernest. Game writing: narrative skills for videogames. Ed. Chris Mark Bateman. Independence: Charles River Media, 2007.
The PDF is only a table of contents. It seems from the contents to be a "how to do what we're currently doing" book, which is not especially interesting to me, since (a) I already know, and (b) I'm trying to get out of this trap.
Pearce, Celia. "Towards a game theory of game." First person: New media as story, performance, and game 1 (2004): 143-153.
This article is very interesting (not sure it actually leads to a "theory of game"), but the PDF I have is very confusing it's hard to tell which bits are the original article and which are responses to that article.
Whereas narrative theorists, academics, and those engaged in a critique of games are obsessed with narrative (my emphasis), many game players find narrative quite problematic. The largest controversy has to do with the use of "cut-scenes"… Many game players associate the idea of "narrative" with this type of enforced linearity (again, my emphasis), which is a throwback to the cinema.
What are much more interesting, and I think are proving to be the so-called "killer apps" of narrative, are various procedural forms of narrative, which combine various levels of metastory and story systems.
She goes on to argue that EverQuest and The Sims are both "procedural forms of narrative" in this sense. And given that World of Warcraft (essentially an even more popular and successful derivative of EverQuest and The Sims are the two most successful games in history, she seems to have been proven right).
The idea is that the story emerges as a direct result of social interaction. As with the Renaissance Faire… players enter a fully constructed three-dimensional world. Rather than selecting fixed characters, they select particular character roles.
The Sims is a cross between a dollhouse, a Tamagotchi, and the television program Big Brother.
Mary Flanaghan's response is very interesting:
Narrative, again, operates at a fundamentally different level in games than it does in other media. A game is most simply described as a framework for structured play. In most cases, this structure will include some type of goal, obstacles to that goal, and resources to help you achieve that goal, as well as consequences, in the form of penalties and rewards (which can often translate into obstacles and resources). … these elements create a generic deconstructed narrative structure of sorts.
and she ascribes this to Pearce (who is less clear):
Metastory: A specific narrative "overlay" that creates a context or framework for the game conflict.
Story System: A rule-based story system or kit of generic narrative parts that allows the player to create their own narrative content; story systems can exist independent of or in conjunction with a metastory.
(Pearce seems to argue that at least some games, e.g. MMORPGs and sandbox games, actually function as a "set of parts" as described above.)
A good game, even one without an obvious "storyline" (or metastory), while being played, will tend to follow something that resembles the emotional curve of a dramatic arc. A great example of this would be basketball…
The "drive-by" definition of "game" seems far more robust than other, far more long-winded definitions elsewhere.
Frasca, Gonzalo. "Simulation versus narrative. Introduction to Ludology" The video game theory reader(2003): 221-235.
Seems to lay out the intellectual landscape Ryan's article inhabits. (E.g. it explains ludos vs paidia, discusses Aristotle's lost book on Poetics. Although Frasca is not cited by Ryan, despite quoting the same people (often more extensively).
Laurel, as well as most "interactive narrative" supporters, focuses on Aristotelian closure as the source ofthe user's pleasure. The biggest fallacy of "interactive narrative" is that it pretends to give freedom to the player while maintaining narrative coherence.
This is a great point. I think that the idea that emergent/procedural/dynamic narrative should somehow correspond to expected literary templates seems to be an unnecessary and unhelpful requirement. To the extent that people require a specific emotional arc, they tend to recast their personal stories into those structures post-hoc, regardless of what actually happened. Stories are how we frame events, we don't need events to structure themselves into stories.
The choice between paidia and ludus structures is ideologically essential for a simauthor because both carry different agendas. The simulated world in ludus games seems more coherent because the player's goals are clear: you must do X in order to reach Y and therefore become a winner.
Combat is simply easier to manage, and provides its own natural structure.
I have just suggested a typology of simulation rules (manipulation rules, goal rules, and meta-rules) that can help us to better understand how the designer's agenda can slip into the game's inner laws. However, this typology is not exhaustive and could certainly be expanded (for example, by analyzing the ideological role of the interface rules or by examining the nuances between games which have both winning and losing scenarios and those where you can only lose). I am convinced that it will take us a long time to grasp the potential of simulation as opposed to narrative, mainly because we are so familiar and proficient with the latter. Simulation contests our notions of authorship and also the boundaries that we are used to [to] apply to works of art.
This is simply excellent (and caps off a long series of excellent observations). When we criticize a simulation (and most games are to some extend simulations) the design decisions are intrinsically political in ways we aren't well-equipped to grasp yet. (In other words, the rules and meta-rules of the simulation are, in themselves, ideologically loaded.)
Representation and narrative may still hold a lot of tricks in their bags, but the promise of the yet unexplored field of simulation and games is so vast and appealing that some of us can hardly wait to start experimenting with it. Whoever slowly walked back home after buying a long-awaited video game knows exactly the kind of excitement that I am talking about.
Lindley, Craig A. "Story and narrative structures in computer games." Bushoff, Brunhild. ed (2005).
Aylett, Ruth. "Narrative in virtual environments-towards emergent narrative."Proceedings of the AAAI fall symposium on narrative intelligence. 1999.
Barber, Heather, and Daniel Kudenko. "Dynamic Generation of Dilemma-based Interactive Narratives." AIIDE 7 (2007): 2-7.
Alvarez-Napagao, Sergio, et al. "Socially-aware emergent narrative." Agents for Educational Games and Simulations. Springer Berlin Heidelberg, 2012. 139-150.
Walsh, Richard. "Emergent narrative in interactive media." Narrative 19.1 (2011): 72-85.
Liu, Changchun, et al. "Dynamic difficulty adjustment in computer games through real-time anxiety-based affective feedback." International Journal of Human-Computer Interaction 25.6 (2009): 506-529.
Wei, Huaxin. "Structuring narrative interaction: what we can learn from heavy rain." Interactive Storytelling. Springer Berlin Heidelberg, 2011. 338-341.
Mateas, Michael, and Andrew Stern. "Interaction and narrative." The game design reader: A rules of play anthology 1 (2006): 642-669.
Ryan, Marie-Laure. "Beyond myth and metaphor: Narrative in digital media."Poetics Today 23.4 (2002): 581-609.
Simons, Jan. "Narrative, games, and theory." Game studies 7.1 (2007).
Gazzard, Alison. "Unlocking the gameworld: The rewards of space and time in videogames." Game Studies 11.1 (2011).
Smith, Greg M. "Computer Games Have Words, Too."
Juul, Jesper. "Games telling stories." Game studies 1.1 (2001): 45.
Miller, Carolyn Handler. Digital storytelling: A creator's guide to interactive entertainment. Taylor & Francis, 2004.
Murray, Janet Horowitz. Hamlet on the holodeck: The future of narrative in cyberspace. Simon and Schuster, 1997.
The following works showed up in Google Scholar and looked relevant, but could not be obtained without paying (and did not show up in the UT Dallas library system).