)URP1DUUDWLYH*DPHVWR3OD\DEOH6WRULHV7RZDUGD

3RHWLFVRI,QWHUDFWLYH1DUUDWLYH
0DULH/DXUH5\DQ

StoryWorlds: A Journal of Narrative Studies, Volume 1, 2009, pp. 43-59

(Article)

3XEOLVKHGE\8QLYHUVLW\RI1HEUDVND3UHVV
DOI: 10.1353/stw.0.0003

For additional information about this article

http://muse.jhu.edu/journals/stw/summary/v001/1.ryan.html

Access provided by University of California, San Diego (13 Feb 2016 22:48 GMT)
From Narrative Games
to Playable Stories
Toward a Poetics of Interactive Narrative

Marie-Laure Ryan

The genres of digital narrative are, if not innumerable

like the genres of narratives, at least very varied. They in-
clude stories generated by artificial intelligence (AI) sys-
tems, which have yet to produce narratives that people
would want to read for the sake of entertainment; hu-
man-generated stories, such as the news, gossips, or auto-
biographical sketches that circulate constantly through
the Internet; and 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. In this article I focus on the third kind,
more particularly on the design problem of integrating
the user’s activity into a framework that fulfills the basic
condition of narrativity: a sequence of events involving
thinking individuals, linked by causal relations, motivat-
ed by a conflict, and aiming at its resolution.
 Whether or not interactive narratives practically exist or are still chi-
meras depends on what is expected of the user’s participation: the more
active and the less constrained the user’s role—in other words, the more
lifelike, though life is not free of constraints—the more problematic its
integration into a well-formed narrative arc. For intensity of user par-
ticipation, freedom of choice, and depth of immersion, nothing can beat
the imaginary Holodeck of the TV series Star Trek, which Janet Murray,
in her 1997 classic Hamlet on the Holodeck, proposes as the model of the
new kind of narrative experience that digital technology will make pos-
sible. The Holodeck is a computer-generated, three-dimensional simu-
lation of a fictional world. 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. The Holodeck may
be the Holy Grail of New Media, but it is also, as Brenda Laurel calls in-
teractive narrative, “an elusive unicorn we can imagine but have yet to
capture” (2001: 72). It would take an artificial intelligence far beyond the
capabilities of existing systems to be able to process whatever the user
decides to do or say, and a creativity far beyond the imagination of the
best novelists and playwrights to be able to integrate this input into a
well-formed plot. (Imagine Shakespeare having to write Hamlet with-
out being in control of all the characters!)
At the other end of the spectrum of ease of implementation is hy-
pertext fiction, a genre that limits the user’s agency to selecting an item
from a menu of possible choices. What hypertext gains in actual fea-
sibility over the Holodeck, thanks to the simplicity of its algorithm, it
loses in ability to create narrative meaning and immersion in a fictional
world: narrative is a linear, causal sequence of events whose significance
depends on their position on a temporal axis, while hypertext is a net-
work of textual fragments that can be read in many different orders.
Unless the user’s choices are severely restricted, it is highly unlikely that
they will produce a sequence that respects narrative logic. This is not to
say that it is impossible for hypertext to tell stories; but if readers are to
construct a causal sequence of events out of fragments presented in a
variable order, they will have to do so by mentally rearranging the frag-

44 storyworlds volume 1  2009

ments into other configurations than the order in which they were ini-
tially presented on the screen.
The purpose of this article is to explore forms of interactive narrative
that offer a compromise between the unrealistic demands made on AI
by the Holodeck and the very programmable but narratively challenged
and interactively limited algorithm of hypertext fiction. In these moder-
ately interactive forms, the user manipulates one or more characters in
the fictional world and affects this world from within, rather than ma-
nipulating fragments of text from a metatextual perspective. My overall
purpose is manifold: to provide reference points and a vocabulary for
narratologists interested in the analysis of interactive narrative, to iden-
tify the problems that face the designers of interactive narrative, and to
survey some of the most viable solutions to what has become known in
the field as “the interactive paradox”: the integration of the unpredict-
able, bottom-up input of the user into a sequence of events that fulfills
the conditions of narrativity—conditions that presuppose a top-down
design. As Aylett and Louchart formulate the paradox (2004: 25): “On
one hand the author seeks control over the direction of a narrative in
order to give it a satisfactory structure. On the other hand a participat-
ing user demands the autonomy to act and react without explicit autho-
rial constraint.”

Narrative Games and Playable Stories

While narrativity is a type of meaning, interactivity, when put in the
service of entertainment, is a type of play. The combination of narrativ-
ity and interactivity oscillates between two forms: the narrative game, in
which narrative meaning is subordinated to the player’s actions, and the
playable story, in which the player’s actions are subordinated to narrative
meaning.1 Or, to put it differently, in a narrative game, story is meant
to enhance gameplay, while in a playable story, gameplay is meant to
produce a story. The concepts of narrative game and playable story re-
flect, in their opposition, the distinction made by the French sociologist
Roger Caillois between two types of game: ludus and paidia (1958/2001:
13). The best example of paidia games is building imaginary scenarios
with toys, using them, in the words of Kendall Walton (1990: 21–24), as

Ryan: Narrative Games to Playable Stories 45

“props in a game of make-believe.” These games do not aim at a specific
goal, and they do not lead to losing or winning. The pleasures of paidia
reside in the free play of the imagination, in adopting foreign identities,
in forming social relations, in building objects, in exploring an environ-
ment, and above all in creating a representation: paidia games are fun-
damentally mimetic activities. If there are rules, they are spontaneously
created by the participants, as when a group of children decides that a
certain tree will be the house of the wolf, and they can be renegotiated
on the fly. Ludus games, by contrast, are strictly controlled by pre-exist-
ing rules accepted by the participants as part of a basic game contract,
they lead to clearly defined states of winning or losing, and their plea-
sure resides in the thrill of competition and in the satisfaction of solv-
ing problems.
What I call a narrative game is a ludus activity. If there is one con-
tribution that digital technology has made to the design of games, it is
their narrativization. By this term I mean the transformation of what
used to be abstract playfields (such a chess boards and football fields)
into concrete fictional worlds populated by recognizable objects and in-
dividuated characters. The main difference between an abstract game
like soccer or chess and a narrativized video game like Half-Life, Max
Payne, or Grand Theft Auto is that in an abstract game the goals of play-
ers are only made desirable by the rules of the game (who would, in real
life, be interested in kicking a ball into a net, or in moving tokens on
a board?), while in a narrativized game the player pursues the kind of
goals that people may form in everyday life or in their fantasies: goals
such as saving the world from invaders and rescuing people in danger,
if you want to be a good guy, or stealing cars and killing people, if you
prefer villainous roles. But in the intensity of the action, players may
forget whether they are terrorists or counter-terrorists, space aliens or
defenders of the earth: in a narrative game, the player plays to win, to
beat the game, and story is mostly a lure into the game world.
While ludus inspires narrative games, the spirit of paidia infuses
playable stories. 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 evolu-
tion of the storyworld. Playable stories induce a much more aesthetic
pleasure than narrative games because the player is not narrowly fo-

46 storyworlds volume 1  2009

cused on goals. For me the essence of the playable story is captured by
what I once heard a little girl say about the game The Sims: “Guess what
I managed to do with my Sims? I made the father and mother drown
in the pool, and now the kids are alone in the house and they can do
whatever they want.” For this little girl, the pleasure of the game did not
come from reaching a state defined by rules, but in coaxing a good story
out of the system. The genres of playable stories include table-top role-
playing games (also known as Dungeons and Dragons), stories based
on decision trees, hypertext fiction, simulation games (like The Sims),
and interactive drama, a digital genre best described as an attempt to
implement the formula of the Holodeck. At the time of this writing, the
only working example of interactive drama is Façade (2005) by Michael
Mateas and Andrew Stern.

The Poetics of Interactivity

The Holodeck, as a whole, may be a castle in the air, but this does not
take anything away from the validity of its individual features as goals
to pursue for researchers and developers of interactive narrative. In this
section I propose to use three of these features as a point of departure
for a poetics of interactive narrativity, measuring them against the de-
vices, interfaces, and design philosophies of actually implemented forms
of narrative games and playable stories.

natural interface
Visitors to the Holodeck interact with the computer-generated world in
exactly the same way people interact with the real world: through lan-
guage and through the gestures of the body. This double means of inter-
action is essential to both social life and to its narrative representation
(or its computer simulation). Our relations to our physical environ-
ment consist mainly of corporeal actions, while our direct relations to
other human beings are based on semiotic transactions, though we can
also affect other people indirectly by performing physical actions that
will please or displease them. While narrative involves both kinds of ac-
tion, the interpersonal, language-based kind is the glue that ties togeth-
er the physical actions. A car chase by itself may be visually stunning,

Ryan: Narrative Games to Playable Stories 47

but it only becomes narratively meaningful if the chaser and the chasee
have reasons to behave the way they do, and these reasons can only be
established through some prior acts of verbal communication, such as
promising, threatening, concluding an alliance, or informing somebody
of somebody else’s plans. Narrative genres may differ from each other
in the prominence they give to physical action and in the complexity
of the interpersonal relations that motivate these actions, but extreme-
ly rare are the stories that focus entirely on problem solving through
physical actions. (Robinson Crusoe may be an exception, until the hero
meets Man Friday.) To rival the narrative richness of other media, then,
a system of interactive storytelling must be able to stage both physical
actions that change the fictional world and verbal acts that affect the
minds of its inhabitants and motivate them to take action.
Both natural language comprehension and whole-body interfaces
have been used in new media art and entertainment, but with limited
narrative efficiency. Interaction through natural language gives the user
a life-like freedom of expression, but there is no guarantee that the sys-
tem will understand the input and respond in a logically (and narra-
tively) coherent way, as the very limited success of computers in passing
the Turing test demonstrates.2 Natural language parsing has been used
in the interactive drama Façade, but it is only because the user plays the
role of a relatively disempowered witness in a domestic fight between
the synthetic characters that she can accept the system’s frequent in-
ability to understand what she is trying to say: inconsistencies can be
attributed to the uncooperativeness inherent to a quarrel and to the
self-centered nature of the synthetic characters. Another form of digi-
tal narrative that uses a natural language interface is interactive fiction
(IF), a purely text-based genre defined by Nick Montfort as “a program
that simulates a world, understands natural language from an interac-
tor, and provides a textual reply based on events in this world” (2004:
316). In contrast to Façade, which allows the user to type whatever she
wants but most of the time fails to respond adequately, the IF parser
sacrifices the naturalness of language by putting severe limitations on
the player’s use of syntax and vocabulary: most of the acceptable inputs
are two-word sentences (like “take knife”), and the parser openly rejects
any input that cannot be processed. This means that the player must

48 storyworlds volume 1  2009

learn the idiom of the system as a code within a code—almost like a
second language. The most extreme limitation in the use of language
occurs in menu-based communication. Many computer games (for in-
stance, Morrowind) allow the player to dialogue with the characters by
selecting an item from a listed of canned utterances. Here the user does
not have to learn a specialized idiom, and the system is able to respond
coherently to any choice, but the price to pay is a loss of fluidity, since
narrative time must stop until a choice is made. The menu also clutters
up the screen with an extra window that takes a toll on the player’s im-
mersion in the fictional world.
Whole-body participation, the other distinctive feature of Holodeck
interaction, is frequent in digital installation art and in sports games,
such as bowling or tennis simulations, or in dancing games. These sys-
tems track the motions of the user’s body and use this data in a variety
of ways: art installations may project an image of the body in a com-
puter-generated display, so that the users can see themselves in the vir-
tual world, while sports games calculate and display the result of the
player’s actions. But sports games and most digital art installations are
limited to a physical interaction with the world, and they do not present
the interpersonal relations that impart narrativity to a designed experi-
ence. The alternative to full-body interaction is the manipulation of a
control pad or the keyboard. In contrast to a full-body interface, this
mode of interaction rests on a non-iconic, arbitrary relation between
the gestures of the body in the real world and the events triggered by
these gestures in the fictional world. The range of actions that can be
performed by manipulating controls without interrupting the unfold-
ing of the events is again limited to physical action, such as moving the
player’s virtual body, picking up objects, and firing weapons. None of
these actions involves interpersonal relations, not even when players
kill enemies in video games: even though these enemies may take the
appearance of human beings, from the player’s point of view they are
mere bodies that need to be eliminated. It is only by resorting to a menu
based on language that the repertory of actions available to players can
affect other characters. In The Sims, for instance, menus allow players
to choose among actions such as “appreciate,” “irritate,” “dance with,”
“play,” “hug,” “kiss,” and “talk to,” all of which take a human being as

Ryan: Narrative Games to Playable Stories 49

patient. The choice of action affects the emotional relations between the
two characters by making the barometer of their mutual affection rise
or drop.
In both the cases of verbal and corporeal action, then, interactive
story systems have to choose between reasonably natural but narratively
limited modes of action and artificial interfaces that suspend game and
narrative time but that allow much richer interpersonal relations.

integration of user action within the story

Just as in real life all of our actions contribute to our life story, in the
Holodeck all of the actions of the visitor move the plot forward. This
degree of integration is almost automatic in a menu-driven system like
The Sims, since the player’s selection counts as the performance in the
fictional world of the action described by words on the menu. The cre-
ation of a narrative out of the user’s choices is facilitated by the nature
of the options: all the items on the menu involve interpersonal relations,
and the succession of choices writes the life story of the Sims family.
In a system with a natural-language understanding system, the degree
of integration depends on the efficiency of the parser: it was my feel-
ing while playing Façade that much of what I typed was not part of the
dialogue, and consequently of the drama, because the characters simply
ignored my input. But how can a story be created when the user’s pos-
sibilities of action are limited to moving, picking up objects, manipulat-
ing them, and solving riddles through this manipulation, as is the case
in shooter and adventure games? The most obvious way to handle this
problem is to choose a type of plot that puts great emphasis on physi-
cal actions. This explains why the archetypal narrative pattern described
by Vladimir Propp (1968) and Joseph Campbell (1968/1973) has been so
popular in computer games: a hero receives a mission, fulfills it by per-
forming various tasks, and gets rewarded in the end. The deeds of the
hero are relatively easy to simulate through the game controls, the basic
sequence of accomplishment-reward can be repeated endlessly, allow-
ing the player to reach higher and higher levels in the game, the script
lends itself to great variations in setting and in the nature of the tasks,
and the solitary nature of the hero’s quest makes interpersonal relations
dispensable. Even when the player needs the assistance of system-cre-

50 storyworlds volume 1  2009

ated characters or of other players to perform the tasks (as happens in
multi-players online games, such as EverQuest and World of Warcraft),
he advances in the game on his own, and other characters are usually
reduced to the fixed roles of either antagonists or helpers.
More intricate plots and backstories can be created through the use
of film clips (known in game jargon as cut scenes), but usually at the
cost of the integration of the user’s actions within the narrative. In the
worst of cases, the story only moves forward in those moments when
control is taken away from the player, and the player’s actions are noth-
ing more than means to unlock the next episode by solving problems
gratuitously thrown along the way to give him something to do. 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” (2000: 109).

dynamic creation of the story

In the Holodeck narrative machine, every action of the visitor affects
the life of his fictional persona, and every different choice leads to dif-
ferent stories. It would be impossible to store in advance all the conse-
quences of all the decisions that can be made by the player. The only
way for the system to deal efficiently with the visitor’s freedom of action
is therefore to compute the effects of her actions in real time, amending
its model of the fictional world, and responding to the player on the ba-
sis of this updated model. This dynamic process is known as a simula-
tion, and it imparts to its output a quality of emergence.
Existing forms of interactive narrative can be broadly divided into
bottom-up, emergent systems that create stories during the run of the
program, as does the Holodeck, and top-down systems that rely on pre-
scripted content. The former can be played many times, with different
results, while the latter are meant for a single traversal, since the story
does not renew itself.

Ryan: Narrative Games to Playable Stories 51

 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 this kind of system the choic-
es offered to the player are too numerous and the interactions of the
various objects too complex for the designer to anticipate all the pos-
sible developments. Bottom-up narrative design is a little bit like the
TV show Survivor: you throw a number of characters with well-de-
fined personality features together in a space, and you wait to see what
kind of story will come out of their interactions. If there is a drawback
to bottom-up systems, it is the lack of closure of their output: with-
out top-down authorial control, it is virtually impossible to create an
Aristotelian curve of rise and fall in tension, or a sequence of events that
stops after a conflict has been resolved. But closure is not indispensable
to narrative pleasure: throughout literary history, from the never-end-
ing Renaissance narrative of Orlando Furioso to the feuilleton novels of
Dickens, Trollope, or Eugene Sue in the nineteenth century and to the
modern TV soap operas, readers have time and again been fascinated
by narratives that go on and on, like life itself.
While the bottom-up approach is favored by playable stories, the top-
down approach is typical of narrative games, such as shooters and ad-
venture 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. There are two ways to create top-
down interactive narrativity. The most common is to start from a set of
problems to solve, actions to take, weapons to use, effects to create—in
short, starting from the design of gameplay—and to wrap this game-
play into a story. This is how, for instance, the game Prince of Persia was
created (Mechner 2007). The other method consists of starting from a
specific storyworld and inserting possibilities of user action to make it
interactive. We see this approach in games based on Harry Potter, The
Matrix, Lord of the Rings, or Alice in Wonderland. But because the plot

52 storyworlds volume 1  2009

of these games must be adapted to the possibilities of action offered by
game controls, it is usually fairly different from its literary or cinematic
source. Many of the games based on a pre-existing story tend to become
stereotyped shooters and quests, with weak integration of the player’s
actions into the storyline. These games attract players much more for
the spatial and visual pleasure of finding themselves in a familiar fic-
tional world and of encountering favorite characters than for the tem-
poral pleasure of enacting a specific sequence of events. In this kind of
design, storyworld takes precedence over story.
The top-down and the bottom-up approaches are not mutually ex-
clusive: scripted elements can be used in bottom-up systems to give
proper narrative form to the output, while top-down systems, as al-
ready noted, would not be interactive if they did not find a way to inte-
grate the bottom-up input of the user in their narrative arc. The Sims,
for instance, sparks interest by occasionally taking control away from
the player in order to stage pre-scripted scenarios that create unexpect-
ed turns of events (such as a male character being kidnapped by space
aliens and returning pregnant), while Façade, a basically top-down de-
sign in which the system-created characters take command of the plot
and bring it toward closure, manages nevertheless to make the dialogue
vary with every performance thanks to the player’s active participation.
Any future solution to the paradox of interactive narrativity will lie in a
novel combination of top-down and bottom-up design.

The Pleasures of Interactive Narrative

What kind of reward can we expect from active participation in a
story? Narrative pleasure can be generally described in terms of immer-
sion in a fictional world, though some kinds of pleasure lie in distan-
ciation. But a distinction should be made between ludic and narrative
immersion. Ludic immersion is a deep absorption in the performance
of a task, comparable to the intensity with which a mathematician con-
centrates on proving a theorem, or a soloist performs a concerto. This
experience is independent of the mimetic content of the game: players
can be deeply immersed in playing chess, go, football, or Tetris—all ex-
amples of abstract games—as well as in Second Life, Doom, or Cops and

Ryan: Narrative Games to Playable Stories 53

Robbers, all games with narrative content. Whereas ludic immersion
presupposes a physically active participant, narrative immersion is an
engagement of the imagination in the construction and contemplation
of a storyworld that relies on purely mental activity. Elsewhere (Ryan
2001) I described three kinds of narrative immersion that relate to dif-
ferent facets of the storyworld: spatial, temporal, and emotional. To
these three forms I now would like to add epistemic immersion.3 These
four types of immersion present variable degrees of compatibility with
the physically active stance of ludic immersion.

spatial immersion
Thanks to the visual and animation resources of digital media, spatial
immersion is the easiest to achieve in an interactive environment. The
ability of digital systems to adapt the display to the position of the play-
er’s virtual body makes them very efficient at simulating movement as
an embodied experience. Whereas hypertext navigation consist of in-
stantaneous jumps from one node to another that deny the existence of
space, the graphic engine of three-dimensional video games and online
worlds creates a smooth evolution of the landscape when the user’s ava-
tar travels from one point to another. Add to this kinetic experience the
visual power of computer graphics, which can make landscapes so beau-
tiful that the exploration of the fictional world becomes an end in itself.
The spatial affordances of the medium are most efficiently exploited in
epic narratives—stories about the travels of a lonesome hero through a
landscape full of dangers—as opposed to dramatic narratives that focus
on human relations. This adds another explanation (besides the adapt-
ability to game controls mentioned earlier) for the predominance of the
quest pattern in video games. But spatial immersion is not only a matter
of experiencing space through movement; it can also consist of an emo-
tional attachment to a certain location. This feeling, which is known to
geographers and phenomenologists as “sense of place,” is also very com-
patible with interactivity. Visitors to online virtual worlds spend lots of
time building themselves a personal retreat, a nest lovingly decorated
with customized objects made with the tools of the system or bought
from other players. While this phenomenon is not strictly narrative, it
echoes the importance of the setting for readers of novels: many people

54 storyworlds volume 1  2009

select narratives on the basis of where the action takes place, and ac-
cording to the cognitive psychologists Mandler and Johnson (1977), set-
ting is the most easily remembered narrative component.

epistemic immersion
The prototypical manifestation of epistemic immersion—the desire to
know—is the mystery story. This effect is relatively easy to achieve in
an interactive environment. The player impersonates the detective and
investigates the case through the standard repertory of computer game
actions: moving the avatar through the game world, picking up tell-
tale objects, and extracting information from system-created characters
through menu-driven dialogues. In this design, as Henry Jenkins (2004)
observes, the fixed and therefore non-interactive narrative of the past is
embedded in an interactive game world, in which the player enacts the
narrative of the investigation.

temporal immersion
Next on the scale of compatibility with interactivity comes tempo-
ral immersion, which, as Meir Sternberg (1992) has shown, includes
three narrative effects: curiosity, surprise, and suspense. The first two
are relatively unproblematic. When participation takes the form of spa-
tial exploration and leads to unexpected discoveries, its motivation is
curiosity, and its reward is surprise. A bottom-up system like The Sims
can create surprise by taking control away from the user and generat-
ing pre-scripted events, such as the abduction scenario described above.
Suspense is much more resistant to interactivity, because it requires
long-range planning by the system and top-down management of the
player’s expectations. Like epistemic immersion, suspense is created
by an intense desire to know, but while epistemic immersion concerns
events that already happened, suspense is focused on the future. People
experience suspense when they can foresee two or more possible devel-
opments 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. And if the system generates an
accidental event to prevent the player from fully controlling the out-
come of the events, the effect will be surprise rather than suspense.

Ryan: Narrative Games to Playable Stories 55

emotional immersion
The combination of emotional immersion with interactivity is the most
problematic of all because it involves interpersonal relations between
the player and computer-operated characters. In real life we experience
two main types of emotions: those directed toward ourselves, and those
directed toward other creatures through a vicarious experience known
as empathy. Self-directed emotions concern our desires and the success
of the actions through which we try to fulfill them. Even when these
emotions involve feelings toward others, such as love and jealousy, the
other is an object in a bipolar relation determined by the desires of the
experiencer. Not so with empathy: it is by mentally simulating the situ-
ation of others, by pretending to be them and imagining their desires as
our own, that we feel joy, pity, or sadness for them.
Narrative has a unique power to generate emotions directed to-
ward others. Aristotle paid tribute to this ability when he described the
effect of tragedy as purification (catharsis) through feelings of terror
and pity inspired by the fate of characters. By contrast, 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. But their range is much smaller than the
self-centered emotions of life: computer game players may fight to res-
cue a princess, and they may receive her hand in reward, but unlike the
heroes of love stories, they are not motivated to act by romantic feelings.
Whereas narrative interest regards characters as persons, ludic interest
regards them as means to an end. Exceedingly rare are the computer-
controlled characters who not only serve a functional role by helping or
hindering players in the pursuit of their goals but also generate interest
and empathy through their own personality. Michael Nitsche mentions
a game situation that involves both emotional involvement and a sense
of guilt in the player. At some point in Deus Ex players must kill a for-
merly friendly game-controlled character who has turned into a zombie
in order to progress in their goal, but some players develop such attach-
ment to this character that they experience extreme discomfort at treat-
ing her like an object that needs to be eliminated.

56 storyworlds volume 1  2009

 While narrative games deliberately sacrifice characters to action,
playable stories have been able to create characters sufficiently lifelike to
generate emotional reactions only by limiting the player’s participation
and hindering self-centered feelings. A case in point is Façade. The in-
teractor experiences intense feelings of dislike and contempt for Grace
and Tripp, a professionally and socially successful couple whose seem-
ingly perfect marriage turns out during the dialogue to be a mere façade
that hides a deeply fractured personal relation. While the user’s agency
allows variations in the dialogue that expose diverse facets of Grace and
Tripp’s personalities, the drama unfolds according to a relatively fixed
script imposed top-down by the system. The interactor may hate or de-
spise Grace and Tripp, but unlike the player of a competitive game or
a participant in an online world, she does not entertain strong feelings
for her avatar, such as caring for her character’s personal relationship
to Grace and Tripp. I certainly did not experience sadness over a lost
friendship when the couple expelled me from their apartment at the
end of the evening to sort out their problems between themselves.

Conclusion
With the epic quest structure of most video games, interactive media
have mastered what could be the oldest form of narrative (or at least
the oldest form of fictional narrative, for gossip must be older): the
struggle of the individual against a hostile world. With the networked
structure of hypertext, its fragmentation into recombinant units, and
its rejection of the linearity inherent to chronology and causality, inter-
activity has made a contribution to the postmodern deconstruction of
narrative, much more than to the construction of a postmodern narra-
tive. 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. Some steps in this direction have been taken with games in-
terspersed with filmic clips, with dialogue systems (Façade), and with
simulation algorithms (The Sims), but in all of these approaches, the
involvement of the player remains peripheral: with film clips he relin-
quishes 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 per-
sonally playing a role in the story. The greatest hurdle to overcome,
if interactive narrative is to combine the self-centered emotions that
come from our active engagement in life and games with the other-
centered emotions of traditional narrative, is the creation of personal
and evolving relations between the user’s character and the synthetic
agents, relations that allow the user not only to interact verbally with
these agents but also to influence their destiny and to feel personally
concerned both for herself and for others.
If interactive narrative is a mountain to climb, we have gone a long
way toward the top in the past thirty years. Yet the hardest part of the
climb is yet to come, because the mountain is a pointed cone, like
Mount Hood, and not a rounded dome like Mount Rainier. But steep
mountains have notoriously been an incentive to alpinists: as Chris
Crawford declares, “To dismiss interactive storytelling on the grounds
that it hasn’t been done before is to reject the entire basis of the human
intellectual adventure” (2004: 50).

Notes
1. I adapt the term playable stories from Noah Wardrip-Fruin’s concept of “play-
able media” (2007).
2. A computer passes the Turing test (devised by British mathematician Alan
Turing) when a human judge interacting verbally with both a computer and a
human cannot tell who is who. Success in the Turing test is measured in terms
of how long it takes for the judge to make the correct identification; so far no
system has lasted as long as five minutes—the conditions needed to receive the
Loebner Prize in artificial intelligence.
3. To these four types of immersion one may add social immersion, an experience
exclusive to multiplayer online worlds. In these worlds players can perform the
quests presented by the game only by forming alliances and sharing resources
with other players. These alliances, known as guilds, lead to a strong sense of
belonging to a virtual community.

58 storyworlds volume 1  2009

Works Cited
Aylett, Ruth, and Sandy Louchart (2004). “The Emergent Narrative: Theoretical
Investigation.” Proceedings of the Narrative and Interactive Learning Environ-
ments Conference, NILE04, Edinburgh, Scotland. 25–33.
Caillois, Roger (1958/2001). Man, Play and Games. Urbana: U of Illinois P.
Campbell, Joseph (1968/1973). The Hero with a Thousand Faces. 2nd ed. Princeton:
Princeton UP.
Crawford, Chris (2004). Chris Crawford on Interactive Storytelling. Berkeley, CA:
New Riders.
Jenkins, Henry (2004). “Game Design as Narrative Architecture.” First Person: New
Media as Story, Performance, and Game. Ed. Noah Wardrip-Fruin and Pat Har-
rigan. Cambridge, MA: MIT P. 118–30.
Laurel, Brenda (2001). Utopian Entrepreneur. Cambridge, MA: MIT P.
Lazzaro, Nicole (n.d.). “Why We Play Games: Four Keys to More Emotions without
Story.” www.xeodesign.com/xeodesign_whyweplaygames.pdf.
Mandler, Jean, and Nancy Johnson (1977). “Remembrance of Things Parsed: Story
Structure and Recall.” Cognitive Psychology 9: 111–51.
Mateas, Michael, and Andrew Stern (2005). Façade. Interactive CD ROM. Down-
load at http://www.interactivestory.net/download/.
Mechner, Jordan (2007). “The Sands of Time: Crafting a Video Game Story.” Second
Person: Role-Playing and Story in Games and Playable Media. Ed. Pat Harrigan
and Noah Wardrip-Fruin. Cambridge, MA: MIT P. 111–20.
Montfort, Nick (2004). “Interactive Fiction as ‘Story,’ ‘Game,’ ‘Storygame,’ ‘Novel,’
‘World,’ ‘Literature,’ ‘Puzzle,’ ‘Problem,’ ‘Riddle,’ and ‘Machine.’” First Person:
New Media as Story, Performance, and Game. Ed. Pat Harrigan and Noah
Wardrip-Fruin. Cambridge, MA: MIT P. 310–17.
Murray, Janet (1997). Hamlet on the Holodeck: The Future of Narrative in Cyberspace.
New York: Free P.
Nitsche, Michael (2008). Video Game Space. Cambridge, MA: MIT P.
Poole, Steven (2000). Trigger Happy: The Inner Life of Video Games. New York:
Arcade.
Propp, Vladimir (1968). Morphology of the Folk Tale. Trans. Laurence Scott. 2nd ed.
Ed. Louis A. Wagner. Austin: U of Texas P.
Ryan, Marie-Laure (2001). Narrative as Virtual Reality: Immersion and Interactivity
in Literature and Electronic Media. Baltimore: Johns Hopkins UP.
Sternberg, Meir (1992). “Telling in Time (II): Chronology, Teleology, Narrativity.”
Poetics Today 13: 463–541.
Walton, Kendall (1990). Mimesis as Make-Believe: On the Foundations of the Repre-
sentational Arts. Cambridge, MA: Harvard UP.
Wardrip-Fruin, Noah (2007). “Playable Media and Textual Instruments.” The Aes-
thetics of Net Literature: Writing, Reading and Playing in Programmable Media.
Ed. Peter Gendolla and Jörgen Schäfer. Bielefeld: Transcript Verlag. 211–80.

Ryan: Narrative Games to Playable Stories 59