A Narrative Theory of Games
Espen Aarseth Center for Computer Games Research
IT University of Copenhagen 2300 Copenhagen, Denmark
+45 7218 5045 Aarseth@itu.dk
ABSTRACT This paper presents a narrative theory of games, building on standard narratology, as a solution to the conundrum that has haunted computer game studies from the start: How to approach software that combines games and stories?
Categories and Subject Descriptors K.8.0. [Personal Computing]: Games General Terms Theory, Design
Keywords Storygames, narratology, ludonarrative model
1. INTRODUCTION Computer games generate many questions and challenges for narrative theory. The very first humanist article on computer games, Interactive Fiction by Anthony Niesz and Norman Hol-land (1984: 125) talks about games as an enigma  to literary theory. Are games a type of narrative? If not, do they contain narratives? Is narratology useful for the study of games? Should the definition of narrative be modified or expanded to incorporate games? Can the study of games yield insights useful for narratol-ogy? Perhaps we instead need a parallel paradigm, a ludology (Frasca 1999) to understand games the way narratology is used to understand narrative? Regardless of the answers to these questions, the fact remains that computer games have emerged as a dominant cultur-al form in the sense that it influences other forms such as cinema, TV, literature, theatre, painting and music. Therefore, the study of these other forms ignores the rapidly evolving field of games at their own peril. But how should such theoretical explorations be carried out? When changing focus from one empirical field to another, it becomes crucial to examine the examination process and the tools used as well as the object of the examination. Do theoretical concepts such as story, fiction, character, narra-tion or rhetoric remain meaningful when transposed to a new field, or do they turn into empty, misleading catachreses, blinding us to the empirical differences and effectively puncturing our chances of producing theoretical innovation? Critical self-reflection is a hallmark of scholarship, and a necessary virtue when we examine a phenomenon with critical tools developed for another type of phenomenon altogether. In other words, when we
study games through the lens of narrative theory, the lens itself must be critically examined as well.
When we apply the perspectives and models from one form onto another, our ability to assess the incongruities as well as the similarities between the two forms becomes critical. Man is a pattern-finding animal. It is extremely easy to find parallels, pre-cursors, and points of overlap, and thus seduce oneself to con-clude that A is a form of B. The responsible theorist, therefore, should take the opposite position as their null-hypothesis: A is not a form of B unless proven otherwise.
Consequently, in the context of games and narratives, one must be careful not to assume the task of proving that games are narrative forms, but to look for evidence and counter-evidence with equal zeal. To further complicate matters, games, as Witt-genstein pointed out in Philosophical Investigations (1953: 66), is not a category that it is possible to define formally. So how can we know that the phenomena we call computer games are even games in the first place? Successful definitions of narratives has a very long history and appear to be easier to come by, so at least there is some fairly firm theoretical ground to stand on: Narrative theory.
As the study of computer games gelled into an emerging institutional practice around 2001, the question of whether games are a form of stories has become a prominent issue among the practitioners, a touchstone used to signal ones familiarity with the new scholarly field. The last ten years have seen a number of comments on the so-called ludology vs. narratology debate, but, ironically, very few have actually engaged the question of the relation between games and stories through a properly narratolog-ical analysis, using the basic concepts and models of modern narrative theory. Instead, this debate has been carried out on a meta-level, through comments on, and characterizations of, the debate itself rather than by direct engagement with it. This meta-debate can be seen as a symptom of the birth pangs of a new academic field. Too often the positions taken have been un-nuanced, untenable, and therefore unproductive: Games are always stories (Murray 2004, p. 2). [T]he computer game is simply not a narrative medium (Juul 1999, p 1).1 Tragically, in the field of game studies the term narra-tology has changed meaning and does not refer to the academic discipline of narrative theory, but to a more or less mythical posi-tion taken by an imagined group of people who are seen to believe that games are stories. It is high time, and hopefully not too late, to reinstate the original meaning and function of narratology, and ground the debate in narratological terminology and theory. Not to prove that all games are narrative (they are not) but to show that
1 A position from which Juul later wisely retreated (cf Juul 2001).
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. FDG '12, May 29-June 1, 2012 Raleigh, NC, USA. Copyright (c) 2012 ACM 978-1-4503-1333-9/12/05... $10.00.
there is much to gain from a rigorous application of narratology to game studies.
What has so far been lacking is a detailed, robust under-standing of the various ways computer software have been used to combine elements from narratives and games into a number of quite different ludo-narratological constructs. What also needs to be realized is that story-game amalgams primarily is entertain-ment software, works that contain many forms of media content and because of their computer-based, Turing-complete existence can emulate any kind of semiotic genre, including, of course, traditional stories. Calling works like Max Payne or FallOut 3 games or stories is a metonymical shorthand usage of the terms that confuses and obscures the composite makeup of these crea-tions.
2. Ludology vs. Narrativism The debate of whether games are stories has suffered from a lack of rigorous, theoretically grounded reflection and also from a basic confusion between normative and descriptive approaches. In reality this is not one, but two debates conflated: one is the design-oriented discussion of the potential and failings of game-based narratives, and another is the discussion of whether games can be said to be stories. The former is normative and partly speculative, partly critical, and the latter is descriptive and theoretical. These two debates have a partial overlap of participants, usually and unfortunately identified by the terms ludologists and narratol-ogists2 but one debate concerns the viability of a semi-utopian hybrid game-story genre, and the other is concerned with the seemingly conflicting definitions of games and (in particular) narratives. The ludologist position was not, as has been claimed, to see the focus shift onto the mechanics of game play (Jenkins 2001) but to emphasize the crucial importance of combining the mechanical and the semiotic aspects and to caution against and criticize the uncritical and unqualified application of terms such as narrative and story to games. In other words, the ludologists critique was a reaction to sloppy scholarship (in which key terms are not defined), one-sided focus and poor theorizing, and not a ban against the application of narrative theory to games as such (an act they all had committed themselves): I wish to challenge the recurrent practice of applying the theories of literary criticism to a new empirical field, seemingly without any critical reassessment of the terms and concepts involved. (Aarseth 1997: 14)
That this challenge has been mistaken for a ban on the use of narrative theory in game studies is nothing less than amazing, and perhaps goes to show that humanist academics are often less astute readers, scholars and interpreters than their training gives them occasion to presume. It could also be suspected that anyone
2 A terminology Henry Jenkins (2001; 2004) introduces to label
the two sides in the debate. This was unfortunate, because it ob-scured the fact that all the so-called ludologists were trained in narratology and used narratology in their studies of games. In his influential article, Jenkins made it sound like the ludolo-gists (Aarseth 1997, Frasca 1998, Juul 1999, Eskelinen 2001) were opposed to the application of narratology to games One gets rid of narrative as a framework for thinking about games only at one's own risk and not merely critical to the weak sto-ry-game hybrids at the time, and weak narratological applica-tions. Jenkins article, published on his web site in 2001, ap-pears to be the first time the word narratologist was used as a label for one side in what he termed a potential blood feud.
who echoes Jenkins misleading nomenclature of ludologists vs narratologists simply has not read the literature itself.
Any attempt to clarify the relations between games and narratives will probably end up addressing both issues, but here the second, theoretical issue will be given priority. The question of whether games can succeed as a narrative medium does hinge on the ques-tion of whether games can be considered narrative at all, but answering it is not a job for theory, but rather for (future) criticism and, above all, creative innovation. As I have noted earlier (Aarseth 1997: 5), the difference between games and narratives is not clear-cut. However, games and stories seem to share a number of elements, namely a world, its agents, objects and events. It is crucial to note that these elements are also the cognitive building blocks of human reality, as well as of medi-ated representations of the same. It is thus fruitful to give priority to neither games nor stories, but rather to base the model in the primary reality that spawned both, and that they both are part of, in somewhat different ways.
As pointed out above, it must be noted that games are not simp-ly games, but complex software programs that can emulate any medium, including film, text/novel, graphic novel, and, for that matter, simulate board games and sports. We often commit the mistake of using the metonymic term games for software that in reality are integrated crossmedia packages, such as Max Payne (2001) which contains graphic novel pages and movie-like cutscenes (short animated movie clips that interrupt the game-play), as well as ludic components. Is Max Payne a story or a game? Is it a hybrid? An amalgam? Whatever the answer, it seems clear that it is not purely a game, but a piece of software that does contain, among other things, a game.
3. The common denominators: What do games and stories have in common? There is not one, but many different techniques which have been applied more or less successfully to make games tell stories, and a ludo-narratological model of this design space must account for the ways in which narrative games differ from one another. There can be no single mode of narrativity in entertainment soft-ware, given the diversity of design solutions. To mention a few examples, MYST, Knights of The Old Republic (KOTOR) and Half-Life occupy very different positions in this design space, and a ludo-narratological model must reflect this diversity. My present approach is to see the ludo-narrative design-space as four independent, ontic dimensions: WORLD, OBJECTS, AGENTS, and EVENTS. Every game (and every story) contains these four elements, but they configure them differently. Game worlds can typically be linear, multicursal, or open, and this has great effect on the games perceived narrative structure. Objects (including avatars and player vehicles) can be dynamic, user-created, or static, and again we see a span between the ludic (dynamic, simulated) and the narrative (static). Agents can be presented as rich, deep and round characters (the narrative pole), or shallow, hollow bots (the ludic pole). The sequence of events can be open, selectable, or plotted, and the narratological notion of nuclei (kernels; events that define that particular story) and satellites (supplementary events that fill out the discourse) can be used to describe four different game types: 1. The linear game (Half-Life): fixed kernels, flexible satellites. 2. The hypertext-like game (Myst, Dragons Lair): Choice be-tween kernels, fixed satellites.
3. The creamy middle quest game (KOTOR, Oblivion): Choice between kernels, flexible satellites.
4. The non-narrative game (Chess, The Sims): No Kernels, flexi-ble discourse: just a game.
Five relevant ludo-narrative examples will be analyzed to illus-trate the variable model: Oblivion, Faade, Fahrenheit (aka Indi-go Prophesy), Half-Life 2 and Knights of the Old Republic. These analyses show that story-games display very different features along the four dimensions outlined above; in other words, there are many ways in which a game can be combined with a story, and so it does not make sense to look for one singular type of ludic story. Together with the metonymical problem, this diversity explains much of the confusion in the earlier debate and the pre-vious lack of success in reaching a good, theoretical understand-ing of the narrative aspects of games. This model and analysis will demonstrate that narratology, properly applied, and combined with a broad sampling of different game types, can provide a fruitful and enlightening perspective on games and game design.
4. What is a Narrative? In the debate (insofar as it has taken place) it is sometimes argued that the standard notion of narrative is outdated and in need of expansion because it is poorly suited to describe games. Narratol-ogist Marie-Laure Ryan suggests that narratology must expand beyond its original territory (2006: 98). Commentators such as Jenkins (2001) have suggested that narratives in games can be spatial, embedded and emergent. This conflation of any kind of diegetic or experienced situation with storytelling is what I have previously labeled narrativism (Aarseth 2004, see also Aarseth 1997:94 and Juul 2001). If an(y) interesting experience in a game is an emergent narrative, where does it end? And why limit this category to game-based situations? At some point it becomes hard to distinguish narratives from...