In Search of the Lost Text: Literary Theory and Computer Games
13 March 2001
This is an English summary of Julian's article Auf der Suche nach dem verlorenen Text: Literaturwissenschaften und Computerspiele, recently published in the online journal www.phin.de (issue 15/01)
Abstract The aim of this paper is the exploration of to what extent the terminology of literary theory is suitable for the analysis of computer games. It also attempts to define the areas of inquiry in which the development of new terminology seems necessary. The critical assessments of the work in this field to date reveal that the majority of philological approaches to this genre fail, precisely because computer games are regarded from a literary perspective. However, approaches acknowledging the influence of the theoretical framework on the object of inquiry are capable of providing us with valuable insights into questions of genre, reader response and narratology. Therefore, this paper relies heavily on models provided by Espen Aarseth, Janet Murray and Marie-Laure Ryan. Though addressing only the most basic questions, it becomes clear that a comprehensive theory of computer games is far from completion. Thus, questions as to what role identification, control and simulation play in the area of computer games must remain unanswered for the time being.
In this paper I discuss the possibilities and limitations of analyzing computer games from a literary perspective. Starting from a critique of the theoretical 'imperialism' of most ventures so far to use philological terminology in the study of computer games - e.g. Anderson 1990, Dalum and Sørensen 1996, and Myers 1991 - I attempt to assess the merits of this perspective and its contributions to a general theory of interactive fiction. The importance of this assessment is founded in the tendency of (postmodern) literary theory to disregard the distinction between the text, or code, of a given game with the individual 'readings' that it allows. Thereby the multilinear structure of computer games is often mistaken for a circular one. Another important aspect of these 'misreadings' is the widespread ignorance of literary scholars toward strategies of 'subversive readings' such as cheats, or manipulation of the source code. For the same reason, a semiotic analysis of computer games such as Andersen's is doomed to fail, exactly because it does not regard the actual text but merely a representation thereof. Dalum and Sørensen's 'close reading' of a computer game, on the other hand, tends to fall back behind the achievements of literary theory to overcome the notion of the author's absolute power over the text.
Upon this critical assessment of the achievements of theoretical approaches to computer games, I base my hypothesis that the shortcomings of these attempts are founded in a misconception of the 'reading direction' in interactive fiction. I try to resolve the resulting dilemma by way of the following suggestion: while in a traditional narrative the fictional world is constructed from the raw data of the text, in an interactive narrative the fictional world is deconstructed to reveal the underlying system of rules formulated in the code. Since this process is an arbitrary logical technique initially guided by trial and error, yet a technique that is acquirable and trainable, this may be characterized, in the words of Charles S. Peirce, as an 'abductive' calculus. In my opinion, this misconception of the 'reading direction gives way not only to mistaking individal readings for the text, but also to an undifferentiated equation of text with narrative. One of the few theorists to address this misinterpretation is Espen Aarseth. From his study Cybertext (Aarseth 1997) I draw the concept of the ergodic text that is consequently employed to develop an analytic terminology that reflects the consequences of applying the theories of literary criticism to the field of computer games.
The following section of my paper attempts to assess what could possibly be gained from a theory of computer games that regards the subject not from a pedagogic, or commercial, viewpoint, but from a broad cultural perspective. In my opinion, one of the most urgent tasks to be undertaken in this field is the development of independent aesthetic criteria, i.e. independent of the criteria established in commercial reviews, and independent of the criteria employed in the criticism of other media. Ted Friedman's plea for the development of a 'software theory' was formulated as early as 1993. However, the humanities have not yet responded to his call in a convincing way. And although Friedman must be credited with being one of the first scholars to bring the subject to academic attention, the scope of his work is narrowed by his approach to formulate a poetics, rather than an aesthetics, of computer games.
The definition of my work's area of inquiry is based on Friedman's concept of an independent software theory, and Aarseth's term cybertext. This rather broad definition aims at a delimitation of a field whose borders have been narrowed by non-empirical attempts to formulate what a computer game is. To further structure this field, I turn my attention towards computer game genres in the next section. Since genres are established by evolution rather than by definition, this part of my study attempts to define the genres employed in the gaming community from within, instead of replacing them with categories just as arbitrary as the 'indigenous' ones. In the following discussion of the different genres - action, adventure, simulation, strategy, and role playing games - interactivity, openness, and narrativity are the guiding criteria. In this context, interactivity refers to the frequency of player interaction, whereas the level of openness determines the range of the player's actions. Narrativity spans the entire spectrum of narrative structures in computer games: from the micro-narratives of many action games to the highly complex plots of role playing and strategy games. Thus, the different genres may be placed in a triangular matrix which reveals that the three criteria are complementary, rather than mutually exclusive.
While the criteria of interactivity and openness are rather easily dealt with, narrativity is more elusive. First of all, the claim to assign a value of narrativity to any given computer game is threatened by games based purely on skill or dexterity; such as Tetris. This dilemma can be resolved, though, by assuming that such games can be regarded solely within the other two criteria, for the time being. Still, this problem helps us keep in mind how limited the perspective of literary theory is in this field, and therefore it cannot be disposed of so easily. Nevertheless, most computer games do have some level of narrativity, so the next step is to determine to what extent the terminology of literary criticism can be employed in order to describe the narrative structures of computer games. A short discussion of George Landow's subsumption of computer games under the category of hypertext (Landow 1997), and Kirksæther's equation of games with interactive films (Kirksæther 1998) reveals that both lack the ability to underline the distinctive quality of computer game narration.
Aarseth's assumption that "[i]n the determinate cybertext [...] the functions of plot (szujet) and story (fabula) appear to have traded places, somehow" is a promising perspective, but in the subsequent discussion of this concept he has to conclude that "[t]he concept of plot is unsettled by the reader (user), who, being strategically within it, is in no position to see through it and glimpse a story behind." (Aarseth 1997). His suggestion to replace story and plot by the terms event plane and progression plane is equally unconvincing, since the model he derives from this terminological hattrick is merely a variation of Chatman's traditional model of narratology yielding no genuinely new insights. However, Aarseth's concept of the ergodic intrigue that he derives from an analysis of Marie-Laure Ryan's and Janet Murray's respective works, makes way for a reassessment of interactivity, openness, and narrativity in a different context: Agency, rapture, and immersion, the criteria Murray derives from her analysis of VR systems. Combining these two triads opens a new perspective on the problem of narrativity: When agency is regarded as the result of openness and interactivity, the combination of narrativity and openness results in rapture, and immersion is created out of the interplay between narrativity and interactivity. Therefore, narrativity has a component of rapture as well as immersion. Thus, rapture and interactivity effectively counteract the critical distance of the 'implicit player' towards the narrative he is confronted with.
The problem of narratological perspective is also discussed within the context of Murray's terms. The questions 'who speaks?' v. 'who sees?' that guide the analysis of the narrator's point of view in traditional narrative cannot be dismissed in regarding interactive fiction. Whatever degree of immersion and rapture is achieved by such a narrative is clearly a function of the perspective from which it is told: the bird's eye view of a strategy game does not allow for the same level of affective participation as a 'first-person-shooter', even though an 'auctorial' perspective might grant a higher degree of agency, or efficiency. However, Stanzel's traditional model of point of view proves to be too rigid for the study of computer games. Therefore, the more flexible concept of focalization (Rimmon-Kenan 1983) is employed to safeguard a differentiated analysis.
Genette's first question - 'who sees?' - provokes the naïve response: 'The player.' But this simple answer immediately reveals multiple implications when the next question is asked: 'Whose eyes does the player perceive the gameworld through?' In a first-person-shooter there usually is an avatar representing the player in the game. But when we turn to role playing and strategy games more and more representations of the player come into view, thus complicating the narratological structure to the point of utter confusion. It is important to keep in mind, however, that in an interactive narrative the roles of the narrator and the narratee cannot be distinguished so easily. Moreover, when dealing with computer games the question of perspective involves not only the avatar's point(s) of view, but also the point of view from which he is seen. This seems to necessitate the addition of an 'implicit author' to the model, despite the difficulties of locating this function in respect to the reader's 'co-authorship'. This dilemma can be resolved by Brenda Laurel's suggestion to subsume computer-mediated communication under the category of dramatic interaction (Laurel 1997). In this model the roles of actor and spectator, rather than the roles of author and reader, are contained in the person of the player.
Unfortunately, Laurel's concept proves incompatible with Rimmon-Kenan's model of focalization when dealing with the question 'who speaks?'. Since Laurel's model does not provide a narrator, we are left in a nonsensical situation in which the player remains the only narrative force. This aporia is intensified when dealing with games in which the identification of the player with the dramatis personae is hardly conceivable - 'software toys', such as SimCity for example, in which the player is the director rather than an actor. Even though this aporia cannot be resolved presently, it highlights the importance of focussing our attention on the question of identification. After all, the location of the narrative function in interactive fiction can be determined only in reference to the position of the player in the communication process. The assessment of this referential model's implications is facilitated by the hypothesis that the narratological perspective of the first-person-shooter corresponds to the second person point of view of 'text adventure games' such as Zork.
Despite its paradoxical appeal, this concept proves to be the key to understanding the communication process in interactive fiction. After all, it is the narratee rather than the player that is addressed in the text adventure game. The equation of the narratee with the audience in dramatic communication makes perfect sense, though. Thus, Laurel's model reveals its compatibility with a modified narratological model. The merits of this conceptualization are obvious: The differentiation between contrary reading directions in traditional v. interactive narratives can then be discarded as trivial.
However, the combination of the dramatic and the narratological model of communication is insufficient to locate the narrative function in this process. Espen Aarseth provides us with a schema of the computer game's components: In this schema a simulation engine and a representation engine mediate the communication between the code and the interface. These 'engines' correspond to the parts of the code that can be manipulated by the player via the interface, or directly. Consequently, the player's identification with the narratee is a necessary precondition for successful communication, regardless of the manifestation of this narratee on the interface level. This narratee must then be conceptualized as both a listener and a responder that is in permanent interaction with the code. This enables us to define the implicit author as a function of the code itself that governs one direction of the information flow in the communication process, while the player controls the other. Therefore, the model incorporates two narrative 'voices' in permanent dialogue, representing a character focalizer from within and a narrator focalizer from without, respectively. From this viewpoint, Geoffrey Rockwell's suggestion to "extend Bakhtin's theory of the novel to provide us with a theoretical framework for understanding computer games", makes perfect sense.
Since these voices are much more likely to contradict each other than to agree, the last section of my paper deals with the question of narrative control as a potential source of conflict. Or, in Rockwell's words: "How are you defined by the choices the game affords?" (Rockwell 1999). The implicit equation of interaction with identification in this question seems to suggest that identification is gained solely through interaction. However, a closer inspection of the process of identification reveals that interactivity both facilitates and counteracts identification. After all, the limited range of actions the game offers to the player, eventually leads to the player's recognition of his representation's limitations in the game. Therefore, the activity of playing a computer game must be characterized as a process of demystification: "Learning and winning [...] a computer game is a process of demystification: one succeeds by discovering how the software is put together" (Friedman 1993). For the same reason, playing a computer game is a genuinely hermeneutic activity: The player interprets the signs on the interface as manifestations of the rules formulated in the game's code. Yet the player will only gain control over the game through a process of demystifiction. From this perspective, the initial hypothesis of 'gaming' as a deconstructive process gains renewed relevancy: only by overcoming the willing suspension of disbelief can the player 'master' the game.
Summarizing the results of this study, I come to the conclusion that the theoretical import of literary criticism to the field of computer games is not without its difficulties, yet a literary approach to interactive fiction that reflects the limitations of its critical terminology can provide valuable insights into these games' narrative, and semiotic, structure. In my opinion, the literary approach to this field can be extended further to include questions of mimesis and simulation, self-referential aspects, as well as the narrative role of computer game characters.
Aarseth, Espen (1997): Cybertext - Perspectives on Ergodic Literature, The Johns Hopkins University Press, Baltimore and London
Andersen, Peter Bøgh (1990): A Theory of Computer Semiotics: Semiotic Approaches to Construction and Assessment of Computer Systems, Cambridge University Press, Cambridge
Dalum, Astrid und Sørensen, Finn V. (1996): Interactive Fiction [http://www.centrum.dk/users/finnv/intfict.htm]
Friedman, Ted (1993/1995/1999): Making Sense of Software: "Computer Games and Interactive Textuality" [http://www.duke.edu/~tlove/simcity.htm],
Kirksæther, Jørgen (1998): The Structure of Video Game Narration [http://cmc.uib.no/dac98/papers/kirksaether.html]
Laurel, Brenda (1991): Computers as Theatre, Addison-Weasley, Menlo Park, CA
Murray, Janet (1997): Hamlet on the Holodeck: The Future of Narrative in Cyberspace, The Free Press, New York
Myers, David (1991): "Computer Game Semiotics", in: Play and Culture, No. 4, p.334-345
Rimmon-Kenan, Shlomith (1983): Narrative Fiction. Contemporary Poetics, Methuen, London
Rockwell, Geoffrey (1999): "Gore Galore: Literary Theory and Computer Games"[http://www.interchange.ubc.ca/winder/papers1999.htm]
Ryan, Marie-Laure (1997): "Interactive Drama: Narrativity in a Highly Interactive Environment", in: Modern Fiction Studies 43/3, 677-707