Towards a theoretical framework for interactive digital narrative

Towards a Theoretical Framework for Interactive Digital Narrative Hartmut Koenitz, Georgia Institute of Technology, School of Literature, Communication and Culture, 686 Cherry St., Atlanta, Georgia 30332 [email protected] Abstract. The emerging artistic practice of interactive narrative in digital media marks a profound departure from traditional narrative. The application of traditional narrative theory for interactive narrative is problematic, since the affordances of digital media challenge many underlying assumptions of theories related to non-digital media. This paper proposes a theoretical framework for interactive storytelling, which addresses these concerns by foregrounding system (the digital artifact) and process (the user interacting with the system) over the product-centered view of legacy media. On this basis, protostory, narrative design, and narrative vectors are proposed as new terms to more adequately describe the structure of narrative in interactive digital storytelling. This move is also relevant for practical design given the influence theoretical concepts have on concrete implementations. Keywords: Interactive Storytelling Theory, Interactive Narrative, Digital Media, Story, Plot, Legacy Media, Instantiation, Protostory, Narrative Design, Narrative Vectors. 1 Introduction Interactive digital narrative (IDN) in its many incarnations as interactive drama (e.g. Façade [1]), hyperfiction literature (e.g. Afternoon [2]), interactive fiction (IF) (e.g. Zork [3]) and other variants such as interactive cinema (e.g. A City in Transition: New Orleans 1983-86 [4]) and narrative games (e.g. The Last Express [5]), heralds not only a change in the technology of representation, and in the opportunities for artistic expression, but also a challenge to existing concepts in narrative theory, such as the role of the author and the concept of a single unified plot. So far, these challenges have been approached by modifications to established theories. A first milestone was set by Brenda Laurel’s re-working of Aristotle’s Poetics [6] based on an understanding of digital interactive narrative as similar to the stage play [7, 8]. Laurel’s theoretical approach was used as the basis for practical experiments by Carnegie Mellon’s OZ group under Joseph Bates [9], which eventually led to the first fully realized interactive drama, Mateas’ and Stern’s Façade [1] Similarly, a poststructuralist perspective articulated by Jacques Derrida, Michel Foucault, Roland Barthes, Umberto Eco, and Jean Baudrillard led to the development of the Storyspace platform and to the creation of hyperfiction works like Michael Joyce's Afternoon [2] and Shelley Jackson’s Patchwork Girl [10]. A third approach has drawn on nonliterary and non-western concepts of narrative – for example African aboriginal or diasporic oral narrative traditions – as a theoretical basis of IDN. Pamela Jennings’ work The book of ruins and desire [11] and Fox Harrell’s GRIOT system [12] implement this approach. Finally, an approach based on narratology as devised by Barthes and Claude Bremond and further developed by Prince, Genette, Chapman and others is proposed by Nick Montfort [13] for IF and by Marie-Laure Ryan [14, 15] as a general model for IDN. To start with any established theory of narrative has clear advantages. Terms, categories, and methods of analysis already well understood can be used to analyze and describe phenomena in interactive digital narrative. On the other hand, analyzing IDN within the frameworks of theories created to describe narrative in traditional media carries the danger of misunderstanding or underestimating the nature of the change. For example, once we understand IDN to be similar to the ancient Greek stage play we can become entrapped in this analogy and overly wedded to the framework of Aristotle’s Poetics. Consequently, aspects that do not fit that particular frame of reference (for example digital media’s capacity for an encyclopedic treatment of a given topic vs. Aristotle’s notion of a complete action that only includes necessary elements) might be misunderstood as minor or even excluded altogether, thus limiting our ability to fully capture the potential of IDN. To overcome these limitations I propose a more adequate framework as a step towards a fully developed theory of IDN. 2 An Initial Approach Towards IDN The analysis in this paper is guided by a framework provided by earlier and contemporary work in the understanding of computers as digital media [8], the affordances and phenomenological qualities of digital media [16] and aspects of the experience and the design of IDN [16, 17] and narrative [18]. This approach takes narrative as a cognitive structure that can be evoked in different ways. It also understands digital media as separate and distinct from legacy media such as the printed page, film, or electronic media. Additionally, digital media is understood to have specific affordances, which consequently make IDN a form of expression that tightly integrates interactivity and narrative. Janet Murray’s descriptive framework starts with her understanding of a computer’s ability to “execute a set of rules” [16] and to be an engine that runs instructions as the procedural affordance. The participatory affordance captures the computer’s ability to react to user input, and respond in a predictable manner. The spatial affordance denotes the ability of computers to represent space and allow a user to traverse this representation on the computer. The encyclopedic affordance is Murray’s term for the computer’s ability to handle and present huge amounts of data. Murray then defines the phenomenological categories of agency, immersion, and transformation to constitute the aesthetics of digital media. She sees agency as the experience a user gains by “making something happen in a dynamically responsive world” [17] if the digital artifact reacts in a coherent and predictable manner. Immersion is the ability of a digital artifact to hold our interest, and minimize distraction by offering an “expansive, detailed, and complete” [17] experience. Unlike Ryan and some game theorists, Murray sees no conflict between “interactivity” and “immersion” or interactivity and narrative. Computer-based narrative is created by exploiting the affordances of the digital medium, and is reinforced by participation, so that the interactor experiences agency that is based on arousing and rewarding narrative expectations, and the active creation of belief in the story world. From this perspective the compound term “interactive narrative” is perhaps misleading, since it can be misunderstood in a way that takes interactivity as an “added feature” for narrative. On the contrary, the perspective taken here understands interactivity and narrativity as inseparable, mutually reinforcing aspects of the emerging expressive form of IDN. David Herman [18] augments narrative theory with additional aspects drawn from cognitive science. Overall Herman describes narrative as a cognitive structure that can result from different coding strategies and forms, a position echoed by Marie-Laure Ryan [15]. In this vein, Herman defines narrative as a “forgiving, flexible cognitive frame for constructing, communicating, and reconstructing mentally projected worlds.” [18] This definition de-couples narrative from specific forms or media and opens up the space for experiments IDN. It also removes the requirement for specific roles of narrator and narratee and is therefore compatible with Murray’s framework of affordances. Consequently, Herman’s definition serves to define narrative in my approach towards a theoretical framework for IDN. 3 Towards a Theory of IDN The starting point for a specific theory of IDN is a change of perspective. Instead of understanding IDN to be similar to narrative in legacy media, interactive digital narrative is taken as dissimilar. Both the material basis in digital media and the conceptual backdrop of IDN as a participatory transformational experience merit this change. This stance does not represent a departure from earlier approaches but rather a radical continuation based on more than two decades of theoretical and practical research. In this fashion, Nick Montfort’s distinction between an IF work and an ordinary narrative is especially productive: “A work of IF is not itself a narrative, it is an interactive computer program” [19]. However, he still considers narratology a useful framework for the analysis of IF works: An IF work is always related to story and narrative, since these terms are used together in narratology, even if a particular work does not have a ‘story’ in this ordinary sense. [13] What is embedded in his observation is a distinction between the material artifact as a computer program and its output as a particular instantiation. This distinction is true for IF and other kinds of IDN. Another important aspect of IDN is in the relation between theses two categories. IDN assumes interaction and thus a participatory process in which a participant engages with the computer program to produce the output. The product of an IDN work – a recording of a single “walkthrough” - might be understood as a narrative in a more traditional sense and could be analyzed with the tools and methods of classical narratology. However, the same theoretical framework does not account for the digital interactive system that enables the production of the narrative in the first place. A crucial step towards an adequate theory of Interactive Digital Narrative is to understand IDN works as comprised of system, process, and product (see Fig. 1). Fig. 1. High-level view of IDN. This model of IDN is inspired by Roy Ascott’s theory of cybernetic art [20]. Ascott advises artists to look at the scientific discipline of cybernetics, the study of “control and communication in the animal and the machine” [21], and create art inspired by cybernetics’ concern with the behavior and regulation of environments, and with organizational structures. Espen Aarseth must be credited with the introduction of cybernetics to IDN. He derives his term Cybertext explicitly from cybernetics and describes the “cybertextual process” [22] in which a user affects the narrative in a cybernetic feedback loop. Ascott’s definition, however, provides a better basis for a theory of IDN, as he improves upon Wiener’s mechanistic concept by merging it with artistic sensibility. His “cybernetic art matrix” [23] proposes a tight integration between art and the computer and foreshadows the importance of interaction for digital media. Furthermore, Ascott understands cybernetic art to represent a change in the artistic focus from product to process and from structure to systems, which will turn the “observer” into a “participant” [24]. Ascott’s vocabulary therefore can be used productively for the definition of a framework for IDN. I propose system as a term to describe the digital artifact, as it exists on a digital storage medium combined with the hardware on which the artifact is executed. This includes the executable programming code and assets - digital representations of pictures, movie clips, sounds, and text, as well as network links to more assets on a local network or the Internet. Additionally, it also includes the connected hardware – keyboards, mice, displays, and other hardware (eg. sensors) used in a digital installation. The system contains “potential narratives”, a term Montfort derives from the Oulipo group’s notion of “potential literature” [see 19]. Once a user starts to engage with the system, a process is created. The actions of the user as interactor, and the opportunities provided by the system define and shape the process. The resulting product of IDN represents an instantiated narrative.1 Instantiation here describes the quality of IDN to produce very different results or narrative products from the same source (the system) through a participatory process. Each single instantiated “walkthrough” could be recorded and may be analyzed in terms of traditional narratology, as a linear narrative. While any single product is an integral element of any IDN, it is important to realize that it represents only one particular instantiation that can and will change as soon as the process changes. In terms of theoretical analysis the product alone is therefore severely limited as a representation of an IDN work. A full analysis of any IDN needs to include an examination of process and system. From this perspective, theoretical approaches based on theories for legacy narratives are problematic since they foreground the analysis of the product of IDN. A potential criticism of this view is the argument that IDN’s process represents the equivalent of the cognitive process of understanding literature and other narratives as described by the reader-response theory [26] and contemporary cognitive narratology [18]. The model proposed here does indeed take the creation of meaning of a narrative in the mind of a recipient as an active process. However, potential narratives in IDN provide an additional mental plane for the participant. Not only does the participant create a mental model of the emergent story, she also speculates about the consequences of her own actions for the narrative, assesses her level of control, and as a result formulates and executes strategies of interaction. This additional plane of consideration and control is an important factor that distinguishes IDN from legacy non-interactive forms such as the novel, or the movie. While this plane does also exist in participatory theater, improvisational performances, story games, and “choose your own adventure” books, these non-digital interactive forms differ from IDN in their material basis in legacy media, and consequently do not share the same affordances as digital media. As a result, IDN can now be defined more clearly as an expressive narrative form in digital media realized in a system containing potential narratives and experienced through a process that results in products that represent instantiated narratives. 1 Noah Wardrip-Fruin [25] shares the concern for process, which he distinguishes from “output.“ He describes the aesthetics of “expressive processes” and foregrounds the evaluation of a work based on these aesthetics. 3.1 Protostory, Narrative Design, and Narrative Vectors Given the flexible and malleable quality of IDN afforded by procedurality and participation, neither story nor plot/discourse can adequately describe an IDN work, as the fixed story (or “content plane of narrative” in Prince’s terms) of legacy media gives way to a space containing potential narratives. At the same time, plot/discourse as the fixed material manifestation gives way to a flexible presentation of narratives while they are being realized. Additionally, a neat distinction between the two categories is no longer possible, since the IDN system contains and encodes aspects of story and discourse by supplying both content and structures of the concrete expression. These aspects need to be reflected in terminology that intends to adequately describe IDN. I propose the term protostory to describe the concrete content of an IDN system as a space of potential narratives. Any realized narrative experience is related to the respective protostory through a process of instantiation. The term protostory shares the aspect of a malleable formation with the concept of prototype-based programming (sometimes also called instance-based programming). In this variant, not only the content (as with classes), but also the behavior and structures (called prototypes) itself can be changed at runtime [27] This model more adequately describes the flexible relationship between an IDN system and a particular realized narrative and clearly distinguishes it from any kind of mechanical reproduction that produces the same copy every time. The protostory then is a prototype, or a procedural blueprint, that describes the space of potential narrative experiences contained in one IDN system. However, protostory is more than just a computer program, as the term encompasses not only the concrete programming code and interactive interfaces, but also the artistic intent that enables a participatory process of instantiation that results in the realization of potential narratives. The concept of plot as separate from protostory is problematic given the compound nature of potential narratives, which contain both structure and content. Instead, I propose narrative design 2 to describe the structure within a protostory that contains and enables a flexible presentation of a narrative. This includes the segmentation and the sequencing of elements and the connections between them. Additionally, the procedural logic applied in the presentation of elements is part of the narrative design (see Fig 2). 2 In contrast, Mateas uses the same term to describe narrative segmentation [see 1]. Fig. 2. Protostory and Narrative Design in an IDN System. The term narrative vectors describe sub-structures in a narrative design that provide a specific direction for the narrative. Narrative vectors work not as isolated structures, but rather in connection to the preceding and the following parts of the narrative. The purpose of such structures is to convey important aspects to the interactor, to prevent an interactor from getting lost and to help to retain a level of authorial control. For example in an IDN murder mystery, a narrative vector could be the occurrence of a murder or the disappearance of an important victim, and also a breakdown of the interactor’s car that prevents her from leaving the crime scene before all clues are gathered. Narrative vectors are roughly functional equivalents to the term plot points in legacy media [see 28]. The term plot point has been used to describe positions within a story that are created by the author in order to propel the narrative experience forward. As a next step I will test this new terminology by applying it to two disparate examples. 3.2 Examples: Afternoon and Façade In Michael Joyce’s Afternoon, the protostory is the space of all lexias and hyperlinks together with the possible paths an interactor can take and the author’s artistic intend to let the interactor experience a fragmented narrative of a psychotic state. An interactor instantiates a particular realized narrative by reading lexias and following hyperlinks. The narrative design in Afternoon describes the segmentation of lexias as well as the hyperlinks connecting them and the guard fields that generate conditional links. Narrative vectors in Afternoon are combinations of lexias and links that are designed to create specific experiences, for example the re-visiting of a particular lexia after the interactor has gathered additional knowledge (see Fig. 3). Fig. 3. Protostory, narrative design, and narrative vectors in Afternoon. Mateas’ and Stern’s work Façade [1] applies sophisticated artificial intelligence to create a richly varied range of narrative possibilities. The protostory in Façade is the space of possible stories described by the contents of the beats (narrative units), the drama manager’s restrictive rules and goals and the artist’s intent to let the interactor experience a marriage falling apart and attempt to save it. By communicating with Grace and Trip, the two other characters in Façade, and by moving within the space of their apartment, an interactor instantiates a realized narrative, which could for example lead to the couple breaking up or throwing the interactor out. The narrative design in Façade is comprised of the different beats, the concept of a story arc and pre-authored goals. Narrative vectors are formed by the drama manager component as a result of the interactor’s input and by consulting pre-authored goals as well as distinct phases in the story arc. Narrative vectors (see Fig. 4) in Façade determine if an interactor is kicked out or if she reaches the therapy part in which Grace and Trip are able to rescue their marriage. Fig. 4. Protostory, narrative design, and narrative vectors in Façade. Understanding the two works in this way also facilitates the exploration of questions about the content of an IDN work outside of the narrative design, which so far has been mostly overlooked. For Afternoon, the aesthetics and participatory possibilities provided by the Storyspace authoring system and its playback component can be analyzed as environment definitions and settings. Similarly, for Façade, the virtual space of the couple’s apartment and the possibilities afforded by the physics engine become an integral part of the examination of the protostory as aspects of the environment and allow a more complete understanding of the work. For both examples, the narrative design is seen as a complete structure comprised of narrative vectors, which enables a classification independently of legacy story structures – Afternoon no longer has to be understood as rhizomic and Façade can be classified independently of legacy dramatic structures. Additionally, narrative vectors comprised of lexias and links or the combination of the drama manager and specific beats allow us to examine the particular narrative strategies of Afternoon and Façade. 4 Conclusion The addition of protostory, narrative design, and narrative vectors to the theoretical vocabulary of IDN, together with the understanding of IDN as comprised of system, process, and product creates a rich descriptive framework for IDN that forms the beginning of a more fully developed theory. The brief analysis presented here of realworld artifacts in the form of two examples (Afternoon and Façade) is an early testimony to the applicability of this framework to diverse works within the IDN spectrum. The inclusion of environment, assets, and settings as integral parts of the examination of digital artifacts enables a more complete understanding of IDN works, while narrative design and narrative vectors allow an understanding of narrative structures beyond legacy notions of story structure and dramatic arc. Further work in this area should analyze the primitives and the segmentation of protostories and create a taxonomy of narrative designs to identify forms and genres. Process should receive additional focus, to arrive at an analysis that enhances the understanding of the interactor’s mental processes while experiencing IDN works. This analysis should also examine the relationship between computational and mental processes in more detail. In practical terms, the clear departure from legacy narrative opens up a space for bold experiments in IDN that do not need traditional narratives as a yardstick to measure against. The theoretical framework proposed here changes the focus of evaluation: not in computational complexity, not in “discourse” or language/images in which a story is told; but in complexity and coherence of protostory, the aesthetics of narrative design, and the richness of narrative vectors as creating expectation and occasions for dramatic agency. References 1. Mateas, M., Stern, A.: Façade: An Experiment in Building a Fully-Realized Interactive Drama, In: Game Developer's Conference: Game Design Track (2003) 2. Joyce, M.: Afternoon, a Story. Eastgate Systems, Watertown (1991) 3. Blank, M., Lebling, D.: Zork I. Infocom, Cambridge (1980) 4 Davenport, G.: Interactive Multimedia on a Single Screen Display. In: Videotechnology Technical Session, Current Applications of Videotechnology. Computer Graphics Systems, National Computer Graphics Association Conference, March 22, 1988 (1988) 5. Mechner, J.: The Last Express. Brøderbund, Eugene (1997) 6. Butcher, S. H. (ed.): The Poetics of Aristotle, With Critical Notes and a Translation. Macmillan, London (1922) 7. Laurel, B.: Toward the Design of a Computer-Based Interactive Fantasy System. Doctoral Thesis. Drama Department, Ohio State University (1986) 8. Laurel, B.: Computers as Theatre. Addison-Wesley, Reading (1991) 9. OZ Group website: http://www.cs.cmu.edu/afs/cs/project/oz/web/ 10. Jackson, S.: Patchwork Girl. Eastgate Systems, Watertown (1995) 11. Jennings, P.: The Book of Ruins and Desire. http://www.pamelajennings.org/PDF/Jenningsruins%20and%20desire.pdf (1996) 12. Harrell, F.: GRIOT's Tales of Haints and Seraphs: A Computational Narrative Generation System. In: Wardrip-Fruin, N. P., Harrigan, P. (eds.) Second Person: Role-Playing and Story in Games and Playable Media. MIT Press, Cambridge (2007) 13. Montfort, N.: Twisty Little Passages: An Approach to Interactive Fiction. MIT Press, Cambridge (2003) 14.Ryan, M.-L.: Narrative and the Split Condition of Digital Textuality. http://www.dichtungdigital.de/2005/1/Ryan/ (2005) 15. Ryan, M.-L.: Avatars of Story. University of Minnesota Press, Minneapolis (2006) 16. Murray, J.: Hamlet on the Holodeck. The Free Press, New York (1997) 17. Murray, J.: Inventing the Medium: Principles of Design for Digital Environments. MIT Press, Cambridge (Forthcoming) 18. Herman, D.: Story Logic: Problems and Possibilities of Narrative. University of Nebraska Press, Lincoln, Nebraska (2002) 19. Montfort, N.: Toward a Theory of Interactive Fiction. http://nickm.com/if/toward.html (2003b) 20. Ascott, R.: The Construction of Change. Cambridge: Cambridge Opinion 41, 37-42 (1964) 21. Wiener, N.: Cybernetics, or Control and Communication in the Animal and the Machine, The MIT Press, Cambridge (1948) 22. Aarseth, E.: Cybertext: Perspectives on Ergodic Literature. Johns Hopkins University Press, Baltimore (1997) 23. Ascott, R.: Behaviourist Art and the Cybernetic Vision. In: Cybernetica, Journal of the International Association for Cybernetics 10, 25-56 (1967) 24. Ascott, R.: Is There Love in the Telematic Embrace? In: Art Journal 49:3, 241 (1990) 25. Wardrip-Fruin, N.: Expressive Processing. MIT Press, Cambridge (2009) 26. Iser, W.: Der Akt des Lesens. Theorie ästhetischer Wirkung. Jauß, München (1976) 27. Noble, J., Taivalsaari, A., Moore, I (eds.): Prototype-Based Programming: Concepts, Languages and Applications. Springer-Verlag, Berlin (1999) 28. Field, S.: Screenplay: The Foundations of Screenwriting; A step-by-step guide from concept to finished script. MJF Books, New York (1998) View publication stats