Structuralist narrative theorists have derived a model of several layers of narrative meaning (Figure 1). The structuralist approach was initially formulated in the structural linguistics of Sausseure (summarized in ). Sausseure made the fundamental distinction between a language (la langue) and the speech acts facilitated by the language (la parole). This general principle inspired a structural approach to general cultural analysis in which specific social forms are seen as manifestations of an underlying generative system.
Structural narrative theory involves more than Saussure’s simple two level system. Specific narratives are understood as the instances in time that express stories. Such an instance can in general be referred to as a text, where “Text is a spoken or written discourse which undertakes the telling of the events in a story” . Only the text is available to the reader. The concept of text has been generalize to cover audio-visual media, since many of the ways narrative functions semiotically are the same across different media forms. Narration is then the act or process of production of the text. The text itself is the narrative. This level of narrative structure has also been referred to as the discourse level .
Figure 1: Layers of meaning in narrative texts.
In general, the text of a narrative expresses a story where “‘Story’ designates the narrated events, abstracted from their disposition in the text and reconstructed in their chronological order, together with the participants in these events” . An impetus behind the identification of the story as a separate level of meaning from the narratives that express it is the fact that the same story may be expressed in many different narratives, either within the same medium or across different media. Between the levels of narrative and story, however, is a further question of which aspects of the story are expressed by a narrative. This is the level of the plot, where
“… plot, story-as-discourse, exists at a more general level than any particular objectification, any given movie, novel or whatever … Its function is to emphasise or de-emphasise certain story-events, to interpret some and to leave others to inference, to show or to tell, to comment or to remain silent, to focus on this or that aspect of an event or character” .
Structuralist narrative theorists have also sought general structures underlying the formulation or generation of stories. Vladimir Propp’s  pioneer work Morphology of the Folktale presents an analysis of the structural generative system underlying a genre of Russian folk tales. Within this system, a typical folktale is built around seven types of character (or more specifically, spheres of action corresponding to performers), namely 1) the villain, 2) the donor, 3) the helper, 4) the princess (and her father), 5) the dispatcher, 6) the hero and 7) the false hero. The names of the characters containing these functions differs from tale to tale, but the type of actions they perform are always the same. A function can be “understood as an act of a character, defined from the point of view of its significance for the course of the action”. Propp presents the system as having afixed number of thirty-one possible plot functions. Not all of the functions are necessary in any given story, but where they occur they always have the same sequential order.
This level of generative substructure is analogous to Saussure’s language level. From the generative substrate it is possible to create a great many stories. Each story can be the source for many plots, and each plot can be expressed in many narratives. Viewing this as a hierarchy, it can be seen that, beginning with the narrative level and going down to the structural substrate, each level down has an increasing generative potential in terms of the number of actual narratives that it facilitates and by which it is expressed. Propp’s specific model has been applied within other narrative genres and analogous systems are plausible for forms of narrative that it does not obviously account for.
Mapping Narrative Terms Onto Computer Games
Irrespectively of the general status of structuralist narratology within the contemporary study of narrative systems, the model is very useful when applied to the analysis and design of interactive narrative and story construction systems, and the identification of several levels of narrative meaning clarifies the relationships between different strategies for interactive narrative and story construction. However, models of narrative meaning developed to account for verbal and textual narrative cannot be applied naively to interactive digital media systems and computer games. How narrative concepts are to be interpreted depends a great deal on the specific form of digital media system in question.
If an interactive media system has what might be called a medium level of granularity in the media chunks that are interactively selected, then the narrative levels can be applied in a more straightforward way. Examples of media systems of this kind include classic hypertexts and interactive movies having a link structure (described in detail below) through which the viewer/reader chooses a path. Interactivity in this case can be created either within or between levels in the classic narrative model. Branching within levels involves providing authored branching pathways from which the reader/viewer may select withina representation of the story, plot or narrative levels. In this case:
- a branching narrative in the strict sense provides interactive selection of narrated elements conveying particular plot elements in particular ways
- a branching plot structure provides alternative pathways through the representation of an overall plot related to a common story; the events, characters and settings of the story remain unchanged, but those narrated to the reader/viewer are interactively determined
- a branching story structure involves interactive selection/determination of a representation of specific set of events, characters and settings constituting a story based upon a predefined set of potential events, characters and settings
The simplest way of structuring interactive story systems is to provide a branching structure connecting media components that represent story and plot elements by narrative components (eg. blocks of text, sequences of video). In principle the branching structure itself could be a representation of the generative substrate, the (interactive) story, the (interactive) plot or be part of the surface narrative. In all cases the components associated with nodes in the branching structure are narrative components.
Systems in which the different narrative levels are represented separately are likely to be research systems in which the user experience is frequently synthesised dynamically (and interactively), rather than using a pre-authored branching structure. Interactive synthesis of the narrative in this case may include one or more of:
- creation of a story from a structural substrate; the substrate may be a system of a-temporal relations from which a specific temporal order will be created in response to reader/viewers interactions or specifications
- creation of a plot from a story
- creation of a specific narrative from a plot
The substrate used for story generation may be the kind of story-oriented structure described by Propp, or it may be a thematic representation. For example, the Auteur system (, ) is a story/plot/narrative generator, while the Automatist  and FRAMES  systems are plot/narrative generators based upon a single story. All of these systems are theme-driven (implicitly or explicitly), and create linear presentations by dynamically linking video segments. The video segments represent the story and plot. Generation of the linear presentation may be based upon additional formalised representations of the meaning of the video segments at one or more of the levels of narrative meaning (structural substrate, story or plot) and/or other interpretations or descriptions of the video components.
Contemporary interactive 3-dimensional computer games, including action games, strategy games and role-playing games, represent a more complex case to which narrative models cannot be applied in a straightforward way. This is because these games are fundamentally and qualitatively different from traditional linear narrative forms in the following ways:
- the player is a joint reader/author at some levels of narrative structure
- more semiotic levels have textual manifestations
- those text levels are generatively interdependent
This is best considered by examining each of the levels of the classic narrative model and looking at how this can be applied to 3D computer games (also considered with different emphasis in ). The analogous relationships are represented on Figure 2.
Figure 2: Relationsips between levels of semiotics structure in verbal/textual language, linear narratives and computer games.
The narration or discourse level can be mapped onto what appears on the computer screen, regarded as an audiovisual artefact extended in time. If one were to watch the screen while someone else is playing, the artefact would effectively be a movie. However, a game is clearly not intended to be experienced (primarily) in this way; it is intended to be played and the player has a major role in determining what occurs in the unfolding history of the screen. The player may be presented with some elements of predefined narrative, potentially including non-interactive cinematic cut scenes and in-game stories and histories (revealed, for example, via in-game simulated books or dialog texts). The player then has a role in creating the context for and bridging sequences between these narrative elements by their play behaviour within the scope of possible behaviours provided by the system designers. The player is, then, a partial author of the narrative constituted by the screen history of play.
Disregarding for the moment the non-interactive elements of a game narrative (such as an overall story arc framing play and delivered via cut scenes), the analogy for the plot level of classical narrative can be regarded as that part of the game story as revealed to the player through the screen. This is, however, the primary focus of the player as a performer within the game world. The plot is not something delivered to the player, but something actively created by the player in interaction with the game system and its (often very large) implied space of possible detailed plots. Since the creation of the plot requires very intensive interaction (this being the focus of player engagement and immersion), it may be more accurate to refer to this level of meaning as a performance level, rather than a plot level. A plot is a representation of a story as revealed through an act of narration by a narrator for a viewer/reader/listener, while the game play performance is an act of partially creating a story of which the performance is understood to be a part and is performed for the pleasure of the performer.
The game story is the total implied game world history as determined by the pre-designed potential of the game in interaction with the game play actions of the player. Not all of the story is revealed to the player; this is most clearly seen in the case of simulation-intensive games like SIM City. The player performs in the plot, as revealed by the computer screen; in the case of SIM City the performance amounts to performing game moves such as creating building zones, setting tax rates and providing budgets for municipal services. Most of these actions are performed on the screen in a medium scale isometric view of the city. However, the total city is continuously simulated. The history of the simulation is analogous to a story, while the history of the screen view through which the player performs control of the simulation is analogous to a plot. Hence it is perhaps more accurate to refer to this as a simulation level, rather than a story level. The sense of this concept is to capture what the player does as a performance within and in relation to the game world, as well as how this performance affects what happens “behind the scenes” that the player may not have direct and explicit knowledge of. The result, in retrospect, may be a story, but the degree to which the player apprehends the unfolding simulation as a story is a function of the design of the game mechanics and the predefined narrative elements of the game, together with the player’s mode of assimilating the experience. Some game designs will encourage the apprehension of the play experience as an experience of “being in a story” more strongly than others. This amounts to game design to support and encourage the immersionist and (dramatic) performance modes of game play described above. Viewing the layers of game time semiotics too strongly in terms of traditional narrative concepts of plot and story loses sight of these being design orientations that may be worked for, against or ignored in the design of the game. However, viewing the layers in terms of performance and simulation acknowledges the scope for designing for different play styles, including those that are not about story: achievement, griefing and socialising.
Looking finally at the level of the structural substrate, this must be subdivided into two different levels in the case of interactive 3D games:
1. the general, cultural structural substrate, as addressed by Propp . This can be understood as a paradigm of distinctions by which the experiences delivered by a game are comprehensible within a culture or a subculture. This level must include not only implicit systems for apprehending narratives, but also for apprehending games as games, and simulations as simulations.
2. the specific generative substrate built into a game engine and providing the basis for computationally realising the player’s experience in relation to the player’s performance choices.
2. is the foundation of all interactive 3D computer games and represents a radical difference from traditional narratives. A system of texts represents it (design documents, high level models, software code, machine code and game data), and it has a physical representation as a pattern of electromagnetic fields interpreted symbolically as ones and zeros, and when translated as a physical pattern into a computer it can be interpreted to realise the form referred to by players as “the game”. At the level of design principles for realising specific play styles, this can be referred to as the game framework (explained in more detail in ).
For a game designer, the value of structural models like those developed by Propp is to inspire design concepts for game frameworks supporting immersive and performative game play. Such design principles are poorly developed, while current methods for achieving narrativity in computer games most strongly favour the audience style of narrative engagement. To show why this is the case, and to map out alternatives for immersive and performative game play, it is necessary to look more closely at the time scales of pre-designed game time structures, beginning with the simulation level at the smallest time scale, passing through game design at the medium scale, and up to narrative design at the largest scale.