Knots & Loops

Images of Videogame Narratives

The following paper was originally delivered at the Readers’ Advisory Forum of the American Library Association in Washington D.C., June 2010.

As you may notice from the title of my discussion, I have, through a minor prepositional adjustment, taken on today’s theme in a slightly oblique manner. The prescribed topic should be ‘Images in Videogame Narrative,’ a suitably rich source of conversation, since games are certainly involved in visual storytelling, but I am far more interested in the images of narrative commonplace in videogames, but marginal or otherwise experimental in other media contexts. So if I have upset you with my slight semantic deviance, I now grant you a few seconds to excuse yourself while I play a short video clip.

Greek artist Miltos Manetas created this video work, titled “King Kong After Peter Jackson.” He appropriated the source footage from a PlayStation 2 game, the succinctly-titled Peter Jackson’s King Kong: The Official Game of the Movie. You, as the player or video viewer, are sharing the perspective of character Jack Driscoll, one of several human and non-human viewpoints you take over the course of the game.

I have a number of reactions to this video. First is familiarity; I’ve played the game and seen the movie, so I recognize the characters and setting. As a game player, I’m also quite familiar with the first-person viewpoint. It’s commonplace in modern videogames, most notably in the so-called first-person shooter genre, known for its fast-paced depictions of disembodied guns traveling and firing through three-dimensional virtual worlds—typically hostile and typically alien-infested.

My second reaction (and hopefully you share this feeling as well) is that, within a few seconds, something feels wrong. Technologically, the players are life-like enough to breathe, blink their eyes, and display other convincing human characteristics (They even share some likeness with their real-life counterparts). Likewise, the surrounding jungle is rendered with discernible detail—we see vegetation hanging from trees, a shroud of fog in the distance, and intermittent flashes of lightning; we hear rain, wildlife, and the crack of thunder. Yet our two companions stare blankly in our direction, at times looking at one another, then back at us, as if to ask, “What should we do?”

For frequent videogame players, Manetas’ representation of a videogame works as an inside-joke. We understand, after years of practice, that the game’s progression relies on some kind of input: push the left stick, move forward, attack, jump, do something. Yet the video medium stymies our control. We are only allowed to watch or not, to puzzle over this strange virtual world that simultaneously ceases to proceed and proceeds infinitely without our interaction.

Manetas has produced a number of similar videos over the past fifteen years, capturing moments of rest or pause in videogames, when players relinquish control and allow the virtual world to proceed without their intervention. I don’t know if his intention is critique, comedy, or a combination of the two. Certainly, the title “King Kong After Peter Jackson” aims to poke fun at the lifelessness, artificiality, or even the exhaustive length of King Kong in the hands of its contemporary director. For my purposes, it serves as an interesting entry into the peculiarities of time and narrative in videogames.

For those of you unfamiliar with the subject, the discussion of videogame narratives was (and is) a controversial topic in computer game studies, and arguably the necessary foil that established ludology (or the study of games) as a viable academic discipline. Near the turn of the century, a series of articles written by a new crop of game-playing scholars attempted to dismantle the links between videogames and literary studies. They argued, to varying degree of success, that the grafting of ‘traditional’ media theory to a ‘non-traditional’ medium (i.e., videogames) failed to acknowledge the unique strengths and weaknesses of that medium. Thus, narrative theory could never adequately describe the types of non-narrative interactions that occurred in videogames. Espen Aarseth, in his 1997 book Cybertext, for instance, argued that narrative elements can be useful for purposes besides narrative, and their function doesn’t always imply storytelling. A World Cup match and Oliver Twist both consist of a sequence of events, for example, but the former’s events do not inherently tell a story. Aarseth coined a new term, ergodic, to describe sports, hypertexts, and computer games alike, a function that “implies a situation in which a chain of events (a path, a sequence of actions, etc.) has been produced by the nontrivial efforts of one or more individuals or mechanisms.”

Texts such as Aarseth’s sparked a division between narratologists and ludologists, with extremists on either side. A new formalism arose in videogame studies, heralding a proliferation of books and articles dismantling videogames into their ‘essential’ parts and incorporating non-narratological strategies for describing their ergodic tendencies. Thankfully, this controversy is largely settled (or at least exhausted), with the strongest proponents of either side conceding the weaknesses of their own essentialism and the merits of the opposing view. In other words, aspects of narrative theory are applicable to videogames and enrich their study, but there are also aspects of videogame play that fail to conform to established narrative categories.

My own conclusions fall somewhere in the middle, with the additional caveat that, in my opinion, the majority of videogame narratives are awful—poorly written, conceived, and executed, especially in relation to literature or cinema. Yet I also believe that this is an unfair comparison across media with divergent narratological intentions. Most sane people don’t argue about whether jazz tells a good story or not. At their best, I believe, videogames can match the intricacies, psychological magnitude, and emotional depth of an exemplary literary or cinematic work, but frequently through divergent, sometimes peculiar means.

I am going to introduce a handful of these ‘peculiar’ examples by way of narrative theory, as a means to extrude a number of models that trouble, counteract, or otherwise enhance known narrative structures. In the article “Order, Duration, and Frequency,” narrative theorist Gerard Genette introduces three problems of narrative discourse, namely time, mode, and voice. His essay only deals with the first of these three, which he further subcategorizes into order, duration, and frequency. Each category describes one portion of the spectrum of relationships that occur between the narrative (or story) and the ‘actual’ procession of events. A given narrative might proceed, for instance, through flashback or recollection, in a non-linear chronological order; its duration might span centuries or several hours in an equal amount of pages. Genette refers to Proust’s Remembrance of Things Past as an exemplary model for both.

The final subcategory, frequency, is the most interesting to our discussion, because its various forms describe some of the more experimental works of literature, yet many of the most common tropes of videogame structure. The first form of narrative frequency is what Genette calls ‘singulative’—the narrative reports a single event once. For instance, “I spoke at the conference today and changed hearts and minds forever.” Its description is so commonplace that it is widely taken for granted.

Genette’s second form is repetitive, where the number of story repetitions exceeds the number of event repetitions. For example, “Today I spoke at the conference. I spoke at the conference today. Today I spoke at the conference.” In practice, this temporal strategy borders on madness, childishness, or blind arrogance. Needless to say, it is not a common literary strategy. The final type, iterative, is a bit more difficult to grasp, as it describes the repetitive form’s inversion. In Genette’s words, “a single narrative assertion covers several recurrences of the same event or, to be more precise, of several analogical events considered only with respect to what they have in common…”

I want to illustrate these latter two types with some familiar examples from videogames. The repetitive type is a ubiquitous characteristic of videogames, persistent across changing technology and genre. The majority of videogames, whether they are narrative-focused or not, involve the (often maddening) repetition of single, discrete events. In arcade or early console terminology, these were called lives, a finite number of available chances to progress through a given game session. In the original arcade version of Donkey Kong, for example, the protagonist Jumpman (later Mario) proceeds through a series of four successive screens to rescue the Lady from the eponymous ape. The game also has a coherent in-game story of sorts, one of the first of its kind. The final (or fourth) screen involves Jumpman removing a series of rivets supporting Donkey Kong’s girder structure, causing the ape to topple and the lovers to reunite. Yet immediately after a short, non-interactive cinematic, the game loops to the first screen, this time at a higher difficulty than before, with Jumpman poised to repeat his Sisyphean task. It’s the literary equivalent of Gregor Samsa awakening in his bed again, but this time with his insect legs plucked off.

The Donkey Kong loop makes no logical narrative sense, but it makes perfect commercial sense as a machine designed to devour quarters, technological sense as a limitation of the hardware that can only produce a finite number of screens, and entertainment sense as a play experience that lasts more than a few minutes. In fact, we can say that the narrative frequency repeats recursively, at multiple interlocked levels, from the larger loop of four successive screens, to the inner loops of dying and level replay, to the loops of individual sprite animations (in other words, how individual graphics are drawn to the screen to simulate movement), and even down to the loops of code that communicate with the arcade hardware.

The repetitive narrative frequency is not limited to the early history of arcade games. It continues in countless examples of modern, narratively sophisticated videogames. Mass Effect, for example, a science-fiction based action/adventure game widely hailed for its realistic fictional world, interactive dialogue, and memorable characters, exhibits many peculiar narrative frequencies that might appear comical or insane in other media. As in most all role-playing games, if your main character (here named Commander Shepard) talks to another character multiple times, the dialogue options are eventually exhausted and the characters begin to have a bizarre, redundant conversation. Within the game’s diegetic world, this behavior is never acknowledged, mocked, or punished. It simply repeats until you choose to proceed.

The videogame equivalent of iterative frequency manifests in several ways. First, it can be used to describe a certain kind of meta-narrative that accounts for the multiple branches possible in certain videogame structures. In Silent Hill 2, a so-called survival horror game based around the narrative of a distraught widower trying to uncover the memory of his wife’s death, there are multiple potential endings to the game—by design— most of which contradict the outcomes of the others. The protagonist might commit suicide, escape the haunted town, or (I kid you not) discover that the game’s events were dictated by a dog behind a video monitor. Consequently, there are numerous iterative possibilities for a player’s individual narratives to describe the game’s fixed sequence of events. Different players uncover their individual narratives in widely different durations (i.e., some players may finish in two hours, while others might take twenty); others may never complete the game at all.

Iterative frequency also manifests in the apparently realistic simulations of virtual worlds. In two modern examples, Assassin’s Creed 2 and Grand Theft Auto IV, we can see a progression of time similar to that of King Kong. Shifting day and night cycles, weather patterns, and the day-to-day activities of non-player characters evoke a sense of immersion and realism, as if the world exists independent of the character’s actions. Yet when the player ceases input, the simulation becomes very unrealistic: the on-screen avatar stands in a fixed position while the simulation proceeds, with inhuman endurance and bladder control; day and night pass at an accelerated pace; weather shifts erratically; and the world’s inhabitants behave as if they exist in radically different time scales—some linger for hours or days in a single conversation, meanwhile ignoring the man standing in the same position for days on end.

These quirks are, of course, concessions to playability and entertainment in a videogame’s narrative structure. No one wants to send their avatar to the bathroom every few hours—with the exception of the tens of millions of people who inexplicably play and enjoy The Sims.

My point is not that videogames exemplify the literary structures of a single, influential French theorist, but that these temporal quirks point to an interesting image of narrative structure that is considered avant-garde or experimental in other media. I like to use an ancient model, the quipu, as a helpful visual reference to what I call knotted narratives. The quipu is an Incan tool for record-keeping made from strands of hair or cotton. Typically, there are one or more base strands from which other branches hang or protrude. Either may contain knots or variation in color that represent some form of unknown data. After centuries of study, we still aren’t sure exactly how they were used. For my purposes, I like that sense of abstraction, because it serves as an elegant visual model for the types of narrative common in videogames.


In Grand Theft Auto IV, there is a primary narrative strand that the player may pursue, guiding the immigrant anti-hero, Nico Bellic, toward his skewed version of the American dream. If a player chooses to follow that strand, she will experience a fairly linear progression of events, much like a traditional film or novel. Yet at nearly any time during the game, the player may diverge from the main strand and proceed along an alternate branch. These branches can take at least two forms: either a ‘side-mission’ scripted with its own narrative arc by the game’s designers, or a branch of ‘free play’ that the player ‘scripts’ for herself according to the rules of the game’s simulation. An example of the former might include taking Nico’s girlfriend out on a date in order to increase her attraction level. An example of the latter might include hijacking a car and launching it from a ramp into the ocean.

The knots along the strands represent looping conditions where a player’s progression becomes impeded, either by their inability to accomplish a task or a fail state that ejects them from the branch and reinserts them along the main branch. Of course, the metaphor strains a bit in both scale and dimensionality. A videogame’s narrative model has the potential to sustain tens of thousands of individual knots and loops along a multi-dimensional array of concurrent strands. And this multivalent aspect is one of the key characteristics that proponents of either narratology or ludology tend to overlook. As I hope you’ve seen in the previous examples, a videogame is frequently comprised of multiple, simultaneous layers of narrative, simulation, ergodic structure, and so on. All have the potential to operate independently or fold and merge into one another. This mutability and dynamism, I believe, is a source of player engagement, interactive experimentation, and often, compelling, emergent behavior within virtual worlds—perhaps even a new image of storytelling.