DHQ: Digital Humanities Quarterly
2009
Volume 3 Number 3
2009 3.3  |  XML |  Discuss ( Comments )

Avatari: Disruption and Imago in Video Games

Philip Sandifer  <philsand_at_ufl_dot_edu>, University of Florida

Abstract

This article offers an analysis of the video game in terms of the experience of gameplay, starting with the concept of the avatar, which is central to all video games. The avatar is typically described as our second self in the virtual world of the game. The paper challenges this theory, suggesting that the avatar is better understood as a set of possible interactions, and proposes a second concept, the avatari. If the avatar is our second self, the avatari is the rebellious figure on the screen that we cannot quite control, and that jumps into the pit, gets hit by the enemy, or otherwise fails despite our best efforts to succeed. Conceptualizing video games in terms of the avatar and avatari enables thinking about video games via a more sophisticated and productive model of interactivity than many of the existing paradigms.

Of late, the field of game studies has been frustratingly focused on debating whether to read video games as games or stories. These approaches have generally collapsed into the increasingly banal poles of narratology (advanced by scholars like Janet Murray) and ludology (championed by Espen Aarseth and Markku Eskelinen).[1] Both schools, upon noticing that the new medium of video games has elements of past media — games or narrative — proceed without paying adequate attention to any aspects of the medium that do not contribute to this similarity. The results are inevitably disappointing, largely because their conclusions are already in place before any actual evidence is marshaled or any argument is made, resulting in a maddening reductionism. We ought instead start with questions not about narrative or play, but about the actual phenomenon of game/player interaction. This question of interaction allows us to zero in on the interface of the video game, allowing us to view video games as a strange particular case of interactive media where the medial tools of interaction are both interface and interference. Video games, in this analysis, become a staging ground for questions of interactivity, allowing us to consider the limitations of interactivity and the imperfections of interfaces in sharp relief and as functional aspects of interactive media instead of as incidental flaws.
The central figure of my analysis will be a concept within the video game known as the avatar, as well as its mirror, which I will call the avatari. The avatar is frequently defined as the procedural representation of the player within the world of the game.[2] It thus, as Bob Rehak has noted in his own study of psychoanalysis and the avatar, “does double duty as self and other,” serving both as an element of the game and as an extension of the player [Rehak 2003, 106]. In serving this double role, the avatar functions as the contact point for the entire dynamic of player/game interaction. The avatari, which I will introduce in more detail later, serves as the avatar's opposite — the eruption of a disjunction between player and game. My contention is that the most salient feature of video games, the disrupted and frustrated attempts of the player to “beat” them, arises from the interplay of these two constructs — avatar and avatari — and that this phenomenon is essential to an understanding of video games.
I will begin with the avatar. The earliest and most common form of the avatar is simply a graphic on screen that responds directly to the player's actions. This form of the avatar is recognizable throughout video games, appearing in classic games such as Pac-Man or Super Mario Bros., and in more contemporary games like Tomb Raider and the Final Fantasy series. Indeed, it is the moment of its first creation — the moment when Steve Russell created a small blip of light on the screen that responded to user control — that Spacewar!, the first video game, took on a meaningful existence.
From that originary moment the avatar has grown steadily more complex, both graphically, as technology improved, and in terms of the complexity of controlling it, as games moved from the simple mechanics of Spacewar! to thoroughly complex games. Where Spacewar! had the basic controls of turn, accelerate, and shoot, all controlled via the existing inputs of the PDP-1 on which it was programmed, a modern video game system like the Xbox 360 has eight buttons, two joysticks, a cross-shaped directional pad, and two trigger-shaped buttons, all of which are likely to cause a given avatar to do something different. And with the addition of motion controls on the Wii and Playstation 3, the potential complexity of using a given controller becomes almost limitless. Meanwhile, a contemporary computer game like World of Warcraft will have a dizzying arrangement of commands that can be typed in, giving the game an interface more familiar to users of DOS or UNIX than to people who hung around arcades. Perhaps the zenith of this complexity of avatars comes from the 2002 Xbox game Steel Battalion, which shipped with a three-foot long controller featuring forty buttons, two separate joysticks, and three foot pedals, all with unique functions. This expansion of the avatar’s formal complexity has moved the avatar from being a single grapheme within the game to being a conceptual unity underlying the game’s entire interface. This shift from visual object to interface has even allowed games to have the avatar remain unseen, offering instead a first-person view in which the screen shows the avatar’s diegetic field of vision.[3]
While Rehak takes the idea of a first-person avatar in directions related to filmic technique, I am more interested in what this move to an avatar without a represented body means for the nature of the avatar as an object. Rehak describes the first-person avatar as being a “relatively inflexible camera, an apparatus that produced a corresponding sense of ghostliness for the player”  [Rehak 2003, 117]. This does not mean that the avatar is disembodied; Rehak emphasizes the bodily nature of the avatar in first-person games such as Quake. But it does suggest an abstract body — a body that does not claim to be immediately represented on screen, but instead symbolically present — hence the claim of ghostliness.
This opens the door to a third type of avatar, which I will refer to as the implied avatar.[4] The implied avatar exists completely invisibly, as an unacknowledged entity that actually enacts the player's commands in the game world. The implied avatar can be found in a game like Warcraft III. In this game, the player commands an army of fantasy characters attempting to destroy other armies. It is tempting to claim that the fantasy characters are the avatar, but this is a tenuous claim. For one thing, there are multiple characters, and the player frequently controls more than one at a time. For another, the death of all of the actual characters does not mean that the game ends — if there are still buildings, more characters can be created. One might also claim that this is an instance of the first-person avatar — that the player controls a general who issues commands to the various units. This claim is supported somewhat by the fact that the game occasionally directly addresses the player by giving them orders. But this doesn't make sense either — no general can physically see the entire battlefield, as the player can, or directly oversee two battles happening in totally different places. One might also claim that there is no avatar in this game, but that doesn't make sense either — something clearly controls the armies, and it is not, diegetically, a 21st century person.
No, Warcraft III, in order to make any sort of coherent sense, requires the assumption of some agent that serves as an intermediate point between the player's physical actions (moving the mouse) and the results in the diegetic world (troops moving). This agent is the implied avatar. The game relies on the existence of some being that exists in the game world issuing commands and controlling things, but it is not embodied within any single graphemic object. This requires a fundamental shift in our understanding of the avatar. Rather than treating it as a graphemic object or as a particular character in a narrative, we must treat the avatar as a general category of actions that can be taken on the player's command. Thus we recognize the avatar of Super Mario Bros. not as Mario Mario, an Italian plumber, nor even as a particular configuration of pixels, but as a subset of moves within the game — moving a specific sprite to the left or right, making it jump, etc. The avatar is, in this case, less an object or body (as Rehak would have it) than a function — the system of allowances and formal constraints that constitutes the player's interactions with the game. The avatar may be graphically represented by a diegetic character, and this representation has significance in interpretation of the game, but the avatar is not the diegetic character as such. Rather, the avatar is a complex figuration — not only a set of graphic elements that signify a diegetic entity, but also a system of rules and capabilities that describe the interactive capabilities of the player.
By considering the avatar not merely as a fictional character but as a function within an interactive system we allow for an analysis that not only effectively handles games with plots and characters, but also handles a game such as Tetris that lacks both. After all, regardless of what purpose the Tetris blocks exist for, we do not treat them as objects that move themselves.[5] As a system, Tetris contains a relationship between things the player controls and things the player does not. Broadening the notion of the avatar allows us to consider this relationship in context with the diegetic division between “your character” and the world that exists in more narrative games. This helps contextualize video games in terms of interactive media in general by connecting the avatar with the broader question of the connection between graphemic elements and interactive capabilities within interface design.[6]
It is important to note, however, that the abandonment of analysis of avatars as a component of diegesis does not remove the issue of motivation and desire. That is, even if we abandon the claim that the avatar of Super Mario Bros. is a fictional character who lives in Brooklyn (as Mario canonically does), we need not — and, in fact, if we are going to understand the game coherently, cannot — abandon the notion that the avatar holds as its end goal the beating of the game, defined, in this case, as the liberation of Princess Toadstool from Bowser. This desire stems not only from the diegetic implications of the avatar but also from the functional system that constitutes the avatar — the fact that Mario can ultimately only move towards Princess Toadstool necessarily creates a goal of rescue.
This is particularly crucial to the implied avatar, which has no narrative component to it, but is still an agent with a particular will, and a particular set of desires. This ascription of desires is at times vital to an understanding of the game. Since the goal of SimCity, for instance, is unlimited expansion of the city's resources, its avatar — the portion of its system that is under the control of the player — can be read as adhering not only to the rules of gameplay, but to the ideological rules of capitalism — seeking to unlimitedly expand itself via the conversion of currency to material goods that produce further currency, and causing the organization of the city to structure according to its will. Notably, however, the avatar is also not the entire game. That is to say, the avatar's will does not simply reign unchecked in the game — Mario does not immediately (or, in practice, always) rescue the Princess. In SimCity, sometimes the unchecked expansion of capital is curbed, whether by player incompetence or by Godzilla. Thus the avatar is always paired with its opposite — the avatari. But before the nature of this opposite can be explored, it is important to see the reasons why existing video game scholarship has failed to address it.
As much as the avatar may be a systemic function of the game, it is important to note that it is not merely a product of the video game's code. It is not an aspect of the game, but rather the point of contact between the game and its player, constituted in the specific act of play. As such, the form and function of the avatar can only be understood through the mechanism of play as an actual phenomenon. It is here we can see the flaw in much existing scholarship on games. This flaw extends across multiple approaches to video games. In the ludology camp, it manifests as zeal to excise all narrative from their discussion. Espen Aarseth, through his determination to understand games as something other than texts, commits the most flagrant version of this error, when he claims that the world of the game is “coincidental to the game”  [Aarseth 2004, 48]. Aarseth’s claim is that, even if you completely change the world of the game, if the underlying game mechanics remain the same, it’s still the same game. In other words, the question of desire is, in Aarseth’s view, determined entirely by the formal properties — the story of Super Mario Bros. is irrelevant, and the goal of rescuing the princess comes entirely from the game’s structural elements. This is consistent with his unwavering criticism of a literary approach to game studies, characterized by references to the literary interest in video games as a “land rush” where “respect for local culture and history is minimal, while the belief in one's own tradition, tools, and competence is unfailing”  [Aarseth 2004, 45].[7]
But this argument does not advance itself as well as Aarseth wants it to. Indeed, when he reaches for an example of this literary error, he attacks the frequent commentary on the size of Lara Croft’s breasts in the popular game Tomb Raider. Aarseth argues that “I don’t even see her body, but see through it and past it”  [Aarseth 2004, 48]. The problem arises when one realizes that, in looking through and past Lara Croft's body, Aarseth is also looking through and past the vast cultural paratext surrounding Lara Croft, which makes it clear that many players do not.[8] The result is that he also avoids actually talking about the general condition in which Tomb Raider is consumed by its devotees. Thus Aarseth's analysis ultimately commits the exact error he decries — a failure to acknowledge the existent contexts of his central texts. The Tomb Raider that Aarseth engages with is a theoretical construct that is necessarily separate from any actual phenomenological experience of play in which Tomb Raider could become manifest to an actual playing subject. Any such subject unlike Aarseth’s theoretical structure would be firmly enmeshed in the social paratext surrounding the game and is inevitably influenced by it.[9] Thus, despite his determined focus on the systems and rules that constrain and shape the act of play, Aarseth's approach forecloses discussion of play as an event that actually happens.
This avoidance of the phenomenological experience of play is not unique to the reductivism of the “it's just a game” approach however, as evidenced by Janet Murray’s Hamlet on the Holodeck, a book deeply invested in the idea of games as narratives. Murray's book is lively, mostly thorough, and presents a clear vision. The vision, however, is not one of any video games that have ever existed. Rather, Murray's vision of video is put primarily in terms of the future: technologies like virtual reality or the Holodeck that do not yet, and may well never exist. Murray describes three aspects of what she calls cyberdrama.[10] They are immersion, agency, and transformation. The most important of these, in many ways, is transformation, and the other two largely stem from it. Transformation, according to Murray, is the way in which the reader/player of a cyberdrama becomes another character — in other words, it is the process by which they become an avatar. I say the other two stem from it because this becoming more or less requires that our avatar do something, and that it have a world to do it in — after all, in Murray, a player becomes something else by having “the opportunity to enact stories”  [Murray 1997, 170].
I should note that my objection to Murray is not that she utilizes narrative concepts in her analysis. This is a productive course of action — just as Aarseth's focus on questions of play is productive. The problem is when the focus on narrative becomes a tool that is employed less to understand the game than to exclude other understandings of the game. In Murray, this problem comes from her investment in a futurism of video games. Murray is not talking about video games as such. Her goal is to explain “what delights or dangers digital narrative will bring to us”  [Murray 1997, 94]. Video games are almost beside the point to her — she refers to them as “Harbingers of the Holodeck” in one of her chapter titles. That doesn't stop her from deploying them regularly as examples to prove her point, though. In short, she's committing an error much like Aarseth’s when he overlooks Lara Croft's breasts — she's declaring that games are part of the development of this futuristic medium, and immediately loses interest in all aspects of games that do not contribute to this — including several that are vital to any understanding of such games. Ultimately, Murray all but admits this, noting that “the term cyberdrama,” which she has spent most of her book defining, “is only a placeholder for whatever is around the corner”  [Murray 1997, 271]. As a result, Murray does not take the road that a satisfying narratology of video games would take — one that works and reworks narrative concepts to serve video games. Rather, she works and reworks video games to serve her futurist vision of narrative.
This problem, admittedly, goes a lot deeper than Murray — the desire to describe video games in terms of virtual reality, the Holodeck, or other futuristic devices is endemic to new media studies at large. It is, for instance, very much important to Marie-Laure Ryan's Narrative as Virtual Reality, which centralizes the notion of immersion in the study of virtual reality. Likewise, the notion of fully immersive media is central to Bolter and Grusin's dualism of immediacy and hypermediation [Bolter & Grusin 2003, 5–6]. Outside of video game studies as such, the futuristic model seems almost to be the default assumption. For instance, Slavoj Žižek describes a hypothetical video game “allowing the player to assume the role of the [Washington D.C.] sniper”  [Žižek 2004, 190], employing the notion that one assumes roles and identities through video game playing. The notion is equally prevalent in art, with films like Tron, eXistenZ, and The Last Starfighter all heavily emphasizing the video game as a gateway to an alternate world.
There are numerous problems here. For one thing, it puts the cart before the horse. The nature of video games is unlikely to be defined primarily in terms of a future medium that did not exist when they were created. More likely, the nature of that distant medium is going to resemble an older one. And, more to the point, it's very difficult to talk about virtual reality without simply indulging speculation and fantasy. As the chronic failures of futurists to in any way accurately predict the future tend to suggest, should any of these fully immersive media ever come into being, they will probably not be what we currently envision them as. Put bluntly, Murray's approach, with its continual turning towards imagined bold and futuristic forms of narrative, has far too much potential to be proven silly and embarrassing when the Holodeck never comes around, or when it is radically different from how we have predicted it.[11] Instead of waiting for the potentially non-existent Holodeck, we should instead note the forms in which video games really do exist, and study them.
More generally, to my mind there is a fundamental flaw with many methodologies that invest themselves in questions of the sort of world that video games create. Even in the hands of a critic such as McKenzie Wark, who is uninvested in the narratology of the medium, this approach suffers from an unhelpful reductionism. In his book Gamer Theory, Wark focuses his inquiry on the social consequences of gaming — on the way in which the gamespace, as he calls it, serves as an allegory for the real world. As Wark puts it, “gamespace is now the very form of the world”  [Wark 2007, 1c7]. Though undeniably provocative, Wark's proclamation seems to me to skip a vital step. For all his nuanced examinations into what gamespace does and what it is an allegory for, and for all his attempts to derive a critical theory of games, Wark is surprisingly uninvested in the actual act of playing the game.
I do not want to understate the strangeness of this. One of the most remarkable things about Wark's book is the lengths he goes to in the book to solicit feedback from actual gamers and to validate his conclusions. The entire format of the project, as an online book that people could comment on and that he would subsequently revise to respond to comments, seems designed to prevent exactly this sort of error. Which is what makes it so strange when Wark proclaimed in an exchange with a reviewer that “this book is called Gamer Theory, not Player Theory”  [Kucklichs 2007]. Here, in a situation where it seems as though everything that could possibly work towards attention to the actual experience of play is present, Wark fundamentally rejects the very notion of studying that act of play.
This is not an idle or accidental rejection — Wark accuses play and related words of being “code words for the labor of validating the commodity”  [Kucklichs 2007]. But it is a rejection that has consequences. These consequences can be seen most clearly when Wark engages the fact that the video game is, by definition, deterministic. Video games, after all, always have code, and will always respond identically to a given input. Wark refers to this code as the game's algorithm, in part so he can contrast the word with allegory. Regarding the algorithm, Wark declares that “what is distinctive about games is that they produce for the gamer an intuitive relation to the algorithm”  [Wark 2007, 30]. This figure of the master gamer, intuitively linked to the algorithm, recalls Wark's earlier figure of the hacker, about whom he says, “Differentiation of functioning components arranged on a plane with a shared goal is the hacker achievement”  [Wark 2004, 83]. Both depend on a vision of mastery — of a gamer or hacker who has seized a fundamental, essential level of power within the system they interact with.
But these two figures share an unlikely cousin — Janet Murray's vision of the reader of a Cyberdrama. In her chapter on immersion, Murray talks about the ways in which the reader of a cyberdrama — she uses the term interactor — is an author of the text. Murray writes of the way in which the cyberdrama's author — its programmer — creates “the rules for the interactor's involvement, that is, the conditions under which things will happen in response to the participants' actions”  [Murray 1997, 152]. In other words, the game's programmer creates the algorithm, and the interactor follows within it. But of this, Murray says, “the interactor can experience... the thrill of exerting power over enticing and plastic materials”  [Murray 1997, 153]. In other words, they can have mastery over the algorithm. Despite the fact that he could not be further in attitude from the utopian vision of Janet Murray, in the end, he falls into the same ecstatic assumptions.
The problem with these ecstatic assumptions is that they attempt to transform the medium of the video game into something immediate. If playing a game transports us into a virtual world or gamespace then the game that creates this world cannot be a medium, much as Raymond Williams argues that language is not a medium [Williams 1997, 158–9]. Lev Manovich recognizes this in The Language of New Media when he talks about the “myth of interactivity,” pointing out that the notion that a new media hypertext does not, in fact, allow interactivity — rather, it curtails it by fixing the number of possible associations that one can have with a concept. He uses the example of a hypertext, in which only a select number of the words or phrases actually produce branching thought and texts [Manovich 2001, 61]. Even Wark recognizes this, noting that “All games are digital. Without exception. They all come down to a strict decision: out or in, offside or onside, goal or no goal”  [Wark 2007, 79]. But this argument can just as easily be applied to a video game — even in one with as much alleged freedom as Grand Theft Auto III: ultimately one's options boil down to driving around within the city and beating people up. You cannot renounce your life of crime, murder specific people outside of the limited set of missions, enter a store and buy a soft drink, or any number of other things. The freedom offered by Grand Theft Auto III is simply the freedom to kill or not kill at a given moment. This freedom is couched in an elaborate network of potential actions, but the world is inexorably oriented towards the inevitable teleology of violence. Manovich likens this situation to the Althusserian concept of interpellation, claiming that “we are asked to mistake the structure of somebody's else [sic] mind for our own”  [Manovich 2001, 61]. Certainly, we can reject this to some extent — one could simply drive around Liberty City listening to the radio, and obeying all the traffic laws. But to do so amounts to putting the game itself — with all of its well-defined missions and goals and purposes — on hold. One cannot drive around Liberty City lawfully without being tacitly aware that one is rejecting the system that the game presents. The player is capable of rejection, not of denial.[12] What happens in the video game is not that we have a second self in a second world, but rather that we are presented with a separate consciousness (or, rather, an imitation of a consciousness) with its own characteristics distinct from ours.
But this should seem clear-cut merely on the linguistic level. After all, if the avatar is actually equivalent to the player, we wouldn't call it the avatar — we’d call it the player. That we call it the avatar, however, means that it is in a fundamental way not equivalent to the player. Any transformation that might be occurring has in fact been a failed transformation — a transformation that aborted somewhere along the line, leaving the player in some state between being completely separate from the avatar and completely merged. What we need, then, is not a concept of transformation but of becoming — a conception not of a thing, but of its disruption.
The obvious thing to turn to is the Heideggerian distinction between object and thing. Heidegger notes that “when we discover its unusability, the thing becomes conspicuous,”  [Heidegger 1996, 68] and in doing so reveal itself more truly than if it were simply there and functioning. In other words, in the Heideggerian view the act of disruption and breakage is a revelatory action — one that in turn makes us aware of the relationship between the broken thing and the world at large. This enables us to identify the point of disruption — the disjunct between the player’s desire and the schema of action delineated by the avatar-function — as a point in which the game genuinely manifests itself as an experience.
In practical terms, this is easiest to notice by going back to the early history of video games. It is difficult to play Space Invaders or Pac-Man today and get any sense of immersion or transformation — the graphics are simply too blocky, the controls too awkward, and the game in general too primitive.[13] For today’s player, the game is simply too awkward an object to present itself handily — it is necessarily conspicuous. But, notably, this does not actually inhibit play — the games remain quite entertaining even when they are conspicuous. In this spirit, I wish to propose a second notion — one that exists in parallel to the avatar. This is the avatari: the dark mirror of the avatar.[14]
If the avatar is our representation in a virtual world — the thing that we control, our second self — then the avatari is the blocky, poorly defined image that won't go where we tell it to. The avatari is what manifests when we can't make a jump in Super Mario Bros., when the pieces come down too quickly to manage in Tetris, or when we can't make sense of what a clue means, what an object is, or how a game is supposed to work. In these moments the interface of the game turns to interference, making itself conspicuously present in its difference from what the player is trying to do. I propose that these moments — the ones where the game resists becoming an integrated part of our consciousness and being — are the moments in which the game is most recognizable as a game being played, and thus the point where the game is most visible as a mediating agent between the player's subjectivity and the ideal of gameplay to which they aspire.
It's important to recognize the avatari as a midpoint — it is not the immersive and transformative avatar, but it's also not an instance of complete breakdown. My point is not that one should play video games with the controller unplugged in order to maximize the avatari function and thus to get the fullest experience. Rather, it is that the avatari exists between these two extremes: at the point of disruption. Disruption can only occur if there is some functioning thing that is being disrupted — a word that, like eruption and rupture which both also make good words to describe the avatari, comes from the Latin for “to break.” The avatari is just that — the moment of breakage, at the transition from functionality and control to anarchy and dissonance. Its presence serves as a further sign of the disjunct between video games and the larger image of virtual reality. This also serves as a major challenge to critics such as Murray and Ryan. Although Ryan (who is the savvier of the two in this area) acknowledges the problematic aspects of immersion, and wisely notes that the term has been unfairly demonized in literary studies [Ryan 2001, 9–11], she still situates it, along with interactivity, as one of the two fundamental aspects of virtual reality. Her notion of immersion, however, still fundamentally depends on the idea of being “transported into the virtual reality of a textual world”  [Ryan 2001, 99]. Even ignoring the many problems with treating the textual world as a coherent place that we could enter, this idea does not quite hold. The video game's most primary function — to provide a challenge to its player — depends on the explicit and active prevention of the act of transportation that Ryan attempts to centralize.
The avatari, then, is that element of the game that is continually frustrating and blocking the ability of the player to simply append the game and controller to our bodies and selves. If the video game controller is an extension of our body, as McLuhan would have it, then there also needs to be an amputation. The medium functions by continually turning that part of our body against us — in a scene similar to the one Žižek describes in his discussion of the movie Fight Club — in which a part of the body, in Fight Club the narrator's hand — attacks him, seemingly with a mind of its own. At this moment, the hand becomes, in Žižek's words, “an organ without a body.” Žižek argues that this “provides the key to the figure of the double... the hero's Ideal-Ego, a spectral/invisible hallucinatory entity...inscribed within the hero's body itself as the autonomization of one of its organs”  [Žižek 2004, 173–4]. Again, given McLuhan, this seems like it can perfectly describe the video game in the case of the avatari. If the avatar is our second self, then the avatari is its manifestation as a thing — in which our second hand turns on us.
Consider an example. You are playing Super Mario Bros.. Controller in hand, you are leaping gamely upon enemies, hopping pits, doing well. The controller is an extension of your hands, and it does your will. Then, you come upon a pit, jump to cross it, and fall short — Mario perishes. At this moment, the controller betrays you — turns on you, destroying your progress. You recoil slightly from surprise. Why? Because the controller, your new appendage, has suddenly betrayed you and become alien to you.
But what, in this analysis, is the double? It's not the avatar, which manifests through the sort of interpolation that Manovich identifies. But all the same, it has emerged, like the avatar, from what Wark calls the algorithm of the game — the structure and rules that make up the player’s interaction with the game. It is for this reason that I call the avatari the mirror of the avatar — because both arise together, out of the same process. Looking again at Grand Theft Auto III, the same avatar system that constrains the player into an available set of actions that point inevitably towards violence is the system that forestalls that violence through the obstacles that are built into the system, and, indeed, into the notion of violence itself. Thus the structure that Grand Theft Auto III invites us to mistake for our own mind is both avatar and avatari. The two are inexorably linked — the game depends not on the triumph of the avatar or the avatari, but on the continual balancing and counter-balancing of them that comes from extended play. This is, after all, the point of the video game — it wants to suck up your time.[15]
Eventually, certainly, you will beat the game, and it will display some congratulations, often even thanking you for playing. But these sentiments are insincere to the extreme. Some games are even explicit about this, offering you a harder “second quest” or challenging you to go back and try again to find more items. The best games, like Tetris, take this to an even greater level, seeming, as you do better, to deliberately undermine you, both by going faster and faster, and seeming to always give you the exact wrong block. The challenge offered by the video game is the elusive “perfect run,” and it can only cause you to keep playing if it makes that perfection ever harder to obtain.
It is here that we can see what really distinguishes video games from other games. The failure involved in the manifestation of the avatari goes beyond that of other failures. It is not merely an issue of failing to achieve something like missing a putt while golfing. It is not even an issue of being beaten by something, of being surpassed by your opponent in chess. It is, rather, an issue of something that has been integrated as a part of you betraying you and defeating you — as though your chess pieces came to life and began making moves that would lose the game. This is why the most popular computer game ever — the version of Solitaire that has been included in every version of Microsoft Windows since 1990 — does not seem to me to be a video game as such. Solitaire is merely a simulation of a regular game — it does not involve the design of a system that is deliberately constructed to undermine you.
Of course, it does not appear that my copy of The Legend of Zelda is likely to come to life. It is my conscious decision to plug in my controller and start trying to control the game that creates the double that's going to betray me. That is to say that to play a video game is to open one’s self up to this potential self-betrayal — of voluntarily annexing an appendage that is going to misbehave. Of course, the same could be said of many games. Golf, for instance, involves annexing an appendage — the club — that you know, generally, is not going to make the ball go exactly where you want it and is generally going to frustrate you until you break it in half over your knee and throw it into the nearest water trap. The point of games, after all, is to be challenging.
What distinguishes video games from regular games, however, is that the video game involves some appearance, as Murray observes, of transformation. That is to say, golf simply involves trying something as yourself and failing. The video game depends on a transformative move — an attempt to become the avatar, which is then frustrated. It is not merely that the video game is hard — it is that the video game has a designed inadequacy to its control scheme that visibly and actively defies attempts to master it. The best analogy among regular games would be a rigged carnival game, or a game of three-card monte, where the game is no longer merely a matter of skill. The attempt to master the video game is doomed, and we know it, but the existence of the attempt is crucial. Central to this attempt is the fact that the video game, at its best, is designed in such a way that no mastery over it is possible. In the best video games, no matter what, the avatari will manifest eventually. The battle of the video game is not to master it and to gain control — it’s to keep from losing control for as long as possible.
In this regard, the medium of the video game is basically the material-circumstantial support of an extended symptom. Playing the video game is a continual re-enactment of the Kleinian schizo-paranoid position, in which a partial object (in this case the material components of the game — controller, screen, interface, etc.) is split into good and bad components — the good game being integrated as an appendage to the body, and the bad game being thrown across the room in frustration, or, equally frustratingly, teasing you with a constant reminder of how far you still are from a perfect game [Klein 1975, 264–8]. But this split game is inconveniently rejoined in a single material form. The game is thus cast out and integrated simultaneously — a double action that constitutes the fundamental act of play.
The avatari, it should be stressed, is not merely the failure that triggers the negative component of this double action. The avatari is a part of the game, embodied at specific points in the material manifestation of it. Consider Super Mario Bros.. Ostensibly, the purpose of the game is to get Mario to Princess Toadstool — an accomplishment that, essentially, involves running a certain distance in a more or less straight line. And it should be noted, this is clearly the desire of the game — the game is set up so that this desire is continually acted upon, with the avatar even being forbidden from backwards motion. But despite this, the game deliberately introduces obstacles to this desire — pits that must be jumped over, enemies that must be avoided or killed, etc. These enemies bring the avatari to bear on the game by disrupting the execution of the avatar's desire. Furthermore, since the game is treated as interactive, the game is considered by the player to have agency. Thus it is not merely that the game is designed to contain its own disruption, it is that the game is an agent that brings about its own disruption.
Perhaps the best example of this conscious design of the avatari is Tetris. On initial inspection, it might seem as though the way in which the avatari manifests in Tetris is through its ever-increasing speed. But there is an upper limit to the speed of Tetris, and no shortage of players who can play well even at that speed. But eventually, even these great players will lose. There is also the speculation that the game deliberately tries to sabotage you with pieces you can’t use. Although this is a satisfying explanation for failure, it’s an untrue one — the piece selection is, in most versions of the game, truly random. The real problem with Tetris, as it turns out, is that it's actually impossible — it has been mathematically proven that no matter how perfectly one controls the pieces, it is impossible to determine any method for where to put them — even if one knows all of the pieces that are coming in advance [Breukelaar et al., 2002]. In other words, in Tetris the avatari manifests simply because the task you're trying to perform is impossible — the nature of the task itself is always to undermine you — to spiral out of control. It is particularly key to note that this impossibility is not a point, but a gradual slippage. That is, it is not that there is some barrier beyond which one cannot pass, as there is with golf where you cannot actually do any better than a hole in one. Rather, the game will, no matter how well you play, always slowly get away from you. No amount of training, improvement of skill, or anything else can forestall it. This fact makes Tetris unique among puzzle games — even a similar speed-based puzzle game like Bejeweled fails to have this active and final undermining that makes Tetris more than an avatar-free game like Solitaire.
A more subtle variation on the inadequacy of control is the deliberate introduction of weaknesses in the control scheme. The most simple of these is the convention in many side-scrolling action games such as Mega Man, Castlevania, and Metroid whereby the avatar can only fire in a finite number of directions — usually just left and right. Although seemingly a small thing, this is an extremely potent restriction, particularly given that all of these games feature a number of enemies that attack from above, where it is difficult to shoot them. This provides a number of occasions for the avatari to manifest in these games. Other similar and common weaknesses exist, such as various levels of inaccuracy in jumping, ranging from the removal of the already physically problematic ability to change course in mid-air to more elaborately awkward jumping motions. A particularly tricky one is in games like Castlevania where, when your avatar is hit by an enemy, he jumps backwards slightly — an uncontrollable phenomenon that can prove fatal if the hit occurs while the avatar is in mid-air or on a narrow ledge. In all of these cases, even though the avatar is what the player seeks to control, the control given is made weak in fundamental ways.
An interesting example of this sort of gap in control comes in many fighting games such as Street Fighter 2. The basic concept of Street Fighter 2 is simple — you control one character, and another player or the computer controls another. The two characters attempt to beat each other up. Eventually, one wins. The controls for this game are straightforward — each of the six buttons on the Super Nintendo controller either throws a punch or a kick. But what is interesting about Street Fighter 2 is that, when your character is actively being pummeled by your opponent, you, to varying extents, lose control. It becomes impossible to throw punches or block while you're under direct attack. Furthermore, when you've been hit a lot of times in rapid succession your character becomes stunned and is unable to move for a period of time, making it even easier to launch another wave of attacks. Particularly valued by Street Fighter 2 players are combos — sequences of attacks that, when performed correctly, give the opponent no opportunities whatsoever to respond. In the earliest versions of Street Fighter 2, one character, Guile, could be used to launch a sequence of attacks that could take an opponent's health from full to zero without any opportunities to respond.
This is a particularly striking effect because fighting games such as Street Fighter 2 are generally characterized by an extremely fast pace, with players pushing buttons on the controller frequently and rapidly — so much so that many of the games are referred to as “button mashers.” When the avatari manifests in these games and you are unable to control your fighter the physical interaction with the controller does not slow. In fact, moments like these tend to result in the player mashing buttons even more frantically as they try to regain control, leading to an effect whereby the more one interacts with the controller, the less one interacts with the game.
The effect is interesting when Street Fighter 2 is considered in terms of its role as a popular two-player game. In the two-player game, each player controls one character and they fight. As a result, the loss of control is caused not by the video game system but by another person. But the appearance of the avatari is not changed — if anything, human players are more likely to use lengthy combos that prevent their opponents from reacting. This clarifies something significant about the avatari — it is not merely something the game creates and confronts the player with to frustrate the player's progress. It is something that arises out of the game as a gestalt — from its control scheme, from its diegetic situation, from its algorithm, from the material controllers being used, and, in multiplayer games, from its other players. The relationship between the avatar and the avatari cannot be collapsed into the relationship between the player/avatar and the enemies, obstacles, and puzzles. The enemies, obstacles, and puzzles in a game are merely one of many methods used to disrupt the link between the player and the avatar. This disruption, not its cause, is the avatari.
Still, it is worth looking at what generally causes the emergence of the avatari. As I have said previously, the avatari is best understood via its relationship to the idea of the interface. That said, the avatari does not generally arise from what is normally considered bad interface design. That is not to say that there are not bad interfaces in video games. But the design principles of a video game interface are not substantially different from those of an interface that is intended to be used seamlessly. Donald A. Norman, in his seminal The Psychology of Everyday Things, lays out a simple two-principle system for usable design. The first principle is that of mapping, which he describes as “the relationship between two things, in this case between the controls and their movements and the result in the world”  [Norman 1988, 23]. He uses the example here of a steering wheel, in which turning the wheel maps logically to the motion of the car. The second principle is that of feedback — “sending back to the user information about what action has actually been done”  [Norman 1988, 27]. Norman’s view is that these two principles must be followed to make a design usable — that is, the user should be able to translate their desire into actions, and receive confirmation that what was intended did in fact happen.
This observation allows us to understand how even a game with no obvious awkwardness in its control scheme such as Super Mario Bros. can manifest the avatari. In Super Mario Bros., the directional pad adequately maps the basic motion of Mario, save for the possible curiosity of “up” not being equivalent to “jump” — an illogical, but sensible exclusion given the importance of adjusting for direction while jumping. Two buttons control the rest of Mario’s movements — one runs, the other jumps. The arrangement of these buttons — running/firing on the left-most button, labeled B, jumping on the right-most A button — is done so perfectly that it has remained the standard for such games across every system since. The logic of this layout is simple — one generally wants to run and shoot almost constantly in most side-scrolling video games, and the thumb can roll easily from holding down the B button to pressing A without releasing B. Similarly, the game world provides good visual cues for key elements of the game. The green pipes sticking out of the ground provide a good map for the concept of going down. The bottomless pits, interrupting the familiar floor, are visibly dangerous. The item blocks, marked with question marks, invite the player to investigate. And, in what was one of the game’s most technically innovative features, the directional pad that controls Mario’s movement remains useful even when Mario is jumping. Unlike the game’s most immediate predecessor, Donkey Kong, Mario’s jumps are not simply parabolas: the directional pad continues to control his movements even in mid-air — an ability that is crucial to beating the game.
The game is similarly impressive in terms of feedback. For instance, it provides good aural cues to let the player know that she has done something. When the player presses A to jump, the game emits a small springing sound as Mario jumps — a sound that remains one of the most recognizable sounds in video games. When Mario grows more powerful for getting a mushroom, he visibly grows in stature. And, on a basic level, because Mario’s response is basically instantaneous, there is always a good layer of feedback coming from the game.
None of this is to say that Super Mario Bros.’s interface is flawless. It is not. But it is a very good interface. Mario generally responds predictably and precisely (at least within the bounds of the game’s physics — Mario does not turn or stop on a dime, instead being carried along by a measure of momentum). What brings forth the avatari in the game is not a badly designed interface, but an inadequate authority on the part of the interface over the world it is trying to control. At its heart, Super Mario Bros. is a game about moving a fast, dynamic character through a set of hazards. The interface allows precision, but the game presents the player with scenarios where careful and precise movements must be made at a high speed in order to survive. In one of the most memorable sections of the game, toward the end, the player must successfully navigate a series of precisely timed jumps across very short ledges while running quickly so as to minimize the amount of time the player can spend on a ledge. It is a notoriously difficult jump — sufficiently so that the game is unusually generous, placing a relatively easy to obtain 1-up shortly before it so that it can be attempted repeatedly. But the challenge comes not from the fact that it is difficult to make Mario do what the player wants, but rather that the action necessary to get across the pit is complex even though all of its components are utterly intuitive.
It is perhaps worth discussing more accurately what sorts of potential manifestations of the avatari are generally avoided by video games, or, when they are not avoided, are generally found unsatisfying. One common example from the side-scrolling genre is the problem of evaluating whether something is safe to stand on or not. Most games will use some sort of general visual signifier that a deadly platform is deadly — visible electricity, fire, or, perhaps most common, spikes. The player intuitively grasps that such surfaces are dangerous. But in Nintendo’s original Donkey Kong, this design technique had not been developed. As a result, the third level, in which the player must navigate a series of elevators, has an unpleasant surprise in the form of a floor that it is fatal to touch. Nowhere previous in the game had anything suggested that the floor might be deadly, and the floor is visually indistinguishable from the safe floors in the level. The resulting death, then, is enormously frustrating for a new player because it seems unfair. The game has given the player no reason to believe that the action they are undertaking is dangerous. This death is an emergence of the avatari, but it is an unsatisfying one, coming as it does from what seems like an unfair trick on the part of the game. The player has been taught a particular mapping of the interface — stay on the floor — only to have that mapping abruptly fail.
A similarly frustrating emergence of the avatari, this one an error of feedback, comes in the adventure game Return to Zork. In this game, early on, the player encounters a plant growing in the dirt. The player can interact with the plant, and several options present themselves, including digging it up and cutting it down. Both of these options result in the player taking the plant with them. But the latter option — cutting the plant — turns out, considerably later in the game, to doom the player to failure because the plant is dead when the player needs it to be alive. Nothing when the player cuts the plant indicates this, and the player has no knowledge when they encounter the plant of what its purpose is. There is, at the time of action, no feedback letting the player know that something has gone wrong. But if the player has made this mistake early in the game their game is effectively over — to beat the game they will have to start from scratch. Such errors certainly have their enthusiasts — the world of interactive fiction is full of people who firmly and passionately believe that games should be aggressively punishing like this. But for the majority of players such an error is merely irritating, and such errors are rarely allowed in games that are intended to be broadly popular.
What is interesting here is that the video game is, when done right, an instance of an interface that is easily understood and manipulated, but fundamentally inadequate to the task at hand. Despite the fact that games are meant to be challenging, a large amount of effort in the design goes into making the interface easily learned and understood. Valve Software’s recent game Portal features, after the game is beaten, a “developer commentary” mode where the player can replay the game with, essentially, a DVD commentary track. This track is fascinating in that it focuses primarily on the way in which the game is carefully structured so that players will gradually learn the interface. The amount of work and detail that the developers have put into this task is staggering: puzzles were carefully designed to not be passable without abandoning specific bad habits that were identified during playtesting, small aural and visual cues were inserted to get players to look in particular directions, and, in one memorable instance, a design decision made to discourage a particular player mistake led to one of the game’s most popular characters, the Weighted Companion Cube. What is remarkable is that even with the obvious effort that went into teaching the player how to play the game intuitively, the game remains genuinely challenging and enjoyable, and was virtually universally praised.
Understanding the emergence of the avatari in terms of a product of successful interface as opposed to a poorly designed one is crucial in seeing both the significance and weirdness of the concept. The shift is significant because it allows us to use video games as a productive model for interactive media at large, inasmuch as the interface and process of interaction become conspicuous through their lack of handiness. This conspicuousness works to bring the overall structure of the game into sharp relief, however, precisely because it exists in contrast to an equally viable and carefully designed dynamic of usability. The video game is both usable and impossible. This dynamic has been observed by Sherry Turkle, who argues for a homogeny between the way in which players of computer games learn the games — “an intuitive sense of what will work without understanding the rules that underlie the game’s behavior” — and what she views as a lack of understanding of the computer systems in widespread use today [Turkle 2005, 268]. What is so interesting about video games, though, is that this lack of understanding is a designed feature. The games are not meant to be understood. Or, more precisely, they are meant to be misunderstood. The player is led towards a failure that is not a failure of design, but an altogether subtler failure of misrecognition.
This is, for instance, what Tetris depends on. The mathematical fact of its impossibility is far removed from the actual heuristic of play — in no small part because its mathematical impossibility calls not for a sudden death point, but for a slow inexorable choking off where, even if the long-term prospects are dim, there is always a moment-to-moment hope that one can make a decision that will extend rather than cut short one’s play time. But more to the point, few people who play Tetris are aware of the mathematics of the game. Thus the game works by misdirection: by presenting one scale of increasing difficulty, the steady speeding up of the blocks, that appears to be blamable for a failure that is really due to a mathematical impossibility.
Another type of example comes when the avatar and what the player tries to control do not match up exactly. For example, in SimCity, your goal is to build and expand your city. The means by which your city actually expands is through the development and expansion of its various SimCitizens. But you do not control these citizens. Instead, you are given control over the physical construction of the city, and over its tax rates. You have perfect control over these things, to the point of being able to freeze time and have a highway or an airport spring up and be fully functional instantaneously. However, these things are not the things you are actually trying to control. Certainly, the physical construction of the city influences the actions of its population. But it's an indirect influence, and, more to the point, an imperfect one. In this case, the avatari is explicitly coded into the game in the gap between what you are able to manipulate and what you are supposed to control.
An example of this interfering interface comes in Nintendo's Pikmin. In it, you are trying to get a horde of up to 100 creatures called Pikmin to perform various tasks for you. Drawing from a tradition of games going back to Lemmings, in Pikmin you do not directly control the Pikmin. Instead, you control a small spaceman named Captain Olimar who walks with the Pikmin. As the spaceman, you have certain interactions with the Pikmin available to you — you can throw them, summon them with a whistle, or tell them to stand still. You can also herd them. But this action is generally inaccurate. One is, after all, controlling a crowd of as many as 100 Pikmin. And so when one herds them in a given direction, there is a certain amount of error to the action — they crowd each other, or go slightly off course. Thus, despite the fact that the Pikmin are the primary target of the player's control, there remains a hard-coded gap between what the player can actually do to control the game and what is supposed to be controlled within the game. The mechanism of control offered to the player is explicitly and deliberately flawed for the task the player is assigned.
This gap, in Pikmin, is also not purely on the level of game structure. It exists on a physical, bodily level. To understand this, it is necessary first to look at the basic design of the Gamecube controller. The primary control mechanism, generally used by the player's left thumb, is an analog stick. The player's right thumb, by default, rests near an arrangement of four buttons. On the Gamecube, this arrangement features an oversized A button surrounded by three smaller buttons, B, X, and Y. The A button, due to its central location, is the default button, and most Gamecube games use it for whatever the most common action is. (In the fighting game Super Smash Brothers, for example, the A button punches. In Super Mario Sunshine and The Legend of Zelda: The Wind Waker, it is the default button for interacting with objects.)
In Pikmin, the left analog stick is used to control Olimar. The Pikmin are primarily controlled by the B button, which sounds the whistle that summons them to your side, and by a secondary analog stick located on an arm extending off the bottom of the area the four buttons are on. This second analog stick, called the C-stick, is used to herd the Pikmin. In terms of the physical act of using the video game controller, then, control of the Pikmin always involves moving off of the default position on the controller. By forcing the right thumb away from its natural resting position, the game makes control of the Pikmin physically secondary to the control of Olimar — a sensation that is physically heightened by the fact that the top of the C-stick is significantly smaller than the main control stick, thus giving a subtle sense of uneasiness and imbalance when the thumb is resting on it. This highlights a key point about the avatari — it is not merely something written into the code of the game, but something that manifests in a distinct and material way during gameplay, and something that is actively shaped by the platform on which the game runs.[16] Pikmin would not have been possible on any earlier Nintendo system — not just because the processing power was not sufficient to handle that many characters, but also because the game could not exist until an analog stick that was clearly and explicitly secondary was put on the controller.
In all of these examples, playing the video game is fundamentally perverse. Using the McLuhan metaphor of adding appendages, to play a game like SimCity, Lemmings, or Pikmin is to deliberately attach a wrong and even monstrous appendage. This is not, it should be noted, an instance of failure — quite the contrary, we continue functioning, but we do so in a fundamentally perverse state. In these cases, the psychoanalytic symptom may resemble the Lacanian mirror stage more than Klein's partial objects. There is the continual confirmation that the objects on the screen are extensions of the player that respond to their controls. This moment of recognition of the self in an exterior position echoes Lacan’s image of the infant gazing into a mirror and recognizing itself reflected. But the recognition of the screen as a coherent body that extends naturally from our own, like the imago in the mirror, is a misrecognition. The infant, like the player, looks in the mirror and recognizes an image of the self — not the actual self. The intervention of the avatari between us and the objects under our control parallels the intervention of the infant's “motor impotence and nursling dependence”  [Lacan 2004, 76] between itself and the ideal-I it envisions in the mirror. Since (as we learn from Heidegger) this entire edifice only becomes manifest to us at the moment of its disruption, the identification of the game as our extension is from the first moment marked by its own negation. Thus we are, from this originary moment, forced to frantically try to reclaim the unity that we have always already lost.
It is in this regard that I break most firmly from the previous model of psychoanalysis and the avatar expressed by Rehak. Although Rehak admits to the fundamental schisms of the psyche expressed by Lacan, he also, ultimately, buys into a version of the futurist fantasy whereby avatars serve, quite improbably, as a force that “complete an arc of desire”  [Rehak 2003, 107] due to the fact that they “meet the criteria of Lacan's objet petit a ”  [Rehak 2003, 106]. It is difficult to overstate how utterly bizarre this claim is from a Lacanian perspective. As Dylan Evans succinctly puts it, the objet petit a is “the object which can never be obtained”  [Evans 1996, 125]. The impossible striving for the objet petit a is, in Lacan, one of the fundamental sources of dissonance and tension among the registers of the Symbolic, Imaginary, and Real. To declare that the objet petit a has suddenly been located in video games is fundamentally impossible.
How, then, does Rehak justify this baffling claim? It appears, for Rehak, that the link between the objet petit a and the avatar is embodied most clearly in his assertion that the central aspect of the avatar is its ability to reincarnate — to “explode into a mist of blood and bone fragments, only to reappear, unscathed, at the click of a mouse”  [Rehak 2003, 107]. Rehak fixates on this image, arguing that the avatar's endless reconstitution allows for an endless confirmation and reconfirmation of the player's ego — an endless repetition of the mirror stage where the avatar serves seamlessly as imago. But Rehak’s explanation glosses over the role of the avatar's death — the explosion of blood and bone fragments — in this cycle. In order to repeat the cycle of rebirth that Rehak identifies, there must be a corresponding cycle of destruction, in which the formulated and imaginary ego is broken apart. But because Rehak is invested in a futurist fantasy whereby the avatar serves as the mythic objet petit a , he, like Murray and Ryan, fails to note that games are not ordered around the repetition of resurrection, but around the repetition of death. Resurrection is a conceit that allows the player to continue playing. But many aspects of games — enemies, traps, puzzles, etc — are designed to produce death, not resurrection.
It is important to note, in fact, that the cycle of death is, at the final moment of the game, the one that wins out. Rehak observes that the death/resurrection cycle of the avatar extends outside of the diegetic gameworld, noting that it takes place “with the flick of a power switch or the selection of a QUIT command on the file menu”  [Rehak 2003, 10c7]. What is vital to observe here, however, is that eventually, every game is, eventually, played for the last time and turned off forever. The cycle of death/resurrection never stops on resurrection. Thus while the dynamic of play may depend on an endless juggling of the avatar and avatari, keeping each in play with the other, it is, in the end, the avatari — the loss of control — that is the most fundamental. In this sense, we have been speaking in error all along: it is not that the avatari emerges forth to disrupt the transformation of the avatar, but rather that the inevitability of the final death is disrupted by the misrecognition of the transformative fantasy. It is the avatar that emerges to disrupt the inevitability of death.
Even in games with definitive endpoints, one cannot really say that the game ends at a point where the avatari has been vanquished. The desire for a more complete unity persists: players go back to games they’ve already beaten, setting up challenges for themselves like beating the game without dying at all, or beating it extremely quickly. The desire, in playing a video game, is always more than just winning. The desire is to completely overcome the avatari function, and to restore the illusory unity promised by the avatari's initial intervention. In which case the real fantasy is that of the elusive “perfect run.” This fantasy of the perfect run is, ultimately, what theorists such as Murray and Wark succumb to when their analyses assume the possibility of any sort of perfect union with or mastery of the game, or what Rehak succumbs to when he declares the objet petit a to be found at last. It provides the entire apparatus for the rhetoric of mastery and of transformation/virtual reality that video games work with, and is a vital part of the culture of gaming.
The perfect run, of course, can functionally never be achieved. There are only a handful of games where it is even theoretically possible to approach it, and in these games only the absolute most devoted players have a chance. But many methods exist to simulate the perfect run. The most classic of these is the cheat code — some particular sequence of buttons or keystrokes that will grant the player some bonus, whether extra lives, invulnerability, the ability to pick a level, unlimited ammo, or something else. These codes function by removing, either partially or fully, the avatari. The result is a curiously sterile game — one where the avatar reigns so supreme that there's little point to the game other than the idle pleasure of seeing parts of the game that the player could not access based on skill alone.[17] What we see from this, then, is that the dream of the perfect run is not simply about the disappearance of the avatari, but about complete mastery over it.
The draw of the perfect run has only increased with the rise of the Internet, and with the popularity of the “speed run” videos. These videos show the completion of a game — usually a classic game released well over a decade ago — done in an extraordinarily short time, usually with other astonishing feats of skill within the game. The videos are, of course, faked, using various tools and cheats to make the display of perfection possible, but still, they provide a glimpse of the objet petit a that is always strived towards. Unlike in Rehak’s formulation, however, the objet petit a remains firmly ineffable in these videos. The videos may show the moment where the ineffable goal is found, but it is never found by the player watching the game. Rather, these videos show the game as played by the Other, for whom the disjunct between ego and ego-ideal never formulated in the first place. The speed run videos are not demonstrations of the avatari's vanquishing and of the reclamation of the objet petit a , but rather demonstrations of the player’s radical non-existence. These speed runs demonstrate a hermeneutic of the video game where the avatari never had been, and where the fusion of player, controller, and game took place without problem or incident, and thus in which the “player” is no longer a meaningful entity as such. They do not show the reclamation objet petit a , but rather rejection of its very separation — the sort of rejection that Lacan labels as verwerfung [Lacan 2004, 323]. This verwerfung, however, is always impossible for an actual player who has always already been cleaved from the tantalizing imago. It is, in short, an untouched and untouchable mirage.
There is, in this image, a dynamism that is absent from existing views of how video games work. The constant and necessarily unending struggle to reclaim the mythic unity of the avatar presents a system where, at any given moment, things are in progress and happening. In this regard, the image of the avatari seems, on a fundamental level, to mirror the video game itself, with its constant motion and interaction. Ultimately, too many of the current theories of video games rely on a model that treats the game as a static object with the property of interactivity — whether as a system of rules or as a cyberdrama. In reality, however, the video game is not a static object — it is a responsive object that is constantly defined and redefined mutually with the subjectivity that plays it.
By bringing this process of mutual resistance within the interface in sharp relief, the video game can finally take an appropriate and useful place in the larger context of digital media. By showing us an interface that consciously comes to a point of disruption the video game allows us to look at the process of interacting and the material instantiation of that interaction. In this view, the interface as an active site of mutual subjectivity between user and system. The video game is a potent reminder that this mutual subjectivity is not a calm or seamless process, but rather a fraught and complex process that makes explicit the dynamic of control and authority that underlies what we have come to understand as “interactive media.”

Acknowledgments

I would like to thank W.J.T. Mitchell and Terry Harpold for their assistance in formulating and fine-tuning this argument, and for the wealth of perspectives and insights that arose during my conversations with both of them.

Notes

[1] Readers seeking background on these camps are encouraged to read the first two sections of Noah Wardrip-Fruin and Pat Harrigan’s First Person collection, and the ensuing responses on the Electronic Book Review. These two sections explicitly stage the narratology/ludology debate. The destructive nature of this debate can be clearly seen in later sections, which, in trying to continue the analysis of game studies beyond those two poles, drift steadily further and further away from video games and into a broader collection of digital media.
[2] The term was first used in 1985 in the video game Ultima IV, where the player's character was simply called “The Avatar.” It was used more broadly in LucasFilm's online RPG Habitat, which ran from 1986-1988 before going through a number of purchases and renamings. (A derivative of the game can still be played under the name VZones). In Habitat, players’ online characters were called “Avatars,” a term that was always capitalized. Its usage as an improper noun originates from Neal Stephenson's seminal novel Snow Crash, in which he independently invented the term (he inadvertently takes credit for it in the acknowledgments) to describe the people who populate the virtual reality world he calls the Metaverse: “The people are pieces of software called avatars. They are the audiovisual bodies that people use to communicate with each other in the Metaverse”  [Stephenson 2000, 33].
[3] Rehak traces the history of this first-person avatar’s evolution, starting with early interactive fictions such as Adventure through to modern games like Battlezone.
[4] This parallels Wayne Booth's concept of the implied author, that is, the author constructed by the reader's inevitable assumptions while reading the text [Booth 1983, 71].
[5] This point may seem controversial. For instance, Espen Aarseth argues against this very idea of narrative, saying that “In Tetris, I do not stop to ponder what those bricks are really supposed to be”  [Aarseth 2004, 49]. This, however, is not actually contrary to my point. My claim is simply there is something that you are controlling in Tetris that is spinning those blocks and clearing them out. This is true regardless of whether that something is involved in a larger narrative.
[6] An interesting special case occurs with games that simply copy traditional games — solitaire or chess, for example. One might make a claim that these games feature a first person avatar, since, in these cases, one can readily imagine an actual person sitting at a table playing chess. This seems to me to be a poor move, however. These games are best understood as analagous to watching a movie on one's television — that is to say, cases where one medium is being emulated in another. There is a lot to do with this distinction, but it is largely outside the scope of this paper, since these imitative games do not contain many of the particular functions of the avatar I am discussing here, in particular that of the Avatari.
[7] 7. This frustration with literary studies can be found, albeit less explicitly, all the way back in his first book, Cybertext. This book begins with a complaint about the inevitable attempts to collapse traditional literature and electronic literature into one topic, in contrast to his desire “to present the perspective of ergodic literature and cybertext to a fresh audience of literary critics and theorists”  [Aarseth 1997, 2].
[8] This paratext includes, but is not limited to, the two movies starring Angelina Jolie, as well as the various “Nude Raider” hacks that allow you to play the game with a naked Lara Croft. The degree to which Croft’s sexuality is weaved deeply into the game is perhaps best captured in Prima Publishing's book on the character, Lara Croft: The Art of Virtual Seduction. The book moves without detectable irony from interviews with the developers of Tomb Raider to digital pin-ups of Lara Croft wearing Gucci bikinis and other designer clothes.
[9] A very good example, albeit one that postdates Aarseth’s argument, is BioShock. Throughout the game, the player encounters characters known as the “Little Sisters,” who appear as small children. The player is, in these encounters, given a choice between rescuing the Little Sisters or killing them. Each choice has certain benefits, with the choice to kill having the greater set of benefits. Were Aarseth correct in his assessment, the player would look past the physical appearance of the Little Sisters and make their choice based entirely on the level of challenge they want in completing the game. In practice, however, the cultural paratext surrounding defenseless children does not disappear so seamlessly, and the choice is not that straightforward. Indeed, numerous reviewers (as well as myself) found themselves unable to kill the Little Sisters because of their resemblance to children in the players’ own lives. This impulse does not come from within the formal system of the game, and Aarseth’s argument is deeply unsuited to the task of accounting for it.
[10] Murray uses this as an “umbrella term” for all of the possibilities for “reinvention of storytelling itself for the new digital medium”  [Murray 1997, 271].
[11] The problems and allure of this futurist perspective are critiqued by Paul Druguid in his essay “Material Matters: The Past and Futurology of the Book.”
[12] This is akin to Peter Rabinowitz's idea of joining the authorial audience as a prerequisite for meaningful engagement or dissent with a text [Rabinowitz & Smith 1998, 28].
[13] This judgment is necessarily retrospective. The rhetoric of transformation was a part of video games from the beginning, as demonstrated in films such as Tron and The Last Starfighter. Furthermore, this rhetoric has always been linked to the idea of the avatar: the first known print usage of the term, a 1986 article on Habitat in Run, a magazine for owners of the Commodore 64 and Commodore 128, suggests that “Once a human being enters Habitat, he or she takes on the visual form of an Avatar, and for all intents and purposes becomes one of these new-world beings”  [Morabito 1986, 24]. I argue below that this ecstatic fantasy is a necessary consequence of gameplay.
[14] Its name is in homage to the company that produced the first widely popular video game, Pong, Atari — and also to the Atari 2600, the first video game system to be widely popular. Both Pong and the 2600 featured blocky graphics — technical limitations in the 2600 prevented a single object from even being more than one color. (A workaround for this existed that allowed you to get multiple colors, but only in individual rows or columns) Their sound was primitive at best, the games had minimal depth, and, for good measure, the controls were frequently unresponsive and sluggish. Despite this myriad of flaws, however, they were wildly popular and successful, to the point where collections of Atari games are a perennial seller on modern video game consoles.
[15] This is an odd aspect of the marketing of the medium. It is largely a throwback to the era where video games were primarily played in dedicated arcades. The profits of arcades depended on people spending a lot of money on games, and so the games needed to have the avatari to manifest as frequently as possible so as to cause the player to insert another quarter to keep playing. As arcades declined and most games were played on the home console, this approach became less sensible, as the home system model depends on a single large upfront cost for buying the game, rather than an extended series of small transactions. Thus profit margins would, in theory, be best expanded by having games end quickly so that players will go and buy a new game. The industry has responded to this curious throwback in three main ways. One approach comes in the various franchises of sports video games, which push players to buy a new version every year so that they can have up to date rosters for various teams. A second approach comes in the form of ancillary merchandise like $20 strategy guides that help get a player through the points of extreme difficulty. And finally, with the rise of Internet-enabled consoles, the old arcade model of making additional purchases has returned as players are encouraged to purchase additional levels and content for the game, thus extending the financial life of their purchase.
[16] One interesting aspect of the avatari that is sadly outside the scope of this paper is the question of how particular video game platforms shape the ways in which the avatari can manifest. For example, the Nintendo Wii, with its motion-sensitive controller, has a unique notion of the avatar. The avatar in a Wii game will, to some extent, physically mirror the player's motions. In Wii Sports, the avatar swings his racket as the player swings the Wii controller. In Elebits, the avatar's capture gun physically shares space with the Wii controller. In these cases, the intervention of the avatari (which is, of course, alive and well in moments when the avatar misses the serve or shoots the wrong thing with the capture gun) emphasizes the ways in which the video game functions through a repetition and echoing of the mirror stage. Other games like Super Mario Galaxy do not depend on this sort of collocation of player and avatar, but on the other hand also do not substantively change the player’s mode of control from that of older systems, incorporating only a few vague motion controls like shaking the WiiMote to produce one effect.
[17] This sterility may explain Nintendo's aggressive reaction in their lawsuit against Lewis Galoob Toys over a product called the Game Genie. The Game Genie was a device that plugged into the Nintendo Entertainment System and allowed cheat codes that were not programmed into the game to be used. Nintendo sued, arguing that the game created illegal derivative works of the original games. This is, on the surface, a strange lawsuit, as the derivative works could not possibly pose any financial threat to Nintendo, since the original cartridge was still required. But the lawsuit makes more sense when one considers the nature of the subsequent hacked game. In essence, the Game Genie makes playing games less fun, and does so in a very tempting way by seeming to offer the dream of the perfect run.

Works Cited

Aarseth 1997 
Aarseth, Espen J. Cybertext: Perspectives on Ergodic Literature. Baltimore: The Johns Hopkins University Press, 1997.
Aarseth 2004 
Aarseth, Espen. “Genre Trouble: Narrativism and the Art of Simulation”. In Noah Wardrip-Fruin and Pat Harrigan, eds., First Person: New Media as Story, Performance, and Game. Cambridge, MA: MIT Press, 2004. pp. 45-55.
Bolter & Grusin 2003 
Bolter, Jay David, and Richard Grusin. Remediation: Understanding New Media. Cambridge: MIT Press, 2003.
Booth 1983 
Booth, Wayne C. The Rhetoric of Fiction. Chicago: University of Chicago Press, 1983.
Breukelaar et al., 2002 
Warning: Biblio formatting not applied. BreukelaarRon. Erik D.Demaine. SusanHohenberger. Tetris is Hard, Even to Approximate. Technical Report MIT-LCS-TR-865, Laboratory of Computer Science, Massachusetts Institute of Technology. 2002.
Cohen 2000 
Cohen, Mark. Lara Croft: The Art of Virtual Seduction. Roseville: Prima Games, 2000.
Duguid 1996 
Duguid, Paul. “Material Matters: The Past and Futurology of the Book”. In Geoffrey Nunberg, The Future of the Book. Los Angeles: University of California Press, 1996. pp. 63-102.
Evans 1996 
Evans, Dylan. An Introductory Dictionary of Lacanian Psychoanalysis. London: Brunner-Routledge, 1996.
Heidegger 1996 
Heidegger, Martin. Being and Time. Translated by Joan Stambaugh. Albany: State University of New York Press, 1996.
Klein 1975 
Klein, Melanie. Love, Guilt, and Reparation and Other Works. New York: The Free Press, 1975.
Kucklichs 2007 
KucklichsJ. The Gospel According to Wark or Why This is Not A Review of Gamer Theory. Gameology. May 7, 2007. http://www.gameology.org/reviews/the_gospel_according_to_wark_or_why_this#comment.
Lacan 2004 
Lacan, Jacques. Ecrits: The First Complete Translation in English. Edited by Bruce Fink. New York: W.W. Norton, 2004.
Manovich 2001 
Manovich, Lev. The Language of New Media. Cambridge: MIT Press, 2001.
McLuhan 1964 
McLuhan, M. Understanding Media: The Extensions of Man. 2nd edition. New York: New American Library, 1964.
Morabito 1986 
Morabito, Margaret. “Enter the Online World of LucasFilm”. Run (August 1986), pp. 24-28.
Murray 1997 
Murray, Janet. Hamlet on the Holodeck: The Future of Narrative in Cyberspace. New York: Free Press, 1997.
Norman 1988 
Norman, Donald A. The Psychology of Everyday Things. New York: Basic Books, 1988.
Rabinowitz & Smith 1998 
Rabinowitz, Peter J., and Michael W. Smith. Authorizing Readers: Resistance and Respect in the Teaching of Literature. New York: Teachers College Press, 1998.
Rehak 2003 
Rehak, Bob. “Playing at Being: Psychoanalysis and the Avatar”. In Mark J.P. Wolf and Bernard Perron, eds., The Video Game Theory Reader. New York: Routledge, 2003. pp. 103-127.
Ryan 2001 
Ryan, Marie-Laure. Narrative as Virtual Reality: Immersion and Interactivity in Literature and Electronic Media. Baltimore: The Johns Hopkins University Press, 2001.
Stephenson 2000 
Stephenson, Neal. Snow Crash. New York: Bantam, 2000.
Turkle 2005 
Turkle, Sherry. “Computer Games as Evocative Objects: From Projective Screens to Relational Artifacts”. In Joost Raessens and Jeffrey Goldstein, eds., Handbook of Computer Game Studies. Cambridge, MA: MIT Press, 2005.
Wark 2004 
Wark, McKenzie. A Hacker Manifesto. Cambridge, MA: Harvard University Press, 2004.
Wark 2007 
Wark, M. Gamer Theory. Cambridge: Harvard University Press, 2007.
Williams 1997 
Williams, Raymond. Marxism and Literature. Oxford: Oxford University Press, 1997.
Žižek 2004 
Zizek, Slavoj. Organs Without Bodies: On Deleuze and Consequences. New York: Routledge, 2004.