My Second Self: Understanding the Relationship between Players and Avatars

The avatar holds a special place in video game design. As the vehicle for the player’s participation in most games, it is central both to the functioning of gameplay and the player’s perception of the game world. Despite this prominent role, the theoretical understanding of the way avatars function seems to me to be underdeveloped. Epsen Aarseth, writing in 1997, says of the avatar that “the user… will not come to see this person as an other, or as a person at all, but rather as a remote-controlled extension of herself” (Cybertext, 113). While over 15 years old, the way avatars are critically examined seems to have changed fairly little. But Aarseth, in this quote, was talking about text-based adventures, and in light of innovations in characterization in video games, this theoretical construct falls apart. Games like The Walking Dead, Heavy Rain, or The Witcher put the player in control of a fully-developed, semi-independent character (or in Heavy Rain, characters). These avatars can’t be said to be simple extensions of the player, and clearly a new theory of the functioning of avatars is needed. And while other critics have taken interesting stances on the avatar, these critics tend to either discuss the issue tangentially, briefly reinterpret its function to serve a larger argument, or use diction to create more of a distance between the player and avatar, like “puppet-master,” rather than making the role of the avatar the center of their analyses. This paper will try to show the inadequacy of the traditional understanding of the avatar by examining its shortcomings when analysing The Walking Dead, and tentatively present an alternative approach based on Jacques Lacan’s theories, as presented in his essay on the mirror stage.

While Aarseth was writing about an invisible avatar that simply follows the commands of the player, games like The Walking Dead give us well-formed and written characters, whose actions we direct. But these characters do more than simply follow our commands. Looking first at dialogue and then at actions, it becomes clear that these avatars function in a much more complicated way.

When talking to another character, it is typically the player who determines what the avatar says, how they respond. In games like Skyrim, the avatar doesn’t actually speak, but is assumed to say precisely what the player chooses. In The Walking Dead, however, things are not so straight forward. Instead of determining what Lee says, the player determines roughly the tone, or the underlying meaning, of what is expressed. The dialogue is Lee’s. At times, the difference between what the player chooses and what Lee says is pronounced. In the opening sequence, Lee is speaking to the police officer who is driving him to prison. When the officer asks a particularly personal question, the player is given the option to respond “Fuck you.” Choose this response, however, and Lee actually says nothing, merely grumbling in protest. It is almost as if Lee is unwilling to say what the player would like to. There is a disconnect here between the will of the character and the will of the player, and we, as players, can’t force Lee to do precisely as we wish. Instead, there is a sort of middle-ground or compromise formed, where Lee expresses his dissatisfaction, at our command, in his own way. The player can thus shape Lee’s speech, and thereby personality, within the confines of the character, but cannot simply determine the character as he/she wills. This is true of any game with dialogue, even games like Skyrim, where the avatar says what we choose but we are still given dialogue options to choose from. The avatar is not simply our mouthpiece: we choose the dialogue option, provided by the script, that best approximates our desired response to the situation. There can also be particularly jarring examples, where the player chooses to say something, but the way it is said by the avatar gives the response a meaning that the player doesn’t intend. In another early sequence of The Walking Dead, for example, Lee is in the middle of a tense conversation with Larry when Clementine tells him “I have to pee.” I distinctly remember this section of the game, as the two possible dialogue options are “In a minute” and “Just go,” neither of which I was particularly happy with. I wanted to be attentive to Clementine’s needs, but there didn’t appear to be a dialogue option that reflected my desire. Instead, I had to choose the choices provided by the script, by the characterization of Lee. In the end, I picked the dialogue option of “Just go.” While I had hoped that this choice was the more caring option, Lee’s actual reaction was something I hadn’t intended: he yells at Clementine. I had meant for Lee to reassure her, be kind to her, demonstrate an attentiveness to her needs. Instead, he shows a disregard for her feelings and needs by reacting violently to her request. Rather than feeling like Lee was my vehicle in the game’s story, I felt disconnected from him, as his actions didn’t match my intentions. Clearly, Lee was not me in this situation, nor was he a simple extension of myself, of my volition. Not only were my dialogue choices severely limited, but the tone and expression of these choices didn’t convey what I had hoped. Perhaps Lee is more my Lee because of the choices I made–that is, perhaps the particular choices I made in dialogue made Lee’s character more to my own liking–but only imperfectly. Instead of him being an extension of me, there is a compromise between my volition and his characterization. As such, the simple understanding of the avatar as an extension breaks down, and is incapable of assessing the complexities of such a player-avatar relationship.

I now want to turn to actions in order to further demonstrate the disconnect between player and avatar. By actions, I essentially mean acts by the player (inputs) which have impacts on-screen (outputs). Of course, the act of choosing between dialogue options is an action, as is any interaction with a digital game. The point I want to raise is that there is a fundamental disconnect between how we act to give inputs and the outputs that are displayed. Video games, as The Stanley Parable parodies, typically demand a particular action on the part of the player: pressing buttons. What occurs on-screen, however, rarely reflects this act. More often, the avatar shoots a gun, swings a sword, interacts with an object, or engages in dialogue with other characters. There is a great deal of disconnect created by the obvious differences between inputs and outputs, which serves to dissociate the player from their avatar. This issue has been discussed recently in a video by Critical Path, presented below. To take one of many possible examples, the second time Lee must take out a zombie in The Walking Dead, he viciously and repeatedly beats it in the head with a hammer. For the player, however, this violent act is simply that of pressing the correct button. Once again, it is clear that Lee is not just an extension of the player, as his/her actions are very imperfectly presented by Lee’s. The fact that Lee acts when we do creates a connection between him and the player, but it is an imperfect one, invalidating the traditional understanding of the avatar.

While there are, of course, a variety of different perspectives presented in this video, the overwhelming tendency is for these designers and critics to seek greater harmony between player inputs, on a controller or otherwise, and the outputs on-screen. But this discussion, near the end especially, branches into something else, into a discussion of how we can circumvent the controller entirely, how we can directly interact with the computer. When this happens, according to Professor Jesse Schell, the final interviewee, video games will become “the dominant art form of the 21st century.” So, the thesis then is that controllers, in disconnecting us from what is happening on-screen, often the actions of the avatar, are holding the medium back. Why, one might ask? Because it prevents the avatar from being a perfect extension of the player, which is a part of the more general quest for “immersion” in video games. As I said in my previous essay, with reference to Eric Swain’s writing, I don’t think that immersion need be the principle goal of game design. More interesting to me is an analysis of how the perceived divide between ourselves and our avatars functions, and indeed how it can be used to create diverse player experiences.

Graphic content in video

If the traditional understanding of the avatar is insufficient, as I hope I have adequately shown, then it is necessary to develop a new one, a theory of player-avatar relations that takes into account new developments in video games. What follows is a tentative attempt to understand avatars in a new way. The psychoanalytic theories of Jacques Lacan seem particularly useful to me in the case of the avatar, specifically his idea of the mirror stage. I will briefly summarize the aspects of this theory that relate most pertinently to the idea of video game avatars, as I understand them, but since Lacan’s essay is only a few pages long and can be interpreted in a number of different ways, I thoroughly encourage you to read it. Alternatively, if the prospect of reading critical theory seems unpleasant, just skip the paragraph, as I sum up the important inferences I draw from Lacan immediately afterwards.

The mirror stage, according to Lacan, is the moment when an infant first understands that the image they see in the mirror is a reflection of themselves. This, he posits, arises from the act of playing in front of the mirror, and seeing his/her own actions mimicked by the reflected image. In Lacan’s words, the child identifies with the image “in a series of gestures in which he experiences in play the relation between the movements assumed in the image and the reflected environment” (Écrits, 1). This image is taken by the child to be their self, which leads to the development of their sense of self, or the ego. So the ego becomes tied to an image, a spatial depiction external to the child’s actual being. Furthermore, this image, which Lacan calls the Ideal-I or imago (Écrits, 2), is not the real self of the child, but is instead a fiction, which the individual never truly becomes: “this form situates the agency of the ego, before its social determination, in a fictional direction… which will only rejoin the coming-into-being of the subject asymptotically” (Écrits, 2). I take this to mean that the ego becomes projected onto a self-image which is inherently a fiction, inherently oversimplified, which the actual individual never truly becomes, but can only approach. Lacan states that there is a division between the individual and their Ideal-I because the image “appears to him above all in contrasting size that fixes it and in a symmetry that inverts it, in contrast with the turbulent movements that the subject feels are animating him” (Écrits, 2). While the individual feels he/she is turbulent, changing, and divided, the Ideal-I is fixed, cohesive, whole. In being so, the individual sees in the image the potential of the “maturation of his power” (Écrits, 2) as well as permanence, which leads him/her to desire to be that image, and strive to do so. The importance of such an image is also seen in the individual’s social relations, as he/she uses this image of a fixed I “to establish a relation between the organism and its reality–or, as they say, between the Innenwelt [inner world] and the Umwelt [environment]” (Écrits, 3). It is through this understanding of ourselves as cohesive wholes that we are able to interact with our environment, to find our place within it, even though our failure to truly be the image may lead to a sense of “organic insufficiency in [our] natural reality” (Écrits, 3).

Needless to say, this relationship between the individual and the Ideal-I of the reflected image is very complex, but what is key for understanding avatars is that 1) we identify with the image because its actions reflect our own, 2) the projection of ourselves onto the image creates a fictional self, 3) this image presents a cohesiveness and power we feel we lack, 4) we can only ever approach this image, be more like the it, and cannot ever fully become it, and 5) it is through this image that we are able to relate to the world, i.e. it forms a bridge between our internal selves and our environment.

Seen through this light, the avatar is a much more nuanced and diverse entity. Lacan’s theories help indicate what draws us to identify with the avatar. Our initial identification comes from mimicry. While the avatar’s mimicry is much less perfect than that of a reflection, we still see the outputs of our inputs reflected on-screen. When I moved the left stick to the left, Lee moved that way. The identification is strengthened as our inputs are seen to have created an impact on the virtual environment through the avatar: I killed the zombie because my input caused Lee to hit it with a hammer, or I consoled Clementine because my input caused Lee to say something nice. The avatar thus gives us a means to cope with and influence the environment, without ever being us. It is enough for us to identify with it in order for it to fulfill its intermediary function, and we identify with it because of this intermediary function, as well as its mimicry of our actions.

But, as I have taken pains to demonstrate, avatars are never completely passive intermediaries between the player and the game world. Instead, just as Lacan’s Ideal-I is said to be a “spatial identification” which is thus exterior to the subject (Écrits, 3), the avatars we play as are fundamentally distanced from us. In the same way the Ideal-I is an other, so too is the avatar. As long as the mimicry of our actions by the avatar is imperfect, as long as we see the avatar before us, even if it is just their limbs, and/or as long as we need a social intermediary to interact with the game world in the form of the avatar, there is always some degree of alienation between the player and the avatar. Our relationship with the avatar therefore exists somewhere between the extremes of pure identification and pure alienation. When the two, player and avatar, approach each other, they do so only asymptotically.

Game design choices can be used to increase the force of identification or, conversely, alienation. The number of tools available for designers is vast. The use of camera angle or perspective, dialogue structure, character creation, and control schemes are just a few examples. I would like to interrogate how a few games implement these design elements to show how designers influence the player’s level of identification or alienation.

Looking at a game like Call of Duty: Black Ops, the level of identification is intended to be very high. While the game does have a single player campaign mode, in which the player takes control of an avatar with a particular personality, the true focus of the game is multiplayer. Here, the avatars controlled by the players are much more generic. They are devoid of personality or narrative, and have pre-set appearances with minimal customization which de-emphasize their particularity, allowing the player to project their identity onto the relatively passive image. The goal, as made apparent in the advertisement campaign, is to make the player feel like they are on the battlefield. So, naturally, identification is key. Combined with the generic character design is the first-person perspective, intended to increase the sense of what is often called “immersion,” but which in this case really seems to mean identification with the avatar. In addition is the input setup, where the “triggers” on the game controller are intended to mirror the physical act of firing a gun, though somewhat imperfectly. This is a basic design feature of first-person shooters made for consoles, but shouldn’t be overlooked, as the greater the difference between input and output, the greater the perceived division between player and avatar. So the design of Call of Duty: Black Ops has aimed to increase identification with the avatar by de-emphasizing the avatar’s particularity and increasing, as much as possible, the mimicry of the player by the avatar. In terms of characterization or personality, the avatar is stripped down almost to its bare minimum, functioning primarily as a means to interact with the game world, and in this case, with other players. Even dialogue supports this, as players are able to communicate to one another directly through microphones and chat. The goal of Call of Duty: Black Ops is clear in both its advertising and its design: you are the avatar, therefore you are fighting. And while the identification that results is necessarily imperfect, several design decisions were made in an attempt to increase it.

The Witcher series of games have very different objectives. Rather than simply playing as a generic avatar, the player takes control of Geralt of Rivia, a witcher (genetically-enhanced monster hunter) with a distinct character taken from a series of novels. Because of the emphasis on Geralt’s personality, evident in the trailer for the newest game in the series, The Witcher 3: Wild Hunt, there will inevitably be more alienation between the player and the avatar. Geralt has a personality of his own, and while the player can shape it, they can’t completely determine it. Design choices reflect the division between player and avatar. Rather than having a first-person perspective, the games in the series use an “over the shoulder” perspective, an altered third person view (though there are different options in the first game). The player is close to Geralt, but the witcher is always visible, and the player is not allowed to see through his eyes. Another alienating element is narrative. As said above, Geralt is a distinct character, complete with his own backstory and personality. The dialogue design reflects this, and takes a form similar to that of The Walking Dead: the player’s choice determines the general meaning, but what Geralt actually says is beyond the player’s control. Character customization also has a much more limited form than in many RPGs. While certain skills can be emphasized, thus altering Geralt’s strengths and weaknesses, the choices are fairly limited: in The Witcher 2, despite the number of melee weapons available for use in gameplay, only skill in swordplay can be upgraded. Geralt the character fights with swords, and Geralt the avatar can only be upgraded in this way, encouraging the player to use swords. Finally, there is the combat system, where inputs on the mouse or controller are far removed from what happens on-screen. In both the first and second games, when an enemy is stunned, Geralt can kill them in one hit. But rather than simply striking the enemy, Geralt goes through a series of more graphic, cinematic moves, as in the video below. While the player presses only one button, the avatar does a whole combination of actions, increasing the distance between the two. Now this reduced level of identification with the avatar is not a bad thing, but is rather a design choice. Geralt is meant to be a distinct, memorable character, and so it is necessary to have a greater amount of alienation between player and avatar. It makes sense, in light of this, for the game’s design to make the player feel more removed from Geralt than from the generic avatar of Black Ops multiplayer, or the nameless “Dragonborn” from Skyrim. Different levels of identification and alienation have their benefits, and the ability to use design elements to fine-tune the relationship between player and avatar is a powerful tool for game designers.

Graphic content in video

As something of a sidenote, I would like to look at the particularly interesting function of character customization in forming a relationship between the player and the avatar. In a recent Polygon article on sexuality in Dragon Age, a fan wrote to the lead writer to say the following:

“I am a lesbian, and Dragon Age is still a beacon to me, when it first came out, so did I, and a lot of people alienated me, and I didn’t have many friends, and there was only my Xbox. Origins was magical, I could have a character that looked like me! And romance a woman! It was life line I desperately needed, for my sanity. Thank You so much!”

This is an interesting point, and one that should be addressed. Polygon recently published a couple of articles dealing with similar subjects. By being able to customize her avatar, the player felt an increased level of identification with it. What is peculiar about character creation in Dragon Age and many other games is that it leads to alienation in one sense and identification in another. We can build our representation on the screen, and so have less difficulty reconciling our being with its. In presenting a fixed image for the avatar the image of the Ideal-I is that much stronger, but the alienating effects of this are mitigated somewhat by the extent to which the player is able to make the avatar that much more their own through customization. Character creation is then one of the few design options that can work in both directions, towards alienation and identification. The level of identification caused by this act can be quite powerful, as evinced by the above quote, and especially in light of the fact that many other design choices in Dragon Age, like dialogue and perspective, are quite similar to those used in the Witcher series. This article also brings up how narrative options, like being able to determine the avatar’s sexuality, can increase the level of identification the player feels for the avatar. Customization, the ability to shape an avatar into whatever is desired, is clearly a means of increasing identification, despite the fact that it usually entails creating a fixed image, an Ideal-I, of the avatar.

The argument of this paper is not intended to be the definitive argument on the subject. There may be other ways of looking at the avatar which are similarly fruitful. Furthermore, this paper doesn’t address numerous other avenues of inquiry raised by the comparison of Lacan’s Ideal-I and the avatar, like issues of power fantasy and escapism, nor does it address games without avatars, like Tetris, the Total War series, and, perhaps most interestingly of all, République. But I hope that this paper has at least given some insight into how nuanced and complicated the relationship between the player and the avatar is, and how design decisions can be used to influence this relationship. What is perhaps most important to me is that we move beyond viewing the avatar, implicitly or explicitly, as a simple extension of the player. Not only is this false, it also leads to some of the more outlandish criticisms of video games, especially for promoting violent behaviour. If the avatar is simply an extension of the player, violence by the avatar reflects a violent attitude held, even if hidden, by the player. This, I think most gamers would agree, is nonsense. But the lack of understanding of the role of the avatar leads many people, even as intelligent a person as Slavoj Žižek, to equate violence by the avatar with the player’s supposed internal desire for violence. Here is a quote taken from Žižek’s movie, The Pervert’s Guide to Cinema, with a video of the relevant excerpt from the film provided below:

“For example, people who play video games, they adopt a screen persona of a sadist, a rapist, whatever. The idea is, in reality I am a weak person, so in order to supplement my real life weakness I adopt their false image of a strong, sexually promiscuous person, and so on and so on. But, so, this would be the naïve reading. I want to appear stronger, more active because in real life I am a weak person. But what if we read it in the opposite way: that this strong, brutal rapist, whatever, identity, is my true self in the sense that this is the psychic truth of myself, and that in real life because of social constraints and so on I am not able to enact it, so that precisely because I think it’s only a game, it’s only a persona, a safe image I adopt in virtual space, I can be there much more truthful. I can enact there an identity which is much closer to my true self.”

Žižek here has completely discounted the role of the avatar, taking it to be a mere stand-in for the player, as per the traditional paradigm presented by Aarseth. It leads him to make some pretty bold claims as to the reasons people play video games. While I don’t want to dispute the specific conclusions he draws, his argument is made fundamentally suspect by the lack of nuance used when dealing with (or failing to deal with) the role of the avatar. Player’s don’t simply “adopt a screen persona,” but rather enter into a very complicated relationship of identification and alienation with the avatar. It is necessary to develop a new paradigm, as I have tried to do here, to avoid such simplifications, which have potentially significant impacts on the medium and its cultural reception. It is my hope that more discussions on how players relate to avatars will follow, and that a more developed theoretical understanding of this relationship will emerge. In a time when video games are implicitly or explicitly being blamed for acts of horrific violence, such a discussion is a pertinent need.