From the time I woke to the time I sat down at my office desk this morning, I counted 24 different social interactions, both face-to-face and mediated by technology. Each one of these interactions required me to make an assumption or rely on known etiquette, in the absence of knowing the true state of affairs. I had many questions, and their answers were not directly perceivable: Is she friendly? Is he going to be trustworthy? Does he remember how I like my coffee, or do I have to tell him again? Will she get the assignment in on time?
In our everyday social interactions, there is a natural, unspoken currency we use to exchange information that often goes unstated but dictates much of how we interpret behavior: our face. Erving Goffman, in his seminal essay “On Face-Work,” defined face as the social value a person claims through patterns of verbal and non-verbal acts: an image of one’s self. “Face-work,” as he describes it, is the actions taken by a person to make whatever he is doing consistent with his “face.” A person, then, establishes a line by which he can communicate during any particular contact, whether face-to-face or mediated . Today businesses are distributed differently, global collaborations among people are more prevalent, and social interactions have different meanings. But that doesn’t mean face matters less. In fact, its definition may have gotten more interesting.
In 1955 Goffman may not have considered a digital translation of his definition, but today that translation is quite meaningful, albeit complicated. As I sit here typing, I count at least six open applications where I must interpret a version of face to which Goffman was referring, most noticeably represented by an avatar. And an avatar, as it happens, is fraught with both social value and perceived meaning by the person using it and the person interpreting it.
The Avatar Descent
“Avatar,” which is derived from the Sanskrit word for “descent,” is a user’s self-representation, whether in the form of a three-dimensional model used in computer games, a two-dimensional icon used in online communities, or a text construct found on early systems such as MUDs. It is an object representing the embodiment of the user that can also refer to the personality connected with the screen name, or handle, of an Internet user .
Neal Stephenson, author of cyberpunk novel Snow Crash, most likely popularized the term in 1992, using avatar to mean “online virtual bodies.” But it wasn’t until 1996, when America Online introduced buddy icons with its free version of instant messaging, that avatars became more wide-spread. Yahoo! was the first to adopt the official “avatar” name for its own instant messaging service . Now with Automattic’s Gravatar (globally recognized avatar) service, which allows people to have one avatar across multiple sites, millions of avatar images have been served billions of times per day since its official launch in 2007 (before that, Gravatars were unmaintained).
There’s etiquette at play both in face-to-face communications and via digital presentation of self with respect to face work. Diplomacy and savoir faire encourage an individual to marry the immediate situation with his or her expressions, drawing on a repertoire of postures. In digital environments, the direct relationship between face and message is often disconnected. While there is direct control over one’s facean obvious advantagethat control takes manual labor. Social toolsinstant messenger, Facebook, Twitter, to name a fewhave avatars that are, by default, static. The access to change is just as possible as it is in face-to-face communications, but there is less fluent movement from thought to action, as the user must go through the non-delicate process of uploading an image. If changing one’s face required a step-by-step process each timeno matter what the mediamy guess is we’d all be a lot less diplomatic. Marshall McLuhan notwithstanding.
“‘Avatar’ asks us to see that everything is connected, all human beings to each other, and us to the Earth. And if you have to go four-and-a-half lightyears to another, madeup planet to appreciate this miracle of the world that we have right here, well, you know what, that’s the wonder of cinema right there, that’s the magic.” James Cameron, Director, ‘Avatar’ 
Recent studies show that avatars are more than just a pretty face. Two separate studies concluded that one’s avatar not only affects how a message is received, but also how individuals interpret it. The higher the attractiveness of the avatar, the more intimate people are willing to be with strangers. People with taller avatars negotiate more aggressively in face-to-face interactions than participants with shorter avatars . Further, people more frequently chose human avatars that match their own gender. Avatars that participants perceived to be more attractive were found to be more credible and homophilous, and participants were more likely to choose them as their own .
Counterintuitive: It isn’t necessary to be realistic in order to be influential. The more realistic an avatar looked, the less likely people were to give out personal information upon a first meeting . Studies in commerce show that the avatar itself, not the information provided by the avatar, was the key to persuasion in sales. This leaves us wondering: What is the role of content when paired with an avatar ?
The avatar does not stand alone.
Add Content, Mix Signals
When someone gives you mixed signals, it’s difficult to know what to read. He is agreeing to the project, but frowning. She is shaking her head, but saying yes. What happens when an avatar, the visual representation of a person’s expression, is paired with content contrary to his statement? Consider a passive statement communicated by an aesthetically aggressive avatar. What about an angry message communicated by a sheepish avatar? Or a snide, biting comment delivered by a sweet, charming avatar? An analysis by OKCupid revealed that dating-profile pictures that showed people doing something (e.g., with animals, travel photos) generated substantially more interesting conversations. Does the visual instance change the impact of the message on its receiver ?
In a 1995 study Clifford Nass and Byron Reeves bestowed computers with “personalities” and tested similar questions in an experiment between computers and humans. They started with dominant and submissive personalities and, assigning these attributes to the computers, paired computers and users together to see if these personalities evoked the same results they would in human-human interactions. (Personalities were assigned through a combination of dialog-box characteristics and content.) The expectation was that dominant users would prefer dominant computers, and submissive users, submissive computers. Their hypothesis was, in fact, correct .
Those of us who have been working with social tools have taken for granted that we understand the juxtaposition of avatar and content. We get it, we think. There’s an assumption that because we’re fluent with digital toolsbecause we’re comfortable with flowing in and out of the complex information landscape of networked mediathat a small detail such as how to read the image of a face or a bodythe most familiar, it might seem, of all human attributesshould not be difficult to parse. Yet we’ve seen messages misinterpreted, information missed, cues questioned, and we’re left puzzled. We’re not seeing the whole picture, so to speak.
As choosers of avatars, we’re one step removed from a face-to-face interaction, not considering the connection between face and message we would otherwise. Is the chance for mixed signals higher? What motivates users then to change or not change their avatars, and what effect does it have on how their content gets interpreted?
Four Core Considerations
I’ll highlight four areas where I think we have work left to do.
Choice. In a social setting, there are clear responsibilities: One has allegiances to a social circle, family, work, perhaps social causes, and all of these contribute to an individual’s behavior in a social encounter. What responsibilities, however, does a person have on these same levels for maintaining an avatar? While IBM has developed an avatar style sheet that outlines behavioral and aesthetic guidelines for use (e.g., “be sensitive about appearances when meeting with clients”), there is no Emily Post-like guidance, quite thankfully, on avatar use for the rest of us . So are there social standards for their use? People tend to change avatars in four ways:
- Events: seasons, holidays (e.g., Christmas, New Year)
- Affiliation: sports team, company
- Social causes: awareness (e.g., breast cancer awareness), national causes, elections, bluebeanie day
- Status: points, color (e.g., Color Wars), demonstration of beauty or wealth
With all, an individual must make a choice as to whether or not to change the avatar. Issues of time, duration, announcement, and scale arise. How soon should it be changed? How long will it remain changed? Will it remain changed across media channels, signaling an identity shift rather than a temporary allegiance?
Ritual. There are those for whom changing an avatar is a nonissue; they observe a single identity across all social mediaflickr, Facebook, personal blogs. There are others still whose avatar is such a mismatched experience that there is no parsing meaning. Yet it is the third group for whom ritual is crucial: the frequent changer. This indecisive group changes their avatar on a whim, and like the eccentric neighbor with eyeglasses to match every outfit, the changer seems to have a different image with every ping, tweet, and status update.
This rate of change leaves the audience in the untenable position of interpretingdoes this signal something new? These obtuse strategies could be as innocent as a play with the tools or as meaningful as an identity crisis. Yet our prior face-work knowledge gives us no other way but to interpret the changed face with rudimentary means.
Homophily. We know that birds of a feather flock together; homophily is the tendency for individuals to bond with those like themselves. It can also correspond to the “perceived degree of psychological similarity between the images and the human psyche, or the extent to which one is perceived to be similar to the perceiver” . In research, people were drawn to avatars that looked most like themselves, and people’s confidence in the content increased when an avatar was perceived to appear more confident. From the similarities, we build up rule sets and interpretations and interpret the content of the corresponding messages accordingly.
Reliability. Even though someone may be wearing a wedding ring, it doesn’t mean she’s married. Even though someone’s status may say “Away,” it doesn’t mean he’s not at his desk. The combination of avatar and content signals reliable or unreliable cues that receivers parse the only way they know how. But the communicators, in the meantime, aren’t considering the combined meaning.
When a person is able to represent herself within media, at least within a virtual world, it fundamentally changes the psychology of interactive technology. Research shows that even though all the action is within a virtual world, people’s hearts beat faster, and the areas of the brain that regulate social interactions are more engagedpeople care how their avatars are treated .
When it comes to interpreting, we’ve navigated meaning fairly well. We can look to fields such as signaling theory, rooted in anthropology, to understand how to interpret some of the juxtapositions between avatar and content. The work of Judith Donath explores why certain signals are reliable and others are not . What happens when signals are not reliable? Where is the threshold for unreliability, and how much can we tolerate before the signals in the digital network become meaningless?
We’re starting to see more and more experiences that weave avatar with message, pairing the expression of intent with content. Get Satisfaction (http://getsatisfaction.com) allows its users to respond to comments by choosing one of four simple faces that represent a choice of 16 emotions, ranging from confident to anxious. Users write a personal message and choose an avatar to visually reinforce that message.
How will the mix of image and message further proliferate through everyday life? Will the image stand for the message or will face work still be work? What will be socially acceptable, and will new etiquettes emerge in segments that cross personal, professional, and mixed boundaries? These are all questions whose answers are emerging as I type alongside the digital expressions of strangers and colleagues. But I do feel confident that the Goffman of today might be in agreement with Errol Morris, who recently said, “I’d rather be an avatar than see ‘Avatar’” .
2. Porter, J. “On the Origin of Avatars.” December 8, 2009. http://bokardo.com/archives/on-the-origins-of-avatar/
3. Avatar (computing) http://en.wikipedia.org/wiki/Avatar_(computing)/
4. Grohol Psyd, J. “The Proteus Effect: How Our Avatar Changes Online Behavior.” http://psychcentral.com/blog/archives/2009/11/24/the-proteus-effect-how-our-avatar-changes-online-behavior/
5. Nowak, K.L., and Rauh, C. “The Influence of the Avatar on Online Perceptions of Anthropomorphism, Androgyny, Credibility, Homophily, and Attraction.” http://jcmc.indiana.edu/vol11/issue1/nowak.html/
7. Holzwarth, M., Janiszewski, C., and Neumann, M.M. “The Influence of Avatars on Online Consumer Shopping Behavior.” http://www.atypon-link.com/AMA/doi/pdf/10.1509/jmkg.70.4.19/
8. “The Four Big Myths of Profile Pictures.” Jan. 20, 2010. http://blog.okcupid.com/index.php/2010/01/20/the-4-big-myths-of-profile-pictures/
10. Reeves, B.J. Read, L. “Avatars in the Workplace.” http://blogs.hbr.org/cs/2010/01/avatars_at_work.html/
11. Donath, J. Berkman Faculty Fellow. http://smg.media.mit.edu/people/judith/
12. Morris, E. http://twitter.com/errolmorris/status/7523256942/
13. “‘Avatar’ wins Golden Globes.” http://rogerebert.suntimes.com/apps/pbcs.dll/article?AID=/20100117/COMMENTARY/100119976/
Liz Danzico is equal parts designer, educator, and editor. She is chair and cofounder of the MFA in Interaction Design Program at the School of Visual Arts and an independent consultant in New York, on the editorial board for Rosenfeld Media, and on the board of Design Ignites Change. In the past, Danzico directed experience strategy for AIGA and the information architecture teams at Barnes & Noble.com and Razorfish New York. She lectures widely and writes at Bobulate.com.
©2010 ACM 1072-5220/10/0300 $10.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2010 ACM, Inc.