notes
portraiture in the online world
7 october 2003
In the world of online forums, portraiture is still alive, but of a very different nature. In a common situation, the "artist" is replaced by an online medium that renders particular fixed attributes of the user. For example, a user may boost the number of posts he has made in order improve that statistic in his profile. Or, a user on friendster.com might try to add as many strangers as "friends" as is possible, in order to boost his statistical popularity. This predictability of how the artificial "artist" will depict a user changes the dynamics of control in the subject-artist relationship, often empowering the subject to much more meticulously controlling how he/she is portrayed. It is often the goal of users in online forums to portray themselves as having the greatest presence, the greatest reputation, and the greatest popularity. When the artistic medium is too easy to manipulate, the portrait looses its information value.
Manipulability by the subject ruins an online portraiture, because credibility is key. The audience in an online forum is particularly interested in being able to accurately assess the presence, reputation, and popularity of others, to help defuse the risk of online dealings and interactions. The audience must assume, by caution, that there are some users out there who will consciously misrepresent themselves, and portraiture that is easily manipulated is not trustworthy. The online audience is much more intent on using portraiture to achieve its information-gathering goals than is a traditional offline audience. The more foolproof a portraiture, the more accurate and informative it is.
So how do we design online portraiture to be complete, informative, and credible? In Irving Goffman's words, a person gives an expression, and gives off an impression. A user's self-representation should be considered because a user will want to speak his/her peace. Also, a user's self-representation is often informative to the audience because it informs them about a user's potential goals. However, we should also design for components of portraiture not easily manipulated by the subject -- the impression given off. For example, a user's browsing history, a purchasing history, messaging statistics, a user's membership and participation in online forums. Not that these histories can't be manipulated, because even Goffman concedes that impressions given off can be manipulated to some extent, but histories in general are not economic to fake because they are so elaborate and take so much effort to fabricate. Thus, these online histories correspond to something between a conventional signal and assessment signal. A third important aspect of portraiture is the audience. In the online portraiture, there is a particular opportunity to explicitly incorporate the feedback of the audience into the portrait. For example, Friendster allows friends to post testimonials (albeit, filtered by the subject) for the user, and these become a part of the online portrait. Also, ebay users are subject to uneditable feedback about them by users who have interacted with them in the past. The audience component of a portraiture could be particularly credible and informative because it is more or less foolproof.
What might such an online portraiture look like? AIM Profiles is a good example of user self-representation in portraiture. Lifelines is a good example of the role of histories in portraiture. And ebay's feedback is a good example of audience input in portraiture. But can we combine self-representation, histories, and audience-input into a single portrait? If we could do so, we should illustrate the contrasts between self-representation, history-representation, and audience-representation, because it is reputatibility and credibility that are central information goals of the audience. Something like a Friendster profile page achieves all three goals. The majority of the profile consists of a user's self-representation. If a user portrays himself as popular, the audience can verify the information to some extent by examining the user's friend-count or testimonial-count. The content of the testimonials can be thought of as audience-input, and either confirm or call into question a user's self-portrayal. It would be neat to bubble up more statistics about a user, such as the people he has messaged, or the number of testimonials he has written, or the number of friend requests he has initiated versus accepted. However, privacy issues often provide hard constraints on the contents of the online portrait.