Warning: this is a rather long post in which I started out exploring more sophisticated interaction possibilities in virtual environments but ended up with philosophical questions about the intrumental possibilities of virtual interactions.
Recently I used SL voice to clarify a question during a meeting in Second Life. Technically it worked fine, but it was a strange feeling anyhow. The question was somewhat delicate and could easily be misunderstood, and while formulating the question I felt like a blind person, not being able to look at the body language of the other person so I could not find out whether he was “with me” in his understanding of the question.
Second Life avatars, for all their originality, are boring to look at during debates or panel discussions. They move a bit, but this is very clearly the result of some repetitive script, the facial expressions remain the same, the lips do not move. The body language we rely so much on, is absolutely lacking.
But things will get better, several projects address this problem. We posted previously on Mitch Kapor who is developing together with Philippe Bossut a 3-D webcam allowing you to control your avatar with simple body movements. This also allows for facial expression, so it is promised.
In another development, in the SL viewer version 1.20 RC8 you find under Advanced, Character, a lipsync option (beta), so I learned on the Dutch language blog ICT & Onderwijs Blog. An example of this lipsync (in Japanese this time):
But I have to admit, not yet very impressive as far as body language and facial expression is concerned.
The blog UgoTrade however is on top of the new developments in Avatar Technology, for instance informing us about the realXtend open source project. The team behind realXtend releases interesting videos about stuff like “inverse kinematics, character morph controls, and clothing physics”. Here you see one such video:
Even more interesting is a video demonstrating a combination of FaceGen with realXtend.
However, I found a truly astonishing video on Milton Broome’s Virtual Psychology blog. Astonishing because of the technological wizardy it demonstrates, but also because of the insights delivered by Jeremy Bailenson of Stanford University’s Virtual Human Interaction Lab.
Bailenson really races through a number of fascinating experiments and issues, you can read this stuff also on the site of the Virtual Human Interaction Lab where the papers are published. Now, watching the video I had a question about the ethics of virtual interaction. Allow me to explain:
- In the video it is demonstrated there is an important difference between video conferencing and a meeting in a virtual environment. At a video conference you send some images of yourself over the network. For the virtual meeting setting as demonstrated in the video however you send information about your behavior to the other party who is than drawing that information on their own machine. Important to note: you select the information you want to send.
- So I can perfectly filter the information about my behavior into “strategic good behavior” and you will render on your machine not necessarily my behavior but the behavior I want you to think I am displaying (transformed behavior) achieving an “optimal self“.
- The technology allows me to use persuasion tactics. Avatars receiving information from your avatar can return this information (for instance head movements) making you believe it is actually their behavior instead of your own returned behavior. The capitalizes on the fact that mimicry is a very effective way to influence others.
This makes one wonder of course about the nature of social interaction in virtual evironments. The arguments in favor of virtual meetings can now be used also against it. If we have sophisticated enough virtual meetings, we can use highly expressive avatars, giving sophisticated feedback during interactions. But if it is true that the other person can easily instrumentalize this feedback by carefully selecting her optimal behavior, using clever algorithms mimicking my behavior for instance, I could very well end up feeling uncomfortable.
Of course, similar situations exist in the physical world. If I have to negociate with a person, known for her great expertise in negociations tactics, I will wonder constantly whether her behahvior is purely instrumental or not.
Let me contrast this with a notion of the German philosopher Jürgen Habermas, the ideal speech situation.These are principles governing such a situation:
- Mutual Understanding. What the speaker says is understandable and comprehensible by others.
- Truthful. There is no deliberate misrepresentation.
- Sincere Expression. This is the point we are discussing here. “When opinions, attitudes, views, and interpretations are being provided, the speaker generally attempts to be sincere and not deliberately mislead.”(cfr. the provided link). Virtual environments, if applied in a sophisticated way, make it perfectly possible to display misleading attitudes and to give a false expression of one’s true beliefs and attitudes.
- The other conditions are “Right to Speak” and “Legitimacy” (cfr the link).
All the conditions mentioned here can be problematic in the physical worls as well as in virtual environments. I only have the feeling that when actually seeing you, even using videoconference, I have more possibilities of detecting small or obvious body language signs which tell me crucial things about our interaction.
The “Optimal Self” possibility seems to imply an impossibility for the discussion partners to detect those small but crucial signs, because they will be filtered away. The sheer possiblity of this extreme instrumental behavior threatens even to undermine the virtual speech act, because I can never be sure whether I am dealing with you or with an Optimal You.
Roland Legrand
