This is a pet idea I came up with while being bored to unconsciousness in my Sociology 101 class (the subject material is, at least conceptually, fascinating, merely the pace of delivery and over-simplification of highly nuanced concepts is boring).

Somewhere whilst I was doodling away explanations as to the theory of socialization as an interaction process came to mind, my neighbour was playing Pokemon with another classmate somewhere else in the lecture hall. Then, deep in my hyper-nerd subconcious an idea for a game/toy emerged.

Socialization is an interactive process and social interaction is both an exchange of “social capital” and a method of communicating culture etc etc. What if we took that to a game scenario where you interact with an agent, a proxy avatar, in an isolated or at least separated setting. Then, based on the nature of that interaction watched that agent perform in a “social” setting with other agents.

It would be a highly complex wind up toy, wherein through methods of play and interaction you train your agent how to behave. Then, using that information you’ve given it it will seek to maximize the gain of social capital whilst minimizing loss.

However, other agents will be seeking to do the same thing. So a game situation presents itself. Your agent can gain favour by being dramatic, telling jokes, being flirtatious, being intelligent, charismatic, demeaning, etc. Your agent can lose favour by capitulating, being demeaned, recipricating flirts, laughing at jokes and responding to other agent’s charisma.

There also presents an interdependence phenomena where a small “knot” of agents woud likely exchange social capital in a cyclical fashion or act as a “team” to affect the social capital of others.

I’m not sure how it would all fix out but, I think, it could prove an interesting study not only in social network construction but on the idea of a strictly capitalist view of social interaction to test how closely it follows reality.

3 Responses to “Socialization Game / Experiment”

Ok, I’m an avid blog reader on my “spare” time and I happened to have stumbled upon your blog and found this entry kind of interesting. So let me get this straight we pick an agent, isolate him, and hope he doesn’t turn feral. Then we socially condition him, like an android, and form him to be the being we want him to be. Then we release him and reward/punish him based on how well he interacts with people and how well he’s able to keep within the means of his social conditioning. It sounds a bit “Brave New World”-ish to me. But then again it’s just a game right. Now in the real world, not only would this be inhumane (it sucks how this word only applies to humans, I mean it’s inhumane to stick a human’s brain with wires and measure the effects of centrifugal force, but totally alright to do it with a monkey???) and illegal but I doubt you can isolate and condition a person without any other human interaction and expect him to be able to interact properly outside of the normal settings he was used too. All you’ll have is a scared looking bunch of humans.

While it’s true that performing such experimentation on human beings is both immoral and illegal, this is a virtualized space.

In essence, the agent (which could, for all intents and purposes be a glowy orb or ambulatory potted plant) would learn from you as you interact with it. Your responses teach it who to react, much like an infant learns basic human interaction from his or her parents.

Now, granted one could be brutal with the agent and teach it to fear, this is unlikely to produce a successful agent when placed in the “social arena.”

The goal of the game and successful strategies would hinge on producing the best min/maxer of social capital. As in minimizing loss and maximizing gain. Since social capital can be earned in a variety of fashions, the “maladjusted” characters may or may not be successful, though it could produce some interesting results (especially if all other agents adopt similar strategies you’d have a rogue-agent scenario, the very stuff of Nash…)

So in essence, the Agent is limited to the limits of your knowledge? Or is there a period where it’ll branch out to learn on it’s own, in order to be the most successful in the social capital part of the game?

Something to say?