Saturday, October 8, 2011

Syncretic Post, Perhaps My Best Ever Contribution to the Social Epistemological Project

I'm going to sketch this out in an effort to get it right. It'll probably be the lense through which I analyze other blogger for awhile.

1. Agent Zero: An agent popularly perceived as having the capacity to make an action-decision that affects the existential odds of the rest of us (the six billion people in our apparent reality).
2. There is an optimal action-decision Agent Zero can make. There is also a worst possible action-decision Agent Zero can make. Along this axis we popularly grade Agent Zero as hero or villian.
3. What follows are a set of Influence Agents. They are not Agent Zero, but they are either popularly perceived as having the ability to influence Agent Zero, or I define them as having the potential ability to influence Agent Zero, even if not popularly perceived, as my way of dealing with strategic framing of agents as not having the ability to (and hence possible villianous framing due to failure to) influence Agent Zero to make the optimal action-decision.
4. Bostrom-Goffman Influence Agent. This is an influence agent who sucessfully influences Agent Zero to make the optimal action-decision. (Bostrom implies motivation for the optimal decision to be made, Goffman implies following the optimal theatrical strategy for the optimal decision to be made).
5. Semmelweiss Influence Agent. A Semmelweis Agent publicly tells Agent Zero to make the (Bostrom) Optimal Action-Decision. But the Semmelweiss agent doesn't follow a Goffman Optimal consequentialist theactrical strategy, so Agent Zero does NOT make the optimal Action-Decision.
6. Heckler Influence Agent. A heckler influence agent publicaly tells Agent Zero not to make the (Bostrom) Optimal Action-Decision, and does NOT do so as part of a Goffman Optimal (double game) consequentialist theatrical strategy, so Agent Zero does NOT make the optimal Action-Decision.
7. Subsidiary Influence Agents. Primary influence agents that directly influence Agent Zero are themselves influence by subsidiary influence agents. Subsidiary influence agents can also be Bostrom-Goffman, Heckler, or Semmelweiss Agents.
8. We are all at least a subsidiary influence agent for every action-decision made in the closed set of all 6 billion people in our apparent reality.
9. Inaction by an agent, be they agent zero, a direct influence agent, or a subsidiary influence agent, is itself a default action-decision.

Moral accounting may be a bit silly, and not grounded in best of breed current scientific understanding of "free will", etc. But I think this is a better more comprehensive framework than the microsocial games played by the best living macrosocial scientists and policy optimizers that I'm aware of.

I'd appreciate being directed toward more rigorous thought on these areas than I've done here. I'm not trying to reinvent a wheel -I'd much rather find one already much better than I've got here.

EDIT: I'm thinking of calling this "The Neurotic Manifesto"

2 comments:

  1. Help me understand what you're talking about here.

    1) "There is an optimal action-decision Agent Zero can make."

    I assume you mean this in a "rolling", constantly updated sense. It's not that Agent Zero has this one super-decision to ever make in its life, on which everything else depends, but rather that Agent Zero is constantly making decision and taking actions, and at any time there is a best possible decision, worst possible decision, and spectrum in between.

    2) "Along this axis we popularly grade Agent Zero as hero or villian."

    But people don't share values, so the goodness or badness of Agent Zero's decisions depends on the standard of judgment. I assume, therefore, that this conceptual framework is meant to be applicable, independently of any particular value system (even though you have "Bostrom" in there to signify optimality of criteria).

    3) I don't understand why there are only three types of primary influence agent listed, when it seems like there ought to be four, but maybe I'm missing the point of Goffman optimality.

    Basically, it seems like there are four types of agent, to be distinguished by whether or not they are Bostrom-optimal and whether or not they are Goffman-optimal. Bostrom-Goffman is optimal in both senses, Semmelweiss is Bostrom-optimal but not Goffman-optimal, Heckler is neither Bostrom-optimal nor Goffman-optimal.

    Since lack of Goffman-optimality apparently translates into ineffectiveness, I would have thought that the Heckler category is one that can be ignored. The Heckler has the "wrong" values (not Bostrom-optimal), but as defined, isn't effective in convincing others to adopt those values.

    The missing category of influence agent is one that is not Bostrom-optimal but which *is* Goffman-optimal. An "Effective Heckler"?

    ReplyDelete
  2. Fuck, another disappeared attempt at a comment.

    Now, more briefly.

    1. The default category is heckler.
    2. All values other than the persistence of the class of living agents (6 billion living people in apparent reality) are heckles.

    Hopefully Anonymous

    ReplyDelete