 |
On Trust, Part IIIMarch 13th, 2006 |
Read Part I
Read Part II
Read A side note
I ended part II by concluding that the reason that the differing models of trust mattered is because at some point, solutions from one model would be applied to the problem domain best suited for the other model.
In other words, bringing in an external authority into a communitarian model (as in the recent examples with guild name trademarks, or even GLBT guilds); or alternatively, trying to use a communitarian model to deal with authority based problems. Many of the latter would likely involve intellectual property concerns, or other intrusions of real life policy into virtual space.
From a commercial perspective, the goal is to get customers to the indifference threshold — which sounds worse than I intend. This isn’t a state of active trust, it’s one of benign neglect. It’s equivalent to a chunking of trust-like behaviors: you want users to behave as if they trusted you, perhaps on the basis of trusting one of your staff, even though it is not true trust (because there’s no assumption of repeated interaction). At this point, we have effectively crossed over into the realm of politics.
In short, the goal is to have an authoritarian model where the players treat you as a member of the community even though you are actually in a position of authority. From the outset, this may seem a doomed enterprise, given that even attempting this pretense could be seen as a betrayal. But practical politics tell us that in fact we live quite happily with this sort of quasi-tribal membership in a wide array of groups, including political parties.
LambdaMOO as an example
The thing about quasi-tribal membership in groups is that they rely on twin loyalties: to a platform which we might term a vision of shared values, and to a person, whom we might call the exemplar of said platform. The tensions between these two loyalties is always evident. Often the current exemplar individual does not live up to all forms of the vision, or exhibits venality in some fashion; often the exemplar is a poor fit, or has a different agenda, but is a prisoner of the platform; often the platform itself may be uninspiring to a given voter/customer/audience member, but a given individual might, through personal charisma, convey a very different sense of what the tribe or organization’s agenda is. (Not to inject politics into things too much, but you might consider both Kennedy and Clinton as examples of the first case, and President Bush as an example of the last case).
It’s a fairly rare thing indeed to have both an agenda and a leader capable of living up to the vision for it. Quite often, in fact, one agenda is sold to the members of a group whilst another is the real operating agenda; the two agendas may not be at cross-purposes, but they are certainly somewhat orthogonal. The classic example of this is any service business, where the agendas of satisfying some need for customers overlaps with, but is not the same as, the agenda of making money.
The case of LambdaMOO and its various new directions is illustrative here. The infamous “Mr. Bungle” Rape in Cyberspace came after the first New Direction. The first form of the New Direction was characterized by the admins abdicating not their abilities (they still had their fingers on the power switch, and they still retained their abilities to affect the code, and take actions not permitted to other players) but rather their authority. They effectively demoted themselves to be the servants of the community, rather than its leaders.
In place of retaining decision-making powers, they instead signed up to do things based on player-driven governmental systems. As documented in Julian Dibbell’s My Tiny Life: Crime and Passion in a Virtual World, what resulted was first and foremost, apathy.
The abdication that the admins had engaged in was specifically directed at social forms of administration, you see. They remained as technicians, but not as parent figures. Without the parental figures around, some players moved into those roles but exercised purely moral authority; and most players simply continued to exist, since the intersection with administrative fiat in social matters was generally limited. Given a small enough community, with no major controversy at hand, the typical forms of communitarian pressure served to regulate the environment, and the result was a retreat to standard communitarian approaches.
It took the crisis of the Mr.Bungle incident to change that into a profusion of politicking, driven starkly by cliques, influence spheres, and yes, even idealism. What seemed like a homogenous community was suddenly fragmented along philosophical fault lines: importantly, philosophical differences in governmental approaches. In other words, the crisis surrounding Mr. Bungle emerged as being a discussion about the forms trust should take.
Dibbell describes the following constituencies:
- Parliamentarian legalist types, who favored the creation of laws by which to govern (trust mediated via contracts)
- Royalists, who favored the repeal of the New Direction and the return of wizard admins (trust in authority)
- Technolibertarians, who felt that governance was best accomplished by empowering users with tools (operating in a world with no trust)
- Anarchists, who in Julian’s retelling actually come across as strict communitarians, arguing for toading as being a form of shunning
The difficulty here, which manifested in several ways in the LambdaMOO environment, is that these groups are not necessarily motivated by even remotely the same issues. Where multiple overlapping but disparate constituencies exist, their needs must be resolved by someone with an understanding of the ecology of the whole; typically, only those in a higher position in some form of hierarchy are in a position to do that.
Consider the Schmoo Wars. In MOO softcode, you have the ability, as a player, to do quite a lot. One enterprising player decided to make an alternate definition of what a player character meant, in its core functionality and data structures. As I recall, he did this for his own purposes: to add sexual functionality to the avatars, as well as other bells and whistles which were not supported by the default definition of virtual “personhood.” However, the Schmoo player type was, according to its detractors, not as robustly constructed as the ones provided originally by the maintainers.
In a situation like this, you have sympathy with the creators of Schmoos, because you want to allow expressivity on their part; on the other hand, you also want the decision on the continued existence of the Schmoos to rest with someone who, if not an actual maintainer, has full access to and the knowledge of, a maintainer. And that is because while the maintainers may not be fully objective (after all, the existence of Schmoos is an implicit critique of their work), they are also best positioned to determine the impact of the presence of Schmoos, what security weaknesses they might, what technical issues might arise, and so on. Rank hath its privileges, yes, but the most undervalued of said privileges is perspective.
This perspective is what we hope our lawgivers have, when they construct and interpret contracts. Seen in this light, it’s frankly astonishing that something like national constitutions ever emerge, and highly unsurprising that they tend to be written by the rich and overeducated. And yet, the contract is exactly what we are developing towards in this story.
LambdaMOOs second take on the New Direction occurred after the the end of the Mr. Bungle incident. One of the wizards decided to use their powers to toad Mr. Bungle, thereby shortcutting all of the debate about what to do regarding the cyber-rape incident. This was exactly the sort of wizardly fiat that the New Direction was intended to eliminate, but it also brought into clear relief the fact that given such widely differing constituencies, none of which actually had the power of shunning, no form of trust could exist. Communitarian solutions were failing to even get a handle on the situation precisely because they lacked any tools.
After this action, the new take on the New Direction was put in place by Pavel Curtis, and it was the step of creating the tools: forcing upon the userbase a system of referenda and votes. In other words, a trust delegation system whereby the hierarchy was restored, albeit subtly. A contract was formed between those in power and those who lacked it.
Eventually, the implicit lesson of this system was made explicit, and “Another New Direction” restored the powers of the wizards.
On December 9, 1992, Haakon posted ‘LambdaMOO Takes A New Direction’ (LTAND). Its intent was to relieve the wizards of the responsibility for making social decisions, and to shift that burden onto the players themselves. It indicated that the wizards would thenceforth refrain from making social decisions, and serve the MOO only as technicians. Over the course of the past three and a half years, it has become obvious that this was an impossible ideal: The line between ‘technical’ and ’social’ is not a clear one, and never can be. The harassment that ensues each time we fail to achieve the impossible is more than we are now willing to bear.
So, we now acknowledge and accept that we have unavoidably made some social decisions over the past three years, and inform you that we hold ourselves free to do so henceforth.
1. We Are Reintroducing Wizardly Fiat
In particular, we henceforth explicitly reserve the right to make decisions that will unquestionably have social impact. We also now acknowledge that any technical decision may have social implications; we will no longer attempt to justify every action we take.
Players will still have a voice, however. Your input is essential. We will keep our existing institutions for now, with the modifications described below…
It’s worth pointing out that the most commonly repeated reason why the wizards felt obliged to do this was not out of a sense of idealism, but out of a sense of self-interest (emphases in the below are mine):
We explicitly reserve
(*) the right to veto any Arbitrator decision, particularly one that significantly impairs the ability of the wizards to do their jobs.
(*) the right to veto any Arbitration Change Proposal that is clearly not a “minor change” in the spirit of *Ballot:Arbitration (#50392) or that significantly impairs the ability of the wizards to do their jobs.
These may be temporary measures, as we hope to facilitate revision or replacement of Arbitration so that it may more adequately meet the needs of the community.
c. Wizardly Actions with Social Implications
The wizards will no longer refrain from taking actions that may have social implications. In three and a half years, no adequate mechanism has been found that prevents disruptive players from creating an intolerably hostile working environment for the wizards.
Thus it is demonstrated that arbitrators, the mediators of trust, are inevitably part and parcel of the community they arbitrate. And yet the need for accreditation remains. In the case of a virtual environment, the ultimate accreditation for the purposes of trust is not the merely technical ability to affect the playerbase, but is in fact the willingness to be considered a full participant in the community that the server represents. Whoever has their finger on the power button must be someone who would be just as devastated as any player to see it flicked to the off position.
The wizards in fact recognized this; in their statement, they proposed that were players not to favor this new direction, they could vote by simple majority to take LambdaMOO down forever. The offering of this choice actually clearly indicated their desire to be community members, and even worthiness to continue in charge of the system. They demonstrated their idealism by being willing to have all their work vanish if the playerbase objected. They effectively signed a contract with their virtual blood.
Building a tentative model
Demonstration of idealism is typically done via yet another accrediting body: the trade association, in the case of businesses. Essentially, once a business recognizes that it is part of two communities (customers and peers), it can be guaranteed to its customers via its peers. This only works, however, if the trade body is itself altruistic.
It hardly bears stating that finding a self-abnegating staff of saints doesn’t scale very well. We must rely on self-interest in some form — market forces, in the case of commercially operated services — to drive the authorities towards saintliness. Even then, there may need to be a guarantor of rights in order to create a trust framework, a contract that flows both ways, interacting with the communitarian aspect of the audience.
This contract is the document that stands above the authorities, and enables the community to enforce behavior on them, reversing the traditional hierarchy. In the real world, it is things like declarations of rights and national constitutions. In the technologically constructed worlds, it must be something that trumps the power switch, because in the end that clearly has the ultimate power.
Given that nobody has yet trumped the power of the power switch, and if we take all the above as a given, we are left with two choices: either we remake the environments themselves to no longer be dependent on the power switch, so that documents like declarations of rights and constitutions can come into their true power; or we accept the ongoing Balkanization of the Internet.
We see this happening today with content filtered on a national level, new root domains, and so on. We alse see it replicated in small in the virtual worlds, as users migrate from world to world, often attempting (somewhat futilely) to recreate lost communities, or just escape from allegedly repressive regimes. We also see the separation into communities of interest. Each user will naturally belong to more than one community, but there is still the risk of homogeneity mentioned previously.
This Balkanization occurs precisely because the various authorities involved (national governments desirous of curtialing the thoughts of their citizens, or protecting them, or protecting themselves; filtering software companies who wish to make money; game operators who wish to make money) and community members involved (various types of players, bloggers, and browsers who have a multitude of agendas and desires) lack unifying principles that bridge the gap between their differing agendas. They do not share one contract.
To recap: we have seen three broad models for trust management.
- Trust uses communitarian models to provide recourse in the event of a breach, and allows autonomy of response by the involved individuals. But it is vulnerable to attacks on the accreditation of the arbitrators. The virtue in trust is in belonging.
- Faith functions much like trust, but frankly does not provide recourse. We oeprate with a veil of confidence that is maintained by the relative rarity of its failure.
- Contract sets the virtue inherent in the contract itself; it may be the social contract, or a document that serves as a statement of principles in which all have placed their faith. The weakness in contracts is actually in identification; should a participant in the system not be someone who has signed up to these principles, then it may be very difficult to enforce anything. This can range from the guy who defrauds using a false credit card, to the cheater who does not provide a real id as they go about wreaking havoc in your virtual environment. Remember, ID leads to accreditation, which then leads to enforcement.
We can plot these on three axes, then: confidence, recourse, and autonomy of response (e.g., does the control rest withthe aggrieved party, or with someone else?). Trust hits all three, faith hits confidence alone, and contract hits recourse alone.
We can see that when plotted on a cube, there are other potential behaviors. In Ultima Online, a huge part of the social landscape was informed by a group called the anti-PKs; these folk operated solely based on autonomy, and had no metrics for confidence nor official recourse or sanction for their actions. They are the textbook vigilante. Other combinations of characteristics lead to other coping strategies. By and large, though, these coping strategies are not really what we consider to be civilized.
In fact, the modern definition of ‘civilized’ revolves around the contract, which is perhaps better termed “the rule of law.” Self-limitation (e.g., giving rights to the interactor, the citizen, the lowly A in our diagrams), is the ethical step of encouraging trust by self-enforcing, thereby putting oneself into the community rather than above it.
But this requires that governing entities be ethical — and in particular, we worry about commercial organizations, given that it has been argued that corporations are psychopathic in terms of how they relate to society.
But I’ll have to talk about that in a part four (!).
Cites from “LambdaMOO Takes Another New Direction” from “help manners”: Cyber-Democracy and its Vicissitudes by Charles J. Stivale, in Enculturation, Vol. 1, No. 1, Spring 1997.
You can follow any responses to this entry through the RSS 2.0 feed. Responses are currently closed, but you can trackback from your own site. [?]Type in a relevant tag, and click the button, and help organize this blog's information.
[More Help]
Unfortunately, as Dibbell has described LambdaMOO, the lack of other community-unifying features or processes was so great that there had to be a “rape in cyberspace” for anyone to come together about anything.
[...] read Dibbell’s work by now… but I haven’t, and I got stung. I’ll get around to that, I suppose…(Post a new comment) Log in now.(Create account, or useOpenID) [...]
Raph’s WebsiteOn Trust, Part III The Sunday Poem: Housebuilding Near Montague Farm Hail in San Diego PowerPoint Presentation - Putting the Fun in Functional Shaping perceptions Midnighters Another prescription/crystal ball Darniaq has seen the future of MMOs
[...] Also see Raph Koster’s analysis on its implications about trust and communities. [...]
[...] Earlier I mentioned the ways in which countries are chopping the Internet apart, creating national walled gardens where the perceptions of reality are literally different, thanks to the filtering of information. [...]
[...] That’s really the problem; of the three, it’s having a relationship with a community that doesn’t scale very well, as I have written about before in my series “On Trust” (1, 2, side note, 3, 3.5. [...]
[...] has been building a series on Trust. I find the meandering through all things real and virtual, personal and sociological, to [...]