The paradigmatic example of an institution which faces this question of political legitimacy is the modern bureaucratic state. As Max Weber famously observed, the state is the entity “that (successfully) claims the monopoly of the legitimate use of physical force within a given territory” (1946:78, emphasis in original; online version here). The legitimacy is important, of course, because ruling by illegitimate force is costly — one wants to release the hounds only when necessary. But on what grounds is this legitimacy claimed? The bureaucratic state, by itself, can refer only to its procedures as right in and of themselves, and this is a weak foundation on which to construct a sense of belonging and legitimate rule amongst a citizenry. The nation, however, is a rich font of such meanings, containing as it does arguments about shared language, culture, history, and territory (all in the context of an “us” that is not “them”). So the story of the nation-state (as it is usually called in the literature) is one that involves an intimate relationship between the institutional apparatus of the state and the symbolic resource of the “nation”. (NB: I am collapsing to an unforgivable degree what is an enormous body of scholarship on the state and its relationship to the nation — this list is a place to start, at least.)

For virtual worlds, however, there is no central metaphor of the nation already in place. Certainly, as they persist they are generating shared languages, meanings, and practices, to a certain degree, and one sees attempts to port offline categories of territorial belonging into virtual worlds, with varying degrees of success. But I have noticed something else, and it leads me to ask whether games and the indeterminate outcomes they generate can be a source of political legitimacy.

Recently, a graduate student of mine, Krista-Lee Malone, posted a paper at ssrn (Malone 2007) based on the master’s thesis she completed at UW-Milwaukee. The thesis was about hardcore raiding guilds in World of Warcraft, and the paper looks comparatively at the adoption of DKP systems by two guilds. Malone suggests that DKP systems create “player obligation to the guild through a rationalized system intended to measure commitment” and that they therefore constitute a core practical component of how guilds as institutions establish legitimate rule. The gameness is relevant in DKP, because player performance in the game is a factor (50 DKP MINUS!).

Similarly, when doing research at Linden Lab I noticed a number of attempts to deploy games as mechanisms to generate outcomes that could stand as legitimate and, importantly, be an alternative to, for example, top-down decision-making, or democratic voting. One such attempt was through the application of the Elo Rating System.

Second Life is a good example of a project – and Linden Lab has been a good example of a company – that is deeply informed by the “left-libertarian” attitude toward technology and its promise which has shaped everything from the structure of the early internet to the proliferation of personal computers, but about which I do not have sufficient space to say more here (see Turner, 2006). Suffice it to say that in a political sense Linden Lab in 2005 was characterized by an almost overpowering faith in technology, matched only by a similarly monumental suspicion of vertical authority – especially bureaucratic authority, although charismatic authority was suspect as well. The paradox for Linden Lab was how to realize the ongoing creation of Second Life in a way that was consistent with their idealized vision of individual creativity and liberty, while remaining indisputably and unavoidably the single most powerful institutional player on the scene.

Over the course of my field research at Linden Lab, I discovered that this tension between vertical control and horizontal, “emergent” governance was not only a key to understanding their struggles to make Second Life, but also of their struggle to make themselves as an organization. For as Second Life grew in size and complexity, so did Linden Lab, and this tension came to be the preoccupying focus of their own organizational lives as well. This preoccupation was the result of the same politically-charged disposition, one which tended to treat top-down or vertical decision-making as the antithesis of empowered and creative collaboration. As people at Linden Lab witnessed their creation and their company growing, this fear of a loss of liberty reached, at times, a fervent pitch, and in this ongoing predicament they are not alone in high-tech circles. Google, as recent coverage by several journalists has revealed, is similarly shaped by a disposition that combines a deep faith in technology with a rejection of vertical authority. Julian Dibbell has also recently charted the strange advent of what he calls ludo-capitalism, wherein labor experience is increasingly framed as (and constituted to be) a game. We should therefore be eager to understand how the entities that have their hands deep into the recesses of our digital lives are going about trying to solve their version of the challenge that famously preoccupied nation-states: how to establish legitimate institutional power in the face of practical and undeniable imperfections and limitations.

Elo rating systems are a group of statistical methods for calculating the relative skill levels of large numbers of players for two player games. Based on a system developed originally by Arpad Elo for generating a ranking of chess players, they have since been both modified and improved within chess and adapted for other two-player games. A ranking system generates a rating for each player, and is seen as legitimate in the degree to which these ratings seem to accord with the matches that do manage to get played. Thus, a key aim of these systems is also to predict the outcomes of matches between rated players, and its accuracy is thereby judged (and thus the system may also come to be modified). In this way, Elo rating systems generate an emergent ordered ranking, and this emergent quality made this technique an attractive solution for the challenge that faced Linden Lab: how to generate a ranked order of prioritization from a heterogenous collection of company tasks.

In mid-2005, one developer at Linden Lab, himself quite familiar with chess ranking systems, set about to code onto Jira a system built on Elo’s algorithm. He created a webpage that pitted two (and only two) tasks against each other for Lindens to choose. These “matches” would over time generate a list of highest-ranked to lowest-ranked Jira tasks. Rosedale enthusiastically supported this effort, and in two days the programmer had created the system and sent an email over the company-wide email list containing a link to the site. Upon arriving at the site, one saw a simple presentation of two Jira tasks, including each one’s title, unique Jira number, and a brief description. Employees were simply to pick one of the two (or pick a “draw”) and the system would record that match result and generate another match of two more tasks (soon after they were each asked to pick winners of ten such “matches” a day). Many Lindens tried out the system with some enthusiasm, as it seemed relatively un-“gameable”. Hundreds of matches were “played” in a short span of time (a matter of days), and a ranked list was generated. For Rosedale, this was a step on the road toward realizing an ideal of company decision-making from the ground up. For others, the system was suspect at the point of participation; presented with two entirely heterogenous tasks (add a urinal to the men’s bathroom vs. add a web browser to the Second Life client), they felt that picking between them was nonsensical. It was eventually abandoned in practice and other game-based (and non-game-based) initiatives to tap into the wisdom of Linden Lab’s crowd were tried.

So, I am left with a lingering question. Was the ultimate illegitimacy of the Elo-ranking system due to something deeper — the fact that it was an attempt to incorporate a game into corporate decision-making? The suggestion I would like to make is that, for institutions, games are in fact quite difficult to domesticate, precisely because they can generate outcomes that challenge or outright contradict any existing, more coherent, narratives. And I would add to this another issue, just as important, and that is how the legitimacy of a game’s outcomes is directly related to the community of its players or the institution which controls it (as in the case of organized sports). When games are mobilized for purposes other than the playing in and of themselves, who gets to interpret the outcomes, and say what they mean? The sponsoring institution, or the participating players? To me, these are central questions as we see more and more institutions attempting to govern through games. While Julian has rightly focused on what this means for labor and exploitation, in the Marxian sense, I think there is a related (Weberian) question that we must attend to as well — how are institutions learning to use games to establish legitimate governance?

References

Malone, Krista-Lee. (2007). Dragon kill points: The economics of power gamers. Available at SSRN: http://ssrn.com/abstract=1008035.

Turner, Fred. (2006). From Counterculture to Cyberculture: Stewart Brand, the Whole Earth Network, and the Rise of Digital Utopianism. Chicago: University of Chicago Press.

Weber, Max. (1946). From Max Weber: Essays in Sociology. H. H. Gerth and C. Wright Mills, eds. Oxford: Oxford University Press.