Play of State: Sovereignty and Governance in MMOGs
Timothy Burke
Department of History
Swarthmore College
August 2004
There is a Bermuda Triangle in the study and design of persistent-world massively-multiplayer online games (MMOGs). The three points that sketch that triangle’s boundaries are virtual community governance, MMOG economies and player-to-player social relations within gameworlds, all areas that have received considerable attention. In between them all lies a near-unknown: the virtual state.
Defining the term “state” is one of the quintessentially contentious tasks of political and social analysis. One intellectual tradition defines the state as the political unit where sovereignty, practical and theoretical control of territory and society, resides. Another, derived from Marxism, tends to view the state as the political entity that maintains both a practical and conceptual monopoly over sanctioned uses of violence. Other traditions see the state as a highly specific and modern form of polity that is always closely tied to the nation, a highly centralized and organized territorial and political unit that expresses or follows the will and identity of an entire culture or society.
The most useful sense of the term to draw upon with MMOGs centers on sovereignty. What holds sovereignty over a persistent virtual world? What controls its virtual society, expresses its public interests, constitutes the ground upon which governance acts? I think there are three possible answers to these questions, and all of them suggest that existing MMOGs still fall far short of their aesthetic and ludological potential as world-simulations, with both practical and theoretical consequences. The first is that the sovereign in MMOGs is the developer; the second, that it is an artifact modeled within the gameworld itself; the third that it resides with player collectivities and organizations.
Developer as Sovereign
The developer is the most obvious and natural holder of sovereignty over the gameworld, just as the real-world owner of a server on which a virtual community resides is its autarch of last resort. If an entity can turn off, sell off, or erase a virtual society with legal impunity, it holds a form of sovereignty that the most absolute of absolute monarchs could only dream of. There are both customary and financial restraints on the use of such power: a steward who arbitrarily (or even justifiably) shuts down a virtual community can suffer a major blow to their reputation capital. A MMOG developer pulls the plug reluctantly, if at all, because the end of the virtual world is the end of the possibility of profit.
This is an inflexible and ethereal kind of sovereignty. A state whose only capacity to control or govern its territory is extinction has no nuanced instruments or practical day-to-day authority. Developers of MMOGs are the state in many other respects as well. One of the few areas where virtual sovereignty has been discussed, most inventively by Julian Dibbell, is in the context of developers as involved in a social contract with players governing property rights inside and outside of the gameworld. Dibbell argues that the developers meld together their role as a virtual state and real-world owner of intellectual property to designate proper and improper forms of economic behavior inside and outside the game, and in so doing, variably define the public interest of the societies over which they claim sovereignty. One MMOG developer may view eBay sales of player characters or virtual items as a violation of the social contract (rooted in the language of the end-user licensing agreement, EULA) and another may permit such activities, and in both cases, the developers act as the state, claiming that it is by their will that the virtual society can or cannot engage in a given activity.
All MMOG developers are the virtual state in an even more hands-on and everyday sense within their virtual worlds, however. The kind of social contract that Dibbell describes in some sense precedes and stands outside of the virtual world itself: a real-world person agrees to a EULA before he enters the gameworld as a character. MMOG developers also set and alter what might be called law in their gameworlds. Richard Bartle notes that in a persistent-world game, “law is code”. Bartle notes the problem that developers also frequently enforce something that seems closer to custom or tradition, that they sanction behaviors or activities that the code permits. But this only underlines how strange the law of a virtual gameworld is, because as code, it regulates and sanctions not just the behavior of human agents but also the physical and biological possibilities of such agency. The law of the developer in a virtual world makes it literally possible (or impossible) to act in particular ways. If a developer-sovereign wishes to forbid violence between its subjects (player-characters) in a virtual world, it can and does. If it wishes to permit them to trade any objects, or forbid them to trade in a particular class of objects, it can and does.
Bartle is right that this calls attention to how strange it is that such sovereigns frequently also claim the right to regulate the customary behavior of their citizens, often through intermediaries like customer service representatives. If a developer-sovereign truly wants to forbid a particular kind of action, why not make it impossible? In many cases, this is because the real-world capacities of human agents essentially outstrip the technical capacities of law-as-code. You can make a language filter that prohibits a player from saying “fuck”, but such a filter is too crude an instrument to deal with the slipperiness of real-world language: a code that stops “fuck” cannot deal with “F U C K”, “fock”, “you mother-forker”, “f*u*c*k” and so on.
This is the least of the reasons why the developer-sovereign does not make all law into code. More important by far is the nature of the relation between the developer-sovereign and the society over which it claims dominion. This is where the virtual state in MMOG gameworlds gets both seriously interesting and seriously weird. Dan Hunter and E. Gregory Lastowka have written about the issue of property rights but also of the rights of avatars within gameworld and in relation to developers. Thinking along these lines raises deeper problems: what kinds of subjects are players, and how are they constituted as citizens in relation to the developer-sovereign? What are the ways in which the developer-as-state tries to assess and express the public interest of their virtual worlds, and how structured is the role of the player-citizen?
If developers do not write all law as code, do not make it literally impossible to act in ways the sovereign prohibits or discourages, that may be partly for the same reason that real-world states do not fully use the powers at their disposal to regulate the behavior of their citizens. In the real world, some laws exist as communicative acts intended to produce interior states of moral inhibition or ethical constraint within the consciousness and everyday practice of human subjects rather than as strict prohibitions because the cost of maintaining an absolute prohibition is recognized by the state as being too high, in financial terms, in terms of lost human potentialities, or in terms of the risk of revolution from subjects who regard such absolute constraints as a violation of their social contract. States also maintain laws or regulations which they use only as opportunistic and circumstantial instruments against particular factions or groups or in times of arbitrary or contingent need, e.g., when a state needs money in the short-term, it may choose to aggressively enforce fines for activities that it normally ignores.
The same is true for developers in gameworlds. Writing code which makes it impossible to swear in any form would be the equivalent of hiring a policeman for every kilometer of roadway to enforce speed limits: an expense unjustified by the public good which that law seeks to enforce. Writing code which makes it impossible to harass another player would require making it impossible to use open-ended communicative language at all, which is what the MMOG Toontown does in order to prevent adult players from saying anything a child (or a child’s parents) might deem offensive. Writing code that comprehensively prevents the exchange of virtual items purchased on eBay might require preventing the exchange of any item. (Many MMOGs do this situationally, to prevent trade in particularly rare and desired commodities, by making those virtual goods “no-drop”, ownable only by the player who did the virtual labor to acquire them.)
Most complicatedly, however, the developer-state must fear the gameworld equivalent of revolution: the mass cancellation of accounts (or the failure to subscribe at all) in response to code-laws which either too tightly constrain the actions of players, or which alter the law of the gameworld in such a way as to forbid that which was previously permitted. However, developer-states also intervene constantly in the law (and custom) of their gameworlds under the banner of fixing bugs and exploits, addressing balance issues, adding new features, and trying to respond to the will of players. Here the question of what the public interest which the sovereign seeks to serve becomes especially acute, and the question of what kinds of subjects the gameworld’s citizens are, and what rights they have, becomes hopelessly muddled in most MMOGs. Here the typical developer-sovereign is revealed as a frustratingly unstable hybrid of deity, absolutist monarch, indifferent bureaucrat and responsive liberal government.
Every change to code is a change in law, just as every new customary prohibition is a change in the culture or norms of the gameworld. A precious few such changes are entirely neutral or positive, affecting all players equally. Everyone is pleased by a reduction in lag or the fixing of bugs that cause servers to crash. Here the developer-sovereign is at its most god-like, remaking the virtual world anthropically, to be a satisfactory home for player life-forms.
However, most changes in the virtual world carry the same potentiality that legal or constitutional changes in real-world states carry: they benefit some and punish others. New laws create new opportunities for transfer-seekers, while often displacing other, established transfer-seekers. Historically, different real-world states have adjudicated such consequences in different ways; more than a few commentators have suggested that contemporary liberal democracies have reached a point of serious crisis due to the influence of organized interest groups of transfer-seekers on the process of governance. Developer-sovereigns in most MMOGs have some of the same problems, aggravated considerably by completely non-transparent mechanisms of decision-making, arbitrary principles of communication and representation between sovereign and subject, and woefully inconsistent and poorly articulated understandings of the public interest of their sovereignties.
Most MMOG developers have treated communication with their customers as something of an afterthought, and have regarded it as a specialized activity to be managed independently from the development of the game itself. And yet, in every MMOG that I can think of, development decisions are clearly paced by both what players say within the game’s public sphere (both official and outside forums) and by what players actually do within the game, both in aggregate and in particular. The failure to make communication an integral part of the development process–the process by which the sovereign actually governs or possesses authority–often means that the game’s public sphere (e.g., forums) is chaotically organized and technically dysfunctional. It also means that the development team usually understands player behavior and opinion using extremely crude quantitative measurements of player activity, rather than direct instruments assessing sentiment and needs. A Roman emperor relying on oracles and the entrails of a goat might be better informed than some developers. The poverty of the public sphere in most MMOGs creates numerous opportunities for clever transfer-seekers: if a collective interest (a guild, a particular class, a group of eBay pharmers) can manage to dominate a particular channel of discussion simply through volume of expression or canny knowledge of the information ecosystem, they can often sway the decisions of credulous or unthoughtful developer-sovereigns. Equally, dedicated transfer-seekers can conceal or confuse information flowing to and from the sovereign.
From the perspective of citizen-players, it is often clear that the sovereign not only cares little for the quality or consistency of communication with its citizens, but also makes decisions and creates law in a manner that is almost entirely concealed from the citizens. It is not merely that the particular details of any given sovereign act are unknown, but even the process by which they are reached. People like Sanya Thomas (Dark Age of Camelot) or Kurt Stangl (Star Wars: Galaxies) who are charged with communicating the decisions of the sovereign (and representing the will of the citizens to the developer) are often one step removed from the process of concrete decision-making, and can offer only general descriptions of what has happened in any particular case, if that. Who decides, how they decide, and when they decide, are almost always unknowns.
Developers defend this non-transparency on various and shifting grounds: that such information is proprietary and must be concealed from business competitors, that if players are given too much information they tend to regard design concepts as contractual promises of future implementation, that players who know too many details find it harder to immerse themselves in the gameworld or get ideas for exploits, that developers are unskilled in communication and tend to antagonize players, or that the provision of such information is too time-consuming.
At least some of these are reasonable enough propositions when a MMOG is seen as a simple business venture, or as a one-to-many publication. MMOGs are not so simple, however, and the position of the developer-as-sovereign is too crucial to their functioning to ignore the deeper problem. MMOG players are notoriously alienated from and antagonistic towards developers precisely in part because they function as citizens in a sovereignty whose workings are permanently veiled. Law changes, custom changes, the physics and morality of whole worlds spin wildly, and with very little predictability or sense. It is not merely that developer-sovereigns hide their governance or manage information and communication poorly, but that in most MMOGs, it is also that the central conceptual underpinning of the sovereignty, its construction of public interest, is either contradictory or weakly articulated.
It can be taken for granted that the bottom line of the public interest in a MMOG virtual state is that it continues to exist. In a MMOG, this means that the revenues from operating it must outweigh its costs by whatever margin the owner deems minimally necessary. A small boutique publisher may set those margins tightly; a large publisher may set them more ambitiously. But even this lowest common denominator of sovereignty can be a difficult thing to judge concretely. What is the best course for a MMOG’s profitability, to satisfy a dedicated core of heavily-involved players or a wider array of more casual players? Profitability seems to suggest the latter, and yet, the developer-sovereigns of almost every MMOG on the market often have chosen to serve the interests of the former. Possibly because they judge that profit is best served by retention of the most reliable customers, but possibly instead because the core audience for MMOGs is symbiotically tied to the mindset of the developers themselves, and the two constituencies, sovereign and citizen, reinforce each other and keep the boundaries of their states tightly drawn. When the fuzzier list of values that a MMOG sovereign might seek to advance is brought to the table–fun, immersion, competition, world-simulation, social connection, community formation, emergent story-telling and so on–almost no MMOG has anything approaching a clear declaration of foundational values or bedrock principles. Contradictory or at least divergent conceptions of the public interest in any given MMOG are promulgated by developer-sovereigns largely as marketing rhetoric and are thrown like scraps to antagonistic communities of citizens who then fight with each other to determine the true foundational principles of the gameworld.
The State as Gameworld Artifact
There are a few important exceptions to the muddled picture of MMOG developers and virtual sovereignty. The game Second Life has established a very clear and purposive social contract with players around property rights and many other important issues that outlines the boundaries of developer sovereignty and even makes relatively transparent the kinds of powers and capacities which that sovereignty reserves unto itself.
More interestingly by far from the perspective of this essay, the game A Tale in the Desert (ATITD) not only explicitly embraces the developer as sovereign of the virtual state, but also works the developer sovereignty into the game as an artifact of the virtual world itself. In so doing, ATITD kills two conceptual birds with one stone. It offers a model of how to clearly and transparently structure relations between player-citizens and developer-sovereigns and it points out a huge game-mechanical gap in the design of most other MMOGs, that gameworlds which should have states within them completely lack anything of the kind.
The first achievement is a simple but important one, in light of the earlier analysis in this essay. Players in ATITD are sovereign over the gameworld within clearly described limits. They can pass laws, and insofar as the developers technically can either implement those laws as code or customary prohibition, they undertake to do so. If the developers interpret a player request as seeking a new feature, they classify it as such and take it under advisement. The limitations of the citizens?? collective power are explicit and transparent, and the process of communication between citizens and developers is highly specific and structured.
Placing this unusually explicit conceptualization of the developer-state within the fictions and mechanisms of the gameworld makes ATITD even more distinctive. All other developer-sovereignties are of the gameworld, but not in it. Moreover, very few other MMOG gameworlds have virtual states functioning within them, even as thinly developed fictions within the back-story of the game, let alone as functioning mechanisms with meaning in the gameplay itself. This is an absence that in many ways muddies the waters of developer-player relations even further, because it removes from the gameworld mechanisms that could very productively reside within it.
MMOGs with few pretensions towards being virtual worlds don’t really need virtual states. City of Heroes (CoH) or Planetside are highly combat-centered, and most of CoH’s content is instanced, meaning it is experienced largely in isolation from most other players. There is nothing within the game’s mechanisms to be governed, though CoH does actually try hard to maintain a story-telling conceit about the governance of Paragon City itself and the way superheroes fit into it.
But MMOGs which have strong internal economies and some sense of being fully-realized virtual worlds (Star Wars: Galaxies, Ultima Online, EVE Online, Anarchy Online, Dark Age of Camelot)–mostly lack models of the state within their gameworlds. EVE Online comes closest in various ways, most crucially with its construction of space sectors in terms of degrees of intrinsic lawlessness.
To take the example I know best as an illustration of why an absent state can be a problem, Star Wars: Galaxies has a classic faucet-sink economy where fees of various kinds are assessed in order to remove currency from the game and where currency enters the game directly through players being paid for missions or quests. It also circulates from player to player as payment for goods manufactured by players from resources taken from the environment or as payment for those resources. The game will pay out currency for goods or resources (destroying them in the process) as well but at rates so low that no player bothers.
The problem is the sovereign entity that sets both fees (the sink) and payment for missions and resource recovery rates (the faucet) is the developer. It is not a state within the game, though player cities are able to assess various taxes directly as an overlay over these developer charges. There is no state within the game as a mechanism (and a fairly poorly developed in-game fictional simulation of the Galactic Empire of the Star Wars films, but that’s an issue for another day). What this does is recurve back the entirety of the virtual economy of the game onto the problems of the developer sovereignty that I have already discussed, even though the economy is notionally player-operated. The total supply of currency, the flow of currency, exist entirely outside of the economy that the game mechanically tries to implement. Value comes from a singularity outside of the gameworld’s space-time and falls back out of it similarly. In practical terms, this is one reason that the game’s early hyperinflation should hardly have been a surprise: mission payouts were set according to some external metric rather than according to an internally generated valuation that made sense to an in-game sovereignty, so missions that were wildly miscalibrated in terms of risk-reward were done repetitively with the exuberantly mechanistic intensity that dedicated powergamers are so well known for, and by the time that the external developer sovereignty noticed, the economy was awash in an amount of currency that the sink fees would have needed ten thousand years to absorb. But after tightening the faucets, with the huge currency overflow inexorably moving to a small handful of plutocrats, it was impossible to raise those fees without grievously punishing average players.
Imagine instead that mission payouts were set entirely by players in a galactic market and so corresponded to actual in-game rubrics of value. Say I need 500 hides of a particular kind, so I set a fee as payment for delivery of hides, with the fee and delivery of goods set entirely remotely, with players able to readily compare pricing for services. Imagine equally that buildings and objects were affected by a transparent, published physical and labor-cost model that was consistent over time. (e.g., that facilities suffered physical damage and decay at a set, highly accelerated rate, and that they had an internal model of the invisible labor costs needed to keep them running.) Whether or how much player cities or individual players charged to cover these costs would be up to them. That’s one in-game model of the state (essentially as absent as a mechanism, entirely devolving onto players and markets.) Another might be making the state a mechanistic part of a so-called closed economy that the faucets and sinks of the economy should be dynamically and automatically related, e.g., that the revenues a virtual state receives from its sink have to be balanced with its faucets, and that the services the virtual state pays for are dynamically and meaningfully related to its environment. So in this case, let’s say that if the city of Corellia does not destroy X number of animal lairs within 2000 meters of the city every month it loses Y functionality or building and is forced to pay to restore that functionality or building. If it loses sufficient functionality, no players will go there any longer, and at a certain point, the virtual state ceases to be. So the state raises payouts for lair destruction until the quota of lairs is reached and then drops it considerably, and charges accordingly on the fee side in order to pay for the mission payouts. Tying high-value resources consistently to highly dangerous environments (where mission payouts and fees are both very high, and the cost of operating businesses in the area accordingly volatile) would help complete this mechanism.
Certainly adding the virtual state in any of these ways would do a lot to complicate and enrich most virtual-world style MMOGs, adding a new layer of gameplay with many opportunities for arbitrage and strategic economic behavior. Just as developers have a combination of god-like powers over the physical and political power over law and culture, in-game states can only work if physical and environmental rules are aligned with their mechanistic functioning. But the lack of in-game states in economies that clearly reference and require them is simply one more way that some MMOGs aggravate their already considerable problems with developer sovereignties, and implementing virtual states within games might be a way to consistently move some of the invisible, non-transparent, non-responsive deliberations of developers into the game itself.
Player Governance
Many MMOG developers play lip-service to player governance and a substantial number of them have tried to provide in-game mechanisms for facilitating governance. However, such mechanisms tend to make the distinction between a sovereign state and a government very clear. Players in MMOGs are often given tools that allow them to structure, organize and even govern their own associational life, but these associations rarely possess even limited sovereignty over the gameworld or even over the players themselves.
The central device of player governance in MMOGs is guilds, and as MMOGs have evolved, not only have they become standard features of the genre, but also they have acquired standard capabilities and structures. Guilds typically are voluntaristic associations of three or more. They differ from groups that form for only the duration of a single adventure or playing session: they are persistent within the gameworld. In most MMOGs, guild formation requires a one-time fee in order to record the guild’s existence. Usually a recognized guild will gain access to a designated emblem, its members will have a guild tag after their names, and a dedicated chat channel that allows guild members to talk exclusively to each other.
Various MMOGs give guilds differing degrees of sovereignty over their members. No MMOG that I know of allows a guild to keep a member from leaving his or her guild, and most require players to voluntarily donate goods, services or currency to the guild rather than to allow them to be involuntarily taxed by the guild at regular intervals. Typically MMOG guilds have powers over membership that can be granted to all, few or only one of the guild members–the power to admit members, expel members, promote members, access collective property and the like. Many MMOGs give guilds the ability (at an in-game cost) to place buildings that belong to the guild rather than individual players.
The limit of guild power over the gameworld is usually the ability to place buildings. Star Wars: Galaxies adds the extra layer of complication that it also allows players to place and build cities with elected mayors who can assess taxes on buildings and sales within city boundaries. In practice, these cities tend to be associated with a single guild but they are not required to be, and some of the largest have three or four guilds within their borders. Informally, in many MMOGs, guilds grow so large and powerful that they begin to exert considerable social power over all players on a given server. This is most famously the case with Everquest, where many high-level activities require large and well-organized groups, and the very largest guilds can essentially control the activities of the rest by threatening to interfere with the adventures of others. In such a case, a player guild becomes the undeclared sovereign of a larger community, but such authority is often extremely unstable, and in any event, has no persistent, structurally permanent effects on the gameworld itself. In Dark Age of Camelot, for example, player guilds can form alliances, but these have no persistent or mechanistic authority within the larger realm or player kingdom to which they belong, even though this would make a certain amount of in-game sense. Even a game like Shadowbane which centers on conflict between player guilds does not map player governance into sovereignty over the gameworld: players govern their own cities and attack others, but do not pass laws or create structurally persistent artifacts which do more than maintain a guild and the buildings it creates within its own city. The disconnect between the virtual state as a gameworld entity and player governance is fairly substantial.
Most developers and players, I suspect, would feel that this is just as it should be, given the demonstrated possibilities for mischief when that connection has occurred. (It will be hard to forget that a group of players who managed to hack developer-level powers for themselves in Shadowbane immediately began teleporting rivals to the bottom of the ocean.) I would generally agree, though there is probably room to experiment with making player governance edge closer to real control over a persistent in-game state or sovereignty. However, this also means that player governance cannot serve as an alibi for the deeper structural problems that exist with sovereignty in MMOG gameworlds. It stands aside from those problems, and is not the resolution to them, not without a much more breathtaking leap into unknown terrain explored to date only by A Tale in the Desert. Even there, ATITD keeps the developer in the picture, not just as passive steward, but as part of a composite sovereignty that is half in-game, half out-of-game, half-player and half-publisher. ATITD is a hybrid, exotic boutique game for a niche market, but its clarity on this key set of issues ought to be closely scrutinized by MMOG developers and players alike.