Charles Lim from the Singapore Attorney-General's department noted that Singapore was taking a generally light touch approach to regulation - generally, holding back from premature regulation. Beyond certain boundaries, however (the examples given were inciting racial or religious hatred), Singapore would act in a heavy-handed manner and punish criminal behaviour. Charles was confident that the common law is sufficiently robust to adapt to challenges faced by virtual worlds, but demonstrated a reluctance to intervene before governments really understood the environments.
Joshua Fairfield noted that the social contract of virtual worlds could not be entirely constructed from contractual documents. The lack of privity rules out contractual arrangements as any real support for participant to participant governance. Fairfield argued that law is already here in most cases and agreed that the common law has the ability to adapt to novel disputes in virtual worlds.
James Grimmelmann argued that software was a large part of the problem for governance in virtual worlds, and while we are not able to solve the problem solely through software, it could help to a large extent. Grimmelmann argued that legitimacy would come from developing software based limits to the power of developers to arbitrarily make changes to the environment. Without this software, we are faced with a series of autocracies with limited legitimacy.
David Post rejected the suggestion that law is already and should be regulating virtual worlds. He asked firstly whose law applied, and questioned why European or US law should dominate these areas. Post beings his discussion from the premise that people have an inalienable right to form their own communities. To the extent that people have voluntarily chosen their law, that law should be respected. Post noted, however, that virtual worlds do not yet have law - they have code and rules, but not the legitimacy which law requires. He asks why developers have not provided a recognised system for the authoritative and legitimate exercise of power.
The common ground for the panellists seemed to be that virtual worlds need to foster legitimacy in their own systems in order to be isolated or immunised from territorial regulation. Fairfield noted that Post's approach was a normative ideal, but ignored the current situation where law has always already been active in virtual worlds.
I believe this is a reasonable proposition. Where there is real consent by participants to a certain meaning making framework, the law should definitely hesitate to attach its own significance to actions within the world. Where consent breaks down, however - where platform owners send conflicting messages to participants, or where the choices made by participants are not fully autonomous - then it is difficult to argue that the internal governance system is legitimate, and even more difficult to argue that it should be granted immunity from territorial regulation. The question of whose regulations would apply tends to draw an artificial and strained distinction between politics and law - virtual world owners always already change their environments to conform with governmental pressures, in proportion with the real power that those governments exercise.
One thing I'd add is that it seems strange to draw a bright line distinction here between 'games' and 'worlds more than games'. A democratic system of governance is neither necessary nor strictly desirable in these environments. Real consent can be given by the participants in environments which are controlled by a platform owner who provides a structured narrative experience. This should not limit the ability of the platform owner and the participants to create meaning in the environment, and does not necessarily involve inviting the law in to the environment. It is real consent which should underpin the choice of whether territorial political systems should impose their own interpretation on in-world actions, and not the specific form of the structure which the participants have agreed to.
Of course, the big caveat here is to be careful not to imply consent where it does not exist. Any analysis must go deeper than merely identifying the contractual or technological framework to critically examine the power relationships in the system.