This presentation is based on my article Virtual Worlds as Comparative Law, 49 N.Y. L. Sch. L. Rev. 149 (2005). In it, I also discuss contract law, criminal law, and constitutional law.

I am taking seriously the claim of virtual worlds to be distinct places, not because it is necessarily true, but because doing so provides a good discipline for thinking about how virtual worlds actually operate. That exercise helps us understand their distinctive problems and solutions, which is very useful for anyone who needs to deal with them professionally. Ultimately, it becomes necessary to talk about the players behind the avatars and to point out the overlaps between virtual worlds and the real one. But if we defer doing so for a while, we'll have a much better sense of what those overlaps actually imply.

Richard Bartle's classic article Players Who Suit MUDs, provides a good introduction to player motivations. Whether or not one agrees with its taxonomy (I would add "profiteers," for example, and note that even those motivated solely by money display a wide range of play styles), it's a good reminder of how complex and varied the psychological forces driving players to virtual worlds are. For every possible motivation for playing, there are corresponding motivations to value virtual property. It's a worthwhile exercise to match up a motivations taxonomy against these categories of valuable goods; one quickly notices that different people will value the same virtual object for different reasons.

This slide is a straightforward application of Lessig's Code: And Other Laws of Cyberspace to virtual worlds. Software sets explicit limits on what can happen in the virtual world; designers are the ones with the power over the software. This gives them enormous power to shape what is even possible in their worlds. Players then create a rich society based on social norms, some of them involving formal software-mediated rewards and punishments, others involving social censure (e.g. "d00d, not K3WL!"). The critical crossover is that the software itself also subtly informs the kinds of interactions people have in virtual worlds. It can help set a tone of friendliness or constant betrayal; it can facilitate self-ordering or hinder it; it can dramatically affect the kinds of groups that players form. The consequences of a change to the software can be both subtle and far-reaching.

Some worlds have tried "steal" commands, or allow avatars to kill each other for looting purposes. But the norm is a very interesting conflation of possession with ownership, as far as the software is concerned. Drop something, and whoever next picks it up can do whatever they want with it. (Of course, the real world operates this way, too. The law of gravity will not protect your dropped wallet from a thief.) The flip side of this is that some kinds of "property" are bound to an avatar. Some worlds have virtual items that cannot be discarded. Also, in a sense, whenver a world allows avatars to change, whatever new attributes they acquire become their inalienable property.

Public property in almost all virtual worlds includes most areas of land. (Worlds with no player-owned real estate at all have a complete commons in land!) Often, there are also large quantities of property implicitly "owned" by non-players, such as virtual "merchants" whose in-world shops buy and sell virtual property to players for in-game currency.

Some worlds, such as There and Second Life, give players more or less extensive freedom to design content that is added to the world. In others, the stock of virtual property is fixed; things come only in the forms that designers have created for them.

I am not kidding about capturing virtual animals being a main source of wealth. The activity known as "gold farming" in many virtual worlds almost always involves repeatedly killing the same virtual monster again and again.

The coordination-problem aspects of virtual hunting are fascinating. The most common paradigm is that monsters beyond a certain level of dangerousness (either because of their intrinsic ferocity or because they travel in groups) require the combined effort of a group of players to kill. Often, the players need to specialize in their roles within this group (e.g., a "tank" is a tough avatar whose job is to distract a monster and stand up to its assault while others in the group work on actually killing the monster) and to use careful coordination. A failure of coordination can be deadly to the avatars involved. A second kind of coordination problem arises when more than one group wishes to "raid" the monsters in a given area. The ensuing chaos can be incredibly dangerous. Indeed, it is possible for a small group of players deliberately to ruin another group's raid. The politics and group dynamics involved are, let me repeat, fascinating.

The problem of dividing up the spoils appears to be one that software on its own will never resolve to players' complete satisfaction. Every world has its own subtly different set of social norms governing who gets what after a successful hunt. Here, we see a perfect example of a layer of norms adapted to the software layer beneath.

For a great eye-opener on the value of virtual real estate, see Julian Dibbell's The Unreal Estate Boom. I predict that within a decade, real-world lawyers will regularly be paid to draw up leases and other transactions for virtual real estate. Second Life is already seeing enormous economic activity (including speculation, subleases, and professonal subdivision and construction) around its land system.

The Sims Online example I am thinking of involves a regular problem of hostile takeovers of virtual real estate by unscrupulous players. They trick others into making them joint owners, and then use their software-granted powers to make enormous changes and to kick the original owners out. Joint ownership systems are not easy to make work in real-life law; making them work with software is even harder.

Nuisance is another perennial issue in land law. It's not clear that it's even capable of good doctrinal elaboration. Each generation's sense of noxious uses is different. Nuisance issues also regularly crop up in virtual worlds. Second Life had issues with dueling political displays directly opposite each other. A Tale in the Desert saw one player post signs that others considered highly offensive. The developers took the interesting response of introducing a voting system. If enough other people flagged a structure for removal, the software would remove it. (By way of comparison, the constitutionality of such systems in the real world is ambiguous and fact-sensitive.)

A group is an institution that systematically generates conflicts it itself is incapable of resolving.

Virtual worlds matter a lot to players. Conflicts among the players are inevitable. Sooner or later, these conflicts will spill over. Real-life lawyers are going to be among the first responders summoned to the scene. To sort out the mess, you'll need to know what's really going on inside virtual worlds. You shouldn't underestimate the power of players and designers to sort out some messes by themselves. To know which messes they can handle on their own, you'll again need to know what's really going on inside virtual worlds.

Thus, my closing message to this audience of lawyers. Pay close attention to virtual worlds. It'll be good for the players and designers to have thoughtful legal help available. It'll be good for virtual worlds to have real-life law working with them, not against them. And it'll be good for you as lawyers to understand the nature of the problems you'll be dealing with.