Imaginary Worlds as Social Experiments

There are perhaps two ways to think about “worlds as experiments”. Making a world is an experiment in itself, an attempt to build anew a place for people to be. In this sense, to experiment is to try something new, to fiddle around with the tools at hand and see what happens. The other sense of experiment is more mundane: Experimenting is a scientific protocol, a recipe for knowledge discovery. To conduct an experiment, you hold certain factors fixed and then vary a specific thing in a specific way, and observe changes in something else, some outcome of interest.Many imaginary worlds, maybe the majority, have been experiments in the first sense. This essay concerns itself with experiments in the second sense. What about using synthetic worlds to conduct experiments along the lines of the scientific protocol? There are very good reasons to think hard about this, as outlined in the following sections. The possibilities and problems of doing synthetic world experiments occupy the rest of the essay. It is exciting to imagine using vast worlds with millions of people to discover solutions to important social problems. Making the discoveries would be quite hard. Games are hard to make; massively-populated games harder still; massively-populated games in which behavior is relevant to broader questions are perhaps beyond our current capacity. On the other hand, the sheer cost of making interactive systems is falling. It may soon be as common for people to make new worlds as it is for them to make short videos today. If so, we might face a bright dawn of new knowledge about how we all live together