VEIL OF IGNORANCE. Cal Pundit suggests extending experimental economics into philosophy. For example:

"What I'm thinking of is a combination of philosophy and experimental economics. Test subjects are presented with three different "worlds," each with 10 people and $100:

World 1: 9 people get $1, one person gets $91.
World 2: 5 people get $3, 3 people get $10, 1 person gets $20, and one person gets $35.
World 3: All ten people get $10.

Each contestant gets to pick a world, the dice are thrown, and the money is doled out. Since your 'position' in the world is randomly chosen, it's a reasonable approximation to the "veil of ignorance" that Rawls proposes as a precondition for the people choosing how society will be ordered.

Has anyone ever tried this? It seems like an interesting little experiment to me.

UPDATE: Actually, how about a little addition to make this fit Rawls's notion a little better: the test subjects are separated into groups of ten and put in a room where they have to decide (any way they choose) which of the three worlds they want. Once they've decided, all ten draw a ticket from a hat and get the amount of money on the ticket. It would be pretty interesting to listen in on the conversations, too!"

I don't know of anyone who has run such an experiment, but Steven Landsburg has some thoughts that I've borrowed from to propose a slightly different experiment. Rather than divide $100, try something like this. Keep Worlds 2 and 3. Replace World 1 with this: 3 people get $15, 7 people get $22. There is more to economic justice than slicing the pie: our actions also determine the pie.

Landsburg gets more provocative, asking, anent Rawls, whether an individual behind the veil of ignorance was ignorant of his or her role: would all be humans, or would some be assigned the role of a baby seal, or of an aborted fetus (Armchair Economist, p. 58).