Here's my ideal protocol for internet debate.

On day 1, Alice and Bob independently send their position statements to Carol the mediator. For example, Alice sends Carol a statement saying "The moon is made of green cheese", and Bob sends Carol his statement as well.

On day 2, Carol forwards Alice's statement to Bob and vice versa. Now they get a chance to see each other's arguments. But they don't write replies to each other. Instead, each of them rewrites their statement to strengthen it and preempt the other's criticisms, and sends it to Carol again. The next day Carol forwards them each other's revised statements, and so on.

This process repeats until Alice and Bob both say: "Ok, I've strengthened my position as much as possible and it is okay to publish as it is". Then Carol publishes both final statements side by side, and the debate ends.

The idea is that after enough rounds, Alice and Bob arrive at a kind of "Bayesian Nash equilibrium" of arguments. Not in the sense that they're both right, but rather in the sense that each debater's argument feels the most convincing according to that debater, and also they are both tired and don't want to continue.

And another part of the idea is that both debaters are trying to persuade the public, but can't do it by attrition or dirty tricks. Their only option is to produce a statement that will look good next to the opposite side's statement. So the public will end up maximally informed about the strongest points for each side.

Back to Vladimir Slepnev's homepage