Aliens And Prisoner’s Dilemma
February 18, 2010 Leave a comment
The problem of whether to commit genocide upon an alien race or not is vaguely related to the famous“prisoner’s dilemma”.
The problem is that the Prisoner’s Dilemma makes it all too likely that Paranoia beats reason. For those unfamiliar with it… here’s the Space version.
Race A and B both have roughly comparable technology, but don’t understand each other. Each race has 2 options: Launch missiles or Ignore each other.
If Both races open fire, both races are devastated but not destroyed.
If one race opens fire and the other ignores it, they’re utterly exterminated.
If both races ignore each other, they live in peace and are fine.
The problem is, neither can really communicate with each other. And although the cooperative choice of ignoring each other is best, the risks of them firing first while you ignore them are too great. Thus, this scenario via game theory, will always result in missiles being exchanged.
|Race B Ignores||Race B Attacks|
|Race A Ignores||Both live constant fear||Race A exterminated
Race B lives free of fear
|Race A Attacks||Race A lives free of fear
Race B exterminated
|Both are devastated but not destroyed|
As the wikipedia article shows, the dilemma comes when you assume that each race is trying to maximize it’s survival.
Say you are Race A. If Race B ignores you, your best outcome is to attack. Then you do not have to live in fear, spend resources on building defenses, and so on. If race B attacks, your best outcome is still to attack, since the alternative is extermination.
And since Race B will make the same determination, both races will attack and be devastated but not destroyed.
An outside observer will note that if the two races are taken as a group, the best outcome of the group is for both races to cooperate. If either attacks, the outcome for the group will be worse. And if both attack, both races receive a worse outcome than if they had both ignored each other.
So if both races selfishly look out for themselves, both will attack and the result is devastation. If both races altruistically think about the group, both will ignore and both will live. And if one race is selfish while the other is altruistic, yet again it will be proven that nice guys finish last.
And it actually doesn’t matter if they can communicate with each other or not, a given race cannot be sure if the other is being truthful. If the two races can communicate, they run into the “cooperation paradox”. Each race must convince the other that they will take the altruistic optiondespite the fact that the race could do better for themselves by taking the selfish option.
Of course the prisoner’s dilemma is a very artificial set-up, in real life the results would not be quite so clean-cut. The basic prisoner’s dilemma matrix looks like this:
|Cooperate||win some-win some||lose all-win all|
|Defect||win all-lose all||lose some-lose some|
In more detail, it looks like this:
|Cooperate||D, D||C, B|
|Defect||B, C||A, A|
where A, B, C, and D are various outcomes, and the relative value of the outcomes are B > D > A > C. If those relative values are true, the prisoner’s dilemma is present. In the first example, B = alive and free from fear, D = alive but in constant fear, A = alive but devastated and C = exterminated.
The prisoner’s dilemma does have some vague similarities to the old cold war doctrine of Mutual Assured Destruction, though they are actually not very closely related. The prisoner’s dilemma also does not work in those cases where what is bad for one player is equally bad for the other.
But, still what we could speculate is that neighbouring civilization would like to live with peace instead wasting their time in wars.