For (b), when I’ve seen the prisoner’s dilemma presented, defecting always benefits the defector and hurts the other person (e.g. reducing the defector’s prison sentence by 1 year, while adding 2 years to the other’s) regardless of the other person’s choice; thus “you have an incentive to defect no matter what”. (In fact, usually it’s presented such that the effects of defecting are always the same down to the exact numbers—in TurnTrout’s article, the example yields (+1, −3) in all cases, although the formalized version with P, R, S, T doesn’t state a requirement that T-R = P-S.) So defection being enabled by the other’s cooperation is not an element of the normal prisoner’s dilemma.
One could say that defection is bad regardless, but that my defection being enabled by your desirable prosocial behavior (which should be rewarded, at least with praise and social capital) adds insult to injury. Then we could say the insult is what makes it rude.
Regarding TurnTrout’s post: I would characterize prisoner’s dilemma cases as “everyone should know that everyone will always be tempted to steal/etc., and therefore we should all be vigilant for that and expect defection to be punished”.
Stag hunt is different—if you’re hunting stag and I know it, then I actually have an incentive to hunt stag too; if I wimp out and don’t do it, that’s because I expected you to wimp out too. In practice, the failures with stag hunt games are down to communication and organization problems—”I didn’t realize hunting stag was an option, didn’t know we were playing this game”, or possibly “I disagree about the payoff, I think the stag will kill us even if we all join the hunt”. Therefore I think it’s inappropriate to use the same word “defect” for wimping out of a stag hunt, implying that people should be punished for it—I do not want e.g. the drunk friend who says it would be awesome if we all did some crazy thing to feel entitled to punish the sober friends who think it’s stupid. (Glancing at the comments, I see jimmy makes a similar point about stag hunt.)
As for the game of chicken… the real-world game involves two people agreeing to do something insanely risky and stupid, which is its own punishment as far as I’m concerned. I’m trying to think of relevant real-life examples. Wiki mentions brinkmanship in military or almost-military conflicts between states, which isn’t very relevant to individuals… and also the Hawk-Dove game, for which the closest modern analogue is “being mugged”, but I would figure muggers usually choose targets who look like they couldn’t fight back very well. I guess “aggressive driving on the highway” would fit. It is a form of defection—trying to get an advantage at the other’s expense—and it is technically enabled by them not defecting as well. But I feel like my main emotional reaction is “you’re crazy if you even try that”; also it is an inherently self-limiting problem, because if the proportion of insanely aggressive drivers gets high enough, they’ll kill each other off frequently enough to prevent it getting higher.
I might indeed consider it “rude” if aggressive drivers’ behavior made it so that, to protect against it, everyone else had to change their behavior, or change the rules, in a way that sucked for everyone. (I tend to disagree with the authorities when they lower speed limits, add speed bumps, and stuff—they seem to err harder in the direction of alleged “safety” than I would—but there are probably interventions that I would agree with if I saw the data.) So X might be “letting everyone drive on this road without stop signs in certain places”, and Y would be “driving on that road, too fast to avoid a crash if someone makes a turn at those places”.
I think the thing I’m getting at is “prosocial behavior should be rewarded, at the very least with social capital; it being met with defection is bad, but it being met with defection enabled by the prosocial behavior is maximally set up to disincentivize the prosocial behavior, and this is especially bad and deserves more focused disapprobation”. And I still suspect “rude” may be the right word.
For (b), when I’ve seen the prisoner’s dilemma presented, defecting always benefits the defector and hurts the other person (e.g. reducing the defector’s prison sentence by 1 year, while adding 2 years to the other’s) regardless of the other person’s choice; thus “you have an incentive to defect no matter what”. (In fact, usually it’s presented such that the effects of defecting are always the same down to the exact numbers—in TurnTrout’s article, the example yields (+1, −3) in all cases, although the formalized version with P, R, S, T doesn’t state a requirement that T-R = P-S.) So defection being enabled by the other’s cooperation is not an element of the normal prisoner’s dilemma.
One could say that defection is bad regardless, but that my defection being enabled by your desirable prosocial behavior (which should be rewarded, at least with praise and social capital) adds insult to injury. Then we could say the insult is what makes it rude.
Regarding TurnTrout’s post: I would characterize prisoner’s dilemma cases as “everyone should know that everyone will always be tempted to steal/etc., and therefore we should all be vigilant for that and expect defection to be punished”.
Stag hunt is different—if you’re hunting stag and I know it, then I actually have an incentive to hunt stag too; if I wimp out and don’t do it, that’s because I expected you to wimp out too. In practice, the failures with stag hunt games are down to communication and organization problems—”I didn’t realize hunting stag was an option, didn’t know we were playing this game”, or possibly “I disagree about the payoff, I think the stag will kill us even if we all join the hunt”. Therefore I think it’s inappropriate to use the same word “defect” for wimping out of a stag hunt, implying that people should be punished for it—I do not want e.g. the drunk friend who says it would be awesome if we all did some crazy thing to feel entitled to punish the sober friends who think it’s stupid. (Glancing at the comments, I see jimmy makes a similar point about stag hunt.)
As for the game of chicken… the real-world game involves two people agreeing to do something insanely risky and stupid, which is its own punishment as far as I’m concerned. I’m trying to think of relevant real-life examples. Wiki mentions brinkmanship in military or almost-military conflicts between states, which isn’t very relevant to individuals… and also the Hawk-Dove game, for which the closest modern analogue is “being mugged”, but I would figure muggers usually choose targets who look like they couldn’t fight back very well. I guess “aggressive driving on the highway” would fit. It is a form of defection—trying to get an advantage at the other’s expense—and it is technically enabled by them not defecting as well. But I feel like my main emotional reaction is “you’re crazy if you even try that”; also it is an inherently self-limiting problem, because if the proportion of insanely aggressive drivers gets high enough, they’ll kill each other off frequently enough to prevent it getting higher.
I might indeed consider it “rude” if aggressive drivers’ behavior made it so that, to protect against it, everyone else had to change their behavior, or change the rules, in a way that sucked for everyone. (I tend to disagree with the authorities when they lower speed limits, add speed bumps, and stuff—they seem to err harder in the direction of alleged “safety” than I would—but there are probably interventions that I would agree with if I saw the data.) So X might be “letting everyone drive on this road without stop signs in certain places”, and Y would be “driving on that road, too fast to avoid a crash if someone makes a turn at those places”.
I think the thing I’m getting at is “prosocial behavior should be rewarded, at the very least with social capital; it being met with defection is bad, but it being met with defection enabled by the prosocial behavior is maximally set up to disincentivize the prosocial behavior, and this is especially bad and deserves more focused disapprobation”. And I still suspect “rude” may be the right word.