3 ms·

The games that we participate in are themselves defined by human actions. Do we rationally choose the games we play? Which of my preferences are compatible with

The games that we participate in are themselves defined by human actions. Do we rationally choose the games we play? Which of my preferences are compatible with your preferences? How do we compare utility between two individuals? Since we live in the same universe and have to deal with some of the consequences of each others' actions, there's some natural correlation between our utilities, but we live in civilizations which have far more social structure built on top of this basic structure, and these social structures inform and structure our possible preferences and the relations between them. In fact, these social structures can have preferences and play their own games! The United States government, for example, is an entity whose preferences and actions are embodied in the coherence of its constituent individuals, and its structure informs the games and preferences of its individuals in a feedback loop we like to call bureaucracy. Does the large-scale behavior of the US government explain the choices of its constituents, or vice versa? Clearly the answer is neither and both: They are two ways of looking at the same dynamical structure. In one view, things can look perfectly rational. In the other view, things can look stark raving mad. I don't think either point of view is distinguished as canonical.

mikeash 11y ago
Sometimes benefits and externalities depend heavily on social structures. Various complex financial arrangements would qualify. Sometimes they're entirely natural, like if I make a fire for warmth and the smoke blows to where you are. Antibiotics in farming looks much closer to the latter than the former to me. I agree that neither point of view is canonical. That is basically my point: it's only madness from a certain perspective, and it's completely rational from another. Neither is right nor wrong, it's simply that the different perspectives are in conflict.
TeMPOraL 11y ago
> In one view, things can look perfectly rational. In the other view, things can look stark raving mad. I don't think either point of view is distinguished as canonical. I'd say that's not the point. The point is, individual selfish choices aimed at producing good outcome for the choicemaker start to produce bad outcomes for the choicemaker when everyone else makes the same choice too, but at this point it's too late to leave the ship. A canonical toy example would be working overtime. Say everyone is working 8 hours, but you decide to stay one hour longer. The boss notices and gives you a bonus for your dedication to company. Your co-workers notice, some of them decide to stay an hour longer too. Since the bonus pool is limited, you all get somewhat smaller bonuses. Ultimately so many people stay the extra hour that it turns into a new normal; 8h/day workers become seen as slackers and are pushed to 9h/day mode to avoid angering the boss. And since it's a normal mode, nobody gets bonuses. End result: you work 9h/day and get paid the same as you used to be before. The article I linked[0] has much better examples of this happening in real-life. The point is, things can look "stark raving mad" pretty much objectively, because the end result is bad in the choicemaker's value system. It's easy to get into a coordination trap and incredibly hard to get out of it. [0] - for convenience, http://slatestarcodex.com/2014/07/30/meditations-on-moloch/ http://slatestarcodex.com/2014/07/30/meditations-on-moloch/

mikeash 11y ago

Sometimes benefits and externalities depend heavily on social structures. Various complex financial arrangements would qualify. Sometimes they're entirely natural, like if I make a fire for warmth and the smoke blows to where you are. Antibiotics in farming looks much closer to the latter than the former to me. I agree that neither point of view is canonical. That is basically my point: it's only madness from a certain perspective, and it's completely rational from another. Neither is right nor wrong, it's simply that the different perspectives are in conflict.

TeMPOraL 11y ago

> In one view, things can look perfectly rational. In the other view, things can look stark raving mad. I don't think either point of view is distinguished as canonical. I'd say that's not the point. The point is, individual selfish choices aimed at producing good outcome for the choicemaker start to produce bad outcomes for the choicemaker when everyone else makes the same choice too, but at this point it's too late to leave the ship. A canonical toy example would be working overtime. Say everyone is working 8 hours, but you decide to stay one hour longer. The boss notices and gives you a bonus for your dedication to company. Your co-workers notice, some of them decide to stay an hour longer too. Since the bonus pool is limited, you all get somewhat smaller bonuses. Ultimately so many people stay the extra hour that it turns into a new normal; 8h/day workers become seen as slackers and are pushed to 9h/day mode to avoid angering the boss. And since it's a normal mode, nobody gets bonuses. End result: you work 9h/day and get paid the same as you used to be before. The article I linked[0] has much better examples of this happening in real-life. The point is, things can look "stark raving mad" pretty much objectively, because the end result is bad in the choicemaker's value system. It's easy to get into a coordination trap and incredibly hard to get out of it. [0] - for convenience, http://slatestarcodex.com/2014/07/30/meditations-on-moloch/ http://slatestarcodex.com/2014/07/30/meditations-on-moloch/