SAE Cooperation Theory Special Paper: Changing the Game / SAE 合作论专论:改局

This paper reads the threshold of cooperation. There are two questions: I want A, you want B; how do I get a result as close to A as possible? I want A, you want B; how do we do C together? A game, as this paper speaks of it, is only a situation in which one end, in this undertaking, asks only the first question; this paper does not define the word for the discipline of game theory. The cooperation this series reads requires that both avow this undertaking and actually take part; this paper reads the stretch before the avowal: one end is headed in the 15DD direction, the other, in this undertaking, asks only the first question, and no shared undertaking has yet been avowed. It does not ask where cooperation in general comes from. 15DD and 14DD here are stances, not labels; at the operation layer this end certifies neither that the other is headed in the 15DD direction nor that the other asks only the first question; what he can recognize is only two things: that there is as yet nothing jointly avowed, and that the other's words and deeds do not yet show the second question. The end headed in the 15DD direction asks both questions. Within the game he reckons as usual and guards his own share; the gains given within the game do not exhaust the direction in which he changes the game. Changing the game is altering the conditions of a matter: who can do what, whose share hangs in whose hands, what words can be said, what can be avowed. The end that leans toward 14DD changes the game too, and may also bring about options there were not before; the difference lies not in whether one changes the game but in where the change heads. When the end headed in the 15DD direction changes the game, what it heads toward is making the second question one that both ends can ask; cooperation here is the setting in which cultivation may happen, not the source of cultivation. In changing the game only direction is reckoned: the game can only invite, and changing the game can only invite too; the share put into the game for the sake of changing it is not written as an account the other must repay. Deciding his direction for him on the ground that it is "better for him," or making him unable to refuse, revise or leave, is the shape of the direction of colonization; putting forward a proposal the other can still decline, change, or stay out of is not among these. Of the share handed over, what is to be checked is whether it has obtained a position that the end that handed it over cannot revoke at will: withdrawing it when the other does not act as this end expects can be exactly hanging the other's share back in one's own hands. Speech is also a step that changes the game: the paths actually open and the paths the other takes to be open are read apart, and truth or falsity does not on its own certify direction. After the game is changed, whether to avow is the other's; once avowed, it is read by the papers of Cooperation Theory; nothing is inferred backward in either direction. The end that changes the game can reckon his own gains, and can be glad for the other; what is to be checked is the step of taking the reckoned gains to declare an avowal on the other's behalf, or occupying the other's share on the strength of one's giving. Not every game is changed into cooperation: competition can be just competition, and when the two ends come together, the third person is checked separately. This paper interfaces with Follett's integration, Pettit's non-domination, Sen's self-imposed constraints, Schelling's commitment, and Brandenburger and Stuart's biform games; for Axelrod it points back to Paper 1. It does not prove that changing the game in the 15DD direction pays, nor that it does not pay; it does not prove that people ought to cooperate; it does not prove game theory wrong. 这一篇读合作的门槛。有两个问句:我要 A,你要 B,我怎样尽量拿到接近 A 的结果;我要 A,你要 B,我们怎样一起做 C。这一篇说的博弈,只指有一端在这件事上只问第一问的局面,不替博弈论这门学问下定义。这个系列读的合作,要两人都认下这件事、实际参与;这一篇读认下以前的那一段:一端在 15DD 方向上,另一端在这件事上只问第一问,共同的事还没有认下。它不问合作的一般来源。15DD、14DD 在这里是姿态,不是标签;操作层上,这一端既不认证对方在 15DD 方向上,也不认证对方只问第一问,他认得的只是两样:还没有共同认下的事,对方的言行还没有显出第二问。 在 15DD 方向上的那一端两问都问。他在局内照算,守住自己那一份;给定局内的所得,不穷尽他改局所朝的方向。改局,是改动一件事里的条件:谁能做什么,谁的那一份挂在谁手上,什么话说得出,什么事认得下。偏 14DD 的一端也改局,也可能改出原来没有的选项;分别不在改不改局,在改局朝哪里去。在 15DD 方向上的那一端改局,所朝的是使第二问对两端都问得出;合作在这里是涵育可能发生的场合,不是涵育的来源。改局只算方向:局只能邀请,改局也只能邀请;为改局在局内放进去的那一份,不写成对方须偿还的账。 以「对他更好」为由替他定下方向,或使他不能拒绝、修订与离开,是殖民方向的样子;提出一个对方仍能不要、能改、能不参与的方案,不在此列。交出去的那一份,要核它是否已取得不由交出的一端随时撤回的位置:对方不照这一端的期待行事时把它收回,可以正把他那一份重新挂回自己手里。话也是一步改局:实际能走的路与对方以为能走的路分开读,真假不独自认证方向。改局以后,认不认下是对方的;认下以后照合作论各篇读;两向都不倒推。改局的一端可以算自己的所得,可以为对方高兴;要核的是拿算出的所得替对方认下、据付出占住对方那一份的那一步。不是每一局都改成合作:竞争可以就是竞争,两端合到一处时另核第三人。 这一篇与福列特的整合、佩蒂特的非支配、森的自加约束、谢林的承诺、布兰登伯格与斯图尔特的双形博弈对接;阿克塞尔罗德回指第一篇。它不证明在 15DD 方向上改局划算,也不证明不划算;不证明人应当合作;不证明博弈论错了。