摘要
刘易斯把因果定义为实际事件之间因果依赖的传递闭包,可以解决困扰规则性理论的诸多问题,但是也确证了因果是传递的。在随后的讨论中因果传递性的反例不断涌现,从而间接地否定了刘易斯因果定义的合理性。希区柯克提出了结构方程模型的因果理论,他也认同通过反事实来理解因果,但和刘易斯不同,他引入了结构方程的工具来清晰地刻画具体的情形,并通过"活跃路径"来定义因果关系,他指出我们可以在不付出因果是传递性这个代价的情况之下,也可以合理地获得刘易斯因果定义的那些好处,而且还能够解决困扰他的定义的那些反例。笔者试图证明所有用以反驳因果传递性的反例都存在一种潜在的概念偷换,并不构成对刘易斯因果理论的反驳,并给出反例说明希区柯克所自信的那些结构因果模型的优点,比如清晰地刻画具体的情形,也有其界限,在具体的问题讨论中也有其无法克服的"言不尽意"处。
In Lewis,he defined causation as a transitive closure of causal dependence between actual events,which can solve many problems that plague regularity theory,but also confirm that causation is transitive.In the ensuing discussion,counterexamples of causal transitivity emerged,indirectly negating the rationality of Lewis’definition.Hitchcock proposed the causal theory of structural equation modeling,and he also agreed to understand causation through counterfactuals,but unlike Lewis,he introduced the structural equation tool is to clearly describe the specific situation and define the causal relationship through the"active route".He pointed out that we can reasonably obtain the Lewis causal definition without paying the cost of causation’s transitivity and the counterexamples that plagues his definition.The author tries to prove that there is a potential concept stealing of all counterexamples used to refute causal transitivity.It does not constitute a rebuttal to Lewis causal theory,and gives a counterexample to illustrate the advantages of Hitchcock’s self-confidence structural causal models,such as clarity,which also boundaries.In the discussion of specific issues,there are also"unsatisfactory"places that it cannot overcome.
作者
吴小安
WU Xiaoan(Department of Philosophy,Tsinghua University,Beijing,100084)
出处
《自然辩证法通讯》
CSSCI
北大核心
2020年第8期1-9,共9页
Journal of Dialectics of Nature
关键词
实际因果
反事实
结构方程
传递性问题
Actual causation
Counterfactuals
Structural equation
Transitivity
