Abstract: The development of game theory can be regarded as a process of continuous improvement and refinement of the solution concept, which imposes  consistency requirements on behaviors and beliefs. Harsanyis contribution is converting a game of incomplete information to a complete but imperfect information game by “Harsanyi transformation”, and proposing the concept of Bayesian Nash equilibrium, which provided an unified analytical framework to cope with information problem and laid the foundation for the economics of information. This paper argues that the Bayesian Nash Equilibrium defined on the Harsanyi game needs to take the common prior assumption as the premise, which is rather demanding. Therefore, any analysis based on the Harsanyi game must make sure the common prior assumptions are satisfied.

Key words: equilibrium solution , highorder belief , common prior , incomplete information game

摘要： 博弈理论的发展过程可以看成是对其均衡解的概念不断完善和精炼的过程，而解的概念反映的是行为和信念的一致性。在不完全信息条件下，在决策时决策者不但要考虑自己的信息，还要考虑别人的信息，别人对自己信息的信息等等高阶信念，此时如何定义信念一致性的问题长期困扰着学者。海萨尼的贡献在于通过构造“海萨尼转换”将不完全信息博弈转换为完全但不完美信息博弈，并基于此提出了贝叶斯纳什均衡的概念，为经济学建立了分析信息问题的博弈框架，并奠定了整个信息经济学的基础。本文认为，海萨尼意义上的贝叶斯纳什均衡的一致性需要以共同先验假设为前提，但是此假设在现实中很难满足。因此，在使用海萨尼博弈分析实际经济问题时，必须首先检验共同先验假设是否可以被满足。

关键词: 均衡解 , 高阶信念 , 共同先验 , 不完全信息博弈