Research on the Philosophical Significance of Connectionism

我们将联结主义与经典计算主义进行比较，试图阐明联结主义的哲学意义。通过仔细研究连接论者对认知活动的解释中“分布式表征”概念的使用，服部认为，真正的“分布式表征”根本不是表征。Mino指出，连接主义和经典计算主义具有相同的表征主义，但两者在经典计算主义是否正确的问题上存在分歧。Kaneko认为动力系统方法不适用于解释认知活动。柴田有保留地表明，在未来，民间心理学可以被认知科学和神经生理学所取代。柏端康成捍卫了命题态度的真实性。他认为，在“没有与命题态度相对应的大脑状态”和“真正的命题态度不存在”这两种观点之间，存在着逻辑上的鸿沟。今天，他讨论了消失论（或消灭唯物主义）。根据他的说法，我们必须更仔细地研究“还原”的概念，以便评估消失理论家的主张。横山以过去的科学革命和类似的例子为例，声称我们必须从更广阔的视角来看待我们的世界。大泽认为，自然语言不仅仅是一种具有句法结构的对象。他主张图画形象具有一种句法结构和表征功能，并提出了这些形象所遵循的一种新的逻辑。

Comparing connectionism with classical computationalism, we attempted to make clear of the philosophical significance of connectionism. By scrutinizing the use of the notion of "distributed representation" in connectionists' explanations of cognitive activities, Hattori argued that really "distributed representations" were not representations at all. Mino showed that connectionism and classical computationalism shared representationalism, while the two positions differed as to whether classical computationalism was right or not. Kaneko argued that dynamical systems approach was not appropriate in explaining cognitive activities. Shibata showed, with some reservations, that in the future folk psychology could be replaced by cognitive science and neurophysioiogy. Kashiwabata defended the reality of propositional attitudes. He argued that there was a logical gap between the claim that there was no brain state corresponding to a propositional attitude and the claim that really propositional attitudes did not exist. Todayama discussed the disappearance theory (or eliminative materialism). According to him, we have to study the notion of "reduction" more carefully in order to evaluate disappearance theorists' claim. Referring to the examples found in past scientific revolutions and the like, Yokoyama claimed that we had to see connectionism in a more broader perspective of our world. Ohsawa claimed that natural language was not only one kind of object which had syntactic structure. He asserted that picture images had a kind of syntactic structure and representative functions, and proposed a new logic which these images followed.