Kasturirangan Rajesh
National Institute of Advanced Study, Indian Institute of Science Campus, Bangalore, India.
Prog Brain Res. 2008;168:105-14. doi: 10.1016/S0079-6123(07)68009-1.
Philosophers as well lay people often think of beliefs as psychological states with dubious epistemic properties. Beliefs are conceptualized as unregulated conceptual structures, for the most part hypothetical and often fanciful or deluded. Thinking and reasoning on the other hand are seen as rational activities regulated by rules and governed by norms. Computational modeling of the mind has focused on rule-governed behavior, ultimately trying to reduce them to rules of logic. What if thinking is less like reasoning and more like believing? I argue that the classical model of thought as rational is mistaken and that thinking is fundamentally constituted by believing. This new approach forces us to re-evaluate classical epistemic concepts like "truth", "justification" etc. Furthermore, if thinking is believing, then it is not clear how thoughts can be modeled computationally. We need new mathematical ideas to model thought, ideas that are quite different from traditional logic-based mathematical structures.
哲学家以及普通大众常常认为信念是具有可疑认知属性的心理状态。信念被概念化为不受约束的概念结构，在很大程度上是假设性的，且常常是虚幻或错误的。另一方面，思考和推理被视为受规则调节并受规范支配的理性活动。对心智的计算建模专注于受规则支配的行为，最终试图将其简化为逻辑规则。要是思考不像推理而更像信念会怎样呢？我认为，将思维视为理性的经典模型是错误的，思维从根本上是由信念构成的。这种新方法迫使我们重新评估诸如“真理”“正当性”等经典认知概念。此外，如果思考就是信念，那么就不清楚如何通过计算对思维进行建模。我们需要新的数学理念来对思维进行建模，这些理念与传统的基于逻辑的数学结构截然不同。
