• (The)Study of the Eastern Classic
• /
• no.35
• /
• pp.275-295
• /
• 2009

The aim of this paper is to reject Graham's interpretation of bi (必) and xianzhi (先知) of Later Mohists' Mojing ("墨經") as logical necessity and a priori knowledge respectively. Graham's interpretations of them are based on his beliefs that Mojing distinguishes lun (論), the art of description from bian (辯), the art of inference in the Mohist disciplines and that the latter art should be seen as such a rigorous proof as Euclidean geometry even though it is not a Western formal logic. His beliefs also start from his distinguishing 'knowledge of names' from 'knowledge of conjunction of names and objects' according to the objects of knowledge. In my reading, the art of description and the art of inference, however, can't be sharply distinguished each other in Mojing and bi and xianzhi should be taken as suggesting both a normative necessity and an empirical necessity. A normative necessity is derived from 'normative theory of definition' which comes form the theory of rectification of names in China. The normative theory of definition, unlike the descriptive theory of definition, defines terms normatively rather than descriptively. For example, although such a definition of father, 'father is beneficient', has the form of being descriptive, but it actually is prescriptive and therefore means 'father should be beneficient'. Through this normative theory of definition, empirical knowledge, as long as it is a knowledge, is seen as necessary and so can't be wrong. To conclude, for Mohists an empirical knowledge is always a basis of an inferential knowledge or a priori knowledge, so Mohists' a priori knowledge is not really a fundamental knowledge and its necessity therefore is nothing but both a normative necessity and an empirical necessity.