ABSTRACT
Carl Gustav Hempel suggests the specification of an artificial empiricist language with restricted vocabulary and explicit rules and, then, the characterization of significance as translatability into such language. Control over the vocabulary of such a particular base language would eliminate nonsensical terms such as 'glubbiflcation', while control over its syntax would eliminate nonsensical combinations of remaining signs. Translatability involves, then, a matching of "predefinitional" with "postdefinitional" meanings. Extensional isomorphism is, thus, the weakest requirement proposed, and must itself be satisfied if any of the others is satisfied. What extensional isomorphism requires of a legitimate definition is that it should enable truth-value preserving translations of certain statements in which one are particularly interested. Intuitive grammaticalness is, as Noam Chomsky has suggested, a matter of degree seems obviously excludable by syntactic restrictions, and, though less ungrammatical, seems also excludable through syntactic means.
