Abstract

It is argued that Davidson's basic objection to the Building Block Method in semantics is neither that it gives the wrong explanation of how a first language is learned nor that it assigns a meaning to Single words prior to interpreting a whole language. The arguments against Fregean concepts and truth-values as the references of predicates and sentences are found to be equally superficial as the arguments against a primitive notion reference defmed in causal terms.Davidson's basic objection turns out to be that thoughts do not have a deep-structure which can be revealed by a correct analysis. His constraints on a theory of meaning do not allow for a distinction, as suggested by Dummett, between analysis and decomposition of thoughts. This forces us to a very general decision about how to do philosophy. As a non-reductivist I think it makes sense to assume a basic thought-structure. From this perspective the use of buildingblocks in semantics is vindicated.