Semantic Systems After 30 Years

Abstract

Semantic systems are sytems with an inherent semantics. An example would be systems showing intrinsic intentionality: if a system is genuinely intentional, it must be able to define its own meanings. Searle was a forerunner of the modern idea of semantic systems in his oft-cited “Chinese Room” paper in 1980. The current author has approached the problem from a different angle 30 years ago in his book Self-Modifying Systems (Pergamon), claiming that minds can define their own meanings by virtue of being “material”, in the sense of freely generating new properties on the fly, as do material objects. In the course of the process, meaning should arise because syntax (and therefore any computation that is entirely syntactical) does not fully describe the rich functioning of such systems, so goes the argument. While still attractive, the author has in the meantime abandoned the idea, for at least two reasons. One, fundamentals-seeking has itself been largely abandoned, because the study of questions like “is the mind computational”? has turned out to be non-productive. Two, brute force approaches such as ANN-based deep learning have been “unreasonably” successful, and this fact puts aside arguments about the limits of computational systems. The paper provides an overview of these developments.