SHANNON AND WEAVER

Abstract

A case is presented for separating Shannon's (1949) paper on information theory from Weaver's introduction, which is shown to contain distortions, as well as proofs by coincidence and homonym. Shannon's mathematical tools and methods are distinguished from his theory, which consists of 23 theorems setting forth the conditions for maximum efficiency in electromechanical signal transmission. Attempts to apply Shannon's theory to our field are reviewed, along with previous critiques, and it is recommended that future uses of Shannon's theory adopt a more methodologically rigorous approach. In particular, it is argued that Shannon's assumptions must be shown to hold before his theorems can be successfully applied.