The Problem of Informational Synonymity

Abstract

In his manifesto ‘On the construction of a machine language for an information machine’, the Soviet pioneer of information science V.A. Uspenskij [179] specified a number of requirements which should be met by the information language to be used in any computerised information system. One of these requirements consists in the formalisation, and hence the possibility of a machine solution, of semantic identification, i.e. of a decision that two differently expressed or recorded items of data are identical in terms of their meaning. This chapter is an attempt to meet that requirement, the concept of ‘semantic identity’ being explicated by the means afforded by the theory of semantic information and on the basis of the concept of ‘transmitted information’. It is in this light that we introduce the concept of ‘informational synonymity’. Since previous attempts to solve the problem of semantic identification have usually been based on the Leibnizian principle of identity, we shall be indicating some of the difficulties and limitations inherent to methods and criteria based on the Leibnizian principle. In principle, even the concept of ‘informational synonymity’ keeps to the requirements which stem from the Leibnizian principle, but in a modified form.