Universal Semantic Code (USC) as a formal semantic language for knowledge representation and inference was shaped at early 80th.
It was time of a first wave of interest to Artificial Intelligence (AI). Many leading researchers in the field: R.Schank, T.Winograd, M.Minsky, D.Lenat and others proposed different approaches to the fundamental task of AI - Knowledge Representation (KR). It is impossible to build human-like AI system without effective KR.
Meantime, a problem has raised. All of the approaches did not give a solution how to represent even common knowledge in a universal way. Statistical methods were not well developed that time, but even if it would be, like today, statistics cannot replace human understanding what is common knowledge.
Actually, common knowledge is a model of the world. Everybody of us has the model overlapping more or less with others. If more overlapped, then more compressed language used and less special efforts to understand each other or for KR needed. If less overlapped, then more special efforts needed. Semantic canonization of natural language is necessary for uncompressing or decoding knowledge kept in the mind in some universal semantic code because it is not depended on the particular natural language.
USC does not resolve all KR questions too, however, it provides formal support to represent knowledge and define inference with axiomatic rules. Author of the USC theory Prof. V.V.Martynov had proposed an approach which is very different from the known and promising in the building of the AI systems.