CONCLUSIVENESS OF NATURAL LANGUAGES AND RECOGNITION OF IMAGES
Zbigniew Marcin Wójcik · Cybernetics & Systems · 1983
General assumption of the paper is that perception results in some internal structures which may be converted into symbols being raphs and natural languages. Conclusiveness (decidability) of any natural language is investigated by analogy to the formal theories: the real world being perceived is treated as metatheory whereas events, images, and scenes are considered as a metalanguage, and all symbols representing the real world (the symbols include all natural languages) are viewed as a theory. Any natural language comprises descriptions (representations) of properties, features, relations, and expressions relating to the physical laws of the real world. Any subset of the theory (e.g., a natural language) forms a language, as in formal theories. The conclusiveness is investigated using recognition processes and one-one correspondence between expressions of a natural language and graphs representing events. The graphs, as conceived in psycholinguistics, are obtained as a result of perception processes. It is possible to generate ' and process the graphs automatically, using computers and then to convert the resulting graphs into expressions of a natural language. Correctness and conclusiveness of the graphs and sentences are investigated using the fundamental condition for events representation processes. Some consequences of the conclusiveness are discussed, e.g., undecidability of arithmetic, human brain assymetry, correctness of statistical calculations and operations research. It is suggested that the group theory should be imposed on mathematical models of any real system. Proof of the fundamental condition is also presented.