A Theory of Language Structure

Abstract

This paper presents a theory of the structure and information of sentenes, which consists of three fundamental relations: the ordered entry of words into the making of a sentence; inequalities of likelihood of occurrence of a particular nth entering word in a sentence in respect to the choices of words entering immediately before the nth; and the reduction in shape of certain words which have, upon their entry into a particular sentence, exceptionally high likelihood of occurrence, i.e. low amount of information. The first two relations determine the structure of a base subset of sentences which have a transparent grammatical structure, and contain little or no grammatical restrictions, and carry all the substantive information carried in any sentence of the language. The last relation is in general optional, and does not materially change the information in a sentence. It introduces a secondary grammatical complexity into sentences, and introduces restrictions, subsets, and grammatical rules into the language. From the base sentences all the remaining sentences of the language can be obtained by a direct application of the stated set of reductions, in a manner suitable to the constructing and, with certain short-cuts, to analyzing of sentences.¹