STATISTICS OF LOW-FREQUENCY KERNEL (SUBORDINATING) MODELS OF THE VERBAL WORD-GROUPS IN THE TEXT CORPUS “RADIO ELECTRONICS”

Abstract

The paper deals with one of the most widely used types of verbal word combinations, namely kernel (subordinating) models of the verbal word-groups functioning in the “Radio electronics” text corpus, which refers to the technical and scientific discourse. The text corpus was compiled on the basis of scientific and technical journals “Radio electronics” published in the USA by the method of a continuous sampling. In order to select this kind of models a theoretical approach describing the problems of combinatorics and valance is considered. This permits to take into account both contact and distance elements, which are necessarily connected with the verb by grammatical (subordinating) bond and form a word-group with it. For compilation of the inventory of kernel models of the subordinating verbal word-groups 10698 sentences are subjected to the distributive analysis. The models occurring with the low frequency (below 180) are exclusively chosen. The total amount of low frequency model is 41 units. Except the distributive methods the statistical ones are applied in the research. The quantitative analysis of kernel models used in the text corpus with a low frequency has shown that the number of different models is much bigger as compared to the same units with a high frequency of usage – 77, 4% and 22, 6%, respectively. But the total frequencies comparison goes in the favour of high frequency models – 0, 4% to 99, 6%. Besides quantitative characteristics the simplicity/complexity parameter of the model structures are analysed. In studying the high frequency kernel models the direct dependence of frequency of usage on the model structure was observed – the simpler the structure is the higher the frequency is. But the low frequency kernel models demonstrate more complex and contradictory functioning as regard to this parameter. In most cases the tendency to indirect dependence is seen quite distinctly, i.e. the simpler the structure is the lower the frequency is, and the chaotic distribution of frequencies occurs in the frequently highest among low-frequency models.