Statistically Exploring Semantic Accessibility Scale Based on English, French and Japanese Corpora: A Comparative Perspective

摘要

The inter-language studies on the textual semantic accessibility scale (SAS) are a new branch of the computational linguistics and the present paper tries to statistically probe into the SASes in English, French and Japanese literature works sampled from the corresponding corpora. Firstly, six control groups are formed by the equidistant texts extracted every 10 pages, 5 pages, 4 pages, 3 pages, 2 pages and 1 page respectively. Secondly, some comparisons are made to verify if there is really a correlation between the sampling ratio (SR) and the SAS in a monolingual language field. Thirdly, the correlation between the SR and SAS is analyzed in multilingual domains. On basis of the afore-mentioned procedures, we find that the difference of sampling space results in different SRs, but the values of the SASes are independent of the SRs. And we also testify that the simple texts boast of a higher value of SAS.