Linguistic profiling of text genres: adventure stories vs. textbooks

Abstract

In this article we test the hypothesis that genre-inherent quantitative linguistic parameters can be reduced to a list of few provided with strictly defined ranges of value. The current research as part of a large project is aimed at contrastive analysis of textbooks on History and Social Studies, and adventure stories. Using RuLingva[1], we identified 18 genre variables, computed their frequencies and employed the Kruskal-Wallis H Test to evaluate the differences significance. The results suggest that the list of the most indicative parameters include sentence length, noun genitive case, future tense, ratio of verbs to nouns, provisionally called ‘narrativity’, and frequency. All the identified parameters have statistically significant differences and three of them (sentence length, genitive noun, and “narrativity”) are implemented in non-overlapping “genre-inherent” ranges of values attributed to (a) History and Social studies textbooks and (b) adventure stories. With the view that the target audience of adventure stories are not expected to demonstrate high levels of professional training but logical skills, we argue that the relatively stable readability of adventure stories, i.e. FKGL = 8-9, can also be attributed to the genre-inherent characteristics. Our results certify that incorporating text complexity indices improve the classification performance of genre quantitative analysis. We also offer our views on linguistic and statistical aspects of the proposed approach for future studies. Further research is needed to see how the same parameters are exploited in texts of other genres and subject domains.