Digital Humanities Based Computational Analysis of Linguistic Evolution and Cultural Narratives in Historical Texts

Abstract

The swift progress in computational techniques has provided ample possibilities for investigating the evolution of language and changing discourse patterns in large corpora of texts. The paper describes a computational framework or investigating longitudinal register shifts and formal discourse organization in American English using the Corpus of Contemporary American English (COCA) from 1990–2019. Lexical, discourse, informational, temporal, and statistical analyses are combined into the framework to detect systematic change from one period to another. Three periods, 1990–1999, 2000–2009, and 2010–2019, were identified in the corpus and analyzed on the level of the document in terms of linguistic indicators calculated through word-count weighting. It was found that character types dropped by 8.48%, whereas expert vocabulary increased by 15.20%. Increases were noted in Information Comparison (+12.88%) and Information Quantity (+15.24%). Larger increases were also found in discourse organization measures, such as Methods-Results-Discussion (+28.82%), Purpose-Plan (+20.26%), and Reasoning (+21.20%). A statistical analysis revealed significant correlations in the time of several variables, including Reasoning (r = 0.113, p < 0.001) and Character Types (r = −0.082, p < 0.001). The operationalized Linguistic–Discourse Evolution Index (LDEI) rose from 0.115 to 0.134 (a 16.4% increase). These results demonstrate systematic evolution toward greater lexical specialization, informational density, and structured academic discourse.