A Study on the Dynamic Evolution of English as an International Language Variant Driven by Big Data

Emerging agri-environmental concepts (e.g., “carbon farming”, “regenerative viticulture”, “drought-smart soils”) increasingly spread through English as a lingua franca (ELF), yet their fine-grained diffusion pathways remain invisible to conventional environmental information systems. The authors integrate the 1.9-billion-word GloWbE corpus (20 regions), 3.5 TB month-over-month Twitter metadata, and OpenSubtitles (2000–2025) to build a cross-platform terminology-tracking framework. Coupling attention-LSTM time-series clustering with geographical graph-convolution networks, the model quantifies lexical innovation rates at the lexical, spelling, and syntactic levels. Geographical distance and community overlap jointly explain 71% of the variance in term diffusion; the proposed “Annual Terminological Information Gain” (ATIG = 0.004 bits word−1) detects neologisms 12.3% more accurately than baseline multi-scale predictors. Results support SDG 2 (zero hunger) and SDG 13 (climate action) by informing interoperable thesauri for FAO and CGIAR digital extension platforms.