Semantic Variation in Digital Academic Texts: A Corpus-Based Study of Technical Terminology

Abstract

Academic communication requires the use of technical jargon to communicate complex concepts, which encourages teamwork, and facilitates interdisciplinary interaction. The differences in the meaning of the words in different fields may pose a hindrance to effective communication. This paper intends to review the semantic difference in the use of the technical terminology in engineering, computer science and social science with interest in the change in the meanings of the common words and the implication of these meanings to academic communication. Semantic overlap was measured using a corpus-based approach, frequency analysis, collocational analysis, and Jaccard similarity coefficient. The information was gathered in three subjects, including peer-reviewed journals, conference papers, and textbooks. The corpus includes 2.5 million tokens of 150 texts in the years 2020-2025. The analysis determined that 67% out of the 50 leading terms revealed multiple meanings across the disciplines. Significant variation in their definition for the terms like “model” and “network” showed Jaccard similarity values of 0.23 and 0.31 between Engineering and Computer Science, 0.18 and 0.27 between Computer Science and Social Sciences, respectively. Disciplinary context explained 72% of the variance in the term definition. Improving academic communication requires technical terminology standardization. Standardizing terminology across academic fields requires knowledge of context-dependent terms in meaning and supports framework development.