Bridging the disciplinary divide: A corpus-based analysis of lexical density and its implications for ESP pedagogy in Saudi higher education

Abstract

The proposed research explores the lexical schism between Hard Sciences (STEM) and Soft Sciences (Humanities) within English-medium instruction, using an 824,560-token corpus. The of peer-reviewed research articles (2020-2025), the study employs a multi-dimensional analysis of five dimensions of linguistic behavior, including lexical saliency, formulaic sequences, epistemic stance, syntactic collocation, and an expert-novice comparative overlay, with statistical analysis via Log-Likelihood (LL) test (p < 0.01). Results show STEM discourse uses certainty, and with the help of high-frequency boosters and research-oriented lexical bundles (e.g., as indicated in the figure), the discourse tries to emphasize empirical objectivity. By contrast, the Humanities emphasize interpretive agency more, applying almost 3.5 times the number of hedging markers and participant-focused bundles (e.g., one might say) to bargain dialogic space. It is also found that there are substantial semantic drifts in common academic lexicon, with verbs like analyze having varying collocational patterns across different fields. The expert-novice overlay also determines an existing linguistic ceiling among EFL learners, characterized by excessive use of general academic words and inadequate application of disciplinary formulaic sequences. The findings respond to the one-size-fits-all concept of Academic English and suggest the need to shift to a new model, differentiated EAP (D-EAP). With the coordination of pedagogical practices and disciplinary epistemology, curriculum designers are better positioned to prepare learners to negotiate the subtle rhetorical requirements of their preferred intellectual realms.