Structural and Functional Differences of Lexical Bundles Between Hard Science and Soft Science Researchers: A Comparative Corpus-Driven Study

Lexical bundles, recurrent word combinations serving essential discourse functions, have gained prominence in the realm of academic writing. A novel perspective that bridges their functional significance and formulaic nature is promising for uncovering intricate features within these recurring language patterns. Drawing from the structural and functional taxonomy introduced by Biber et al. (1999) and Hyland (2008), this comprehensive study aims to find the existence of any significant difference in the frequency and use of lexical bundles from both structural and functional perspectives between research articles authored by individuals in soft and hard science disciplines. The corpus, rich in academic content, encompasses a total of 954,615 words, featuring 90 research articles in each sub-corpus. The investigation extends beyond mere structural classification to encompass functional analysis, unveiling insightful findings. The findings indicate while structural distinctions between authors in hard and soft sciences appear negligible, substantial variations emerge in the pragmatic deployment of lexical bundles. Authors in soft sciences exhibit a predilection for noun phrases combined with of-phrase fragments. In stark contrast, authors in hard sciences predominantly employ passive verb + prepositional phrase fragments. Additionally, the divergence in the functional classification of lexical bundles is noteworthy. In the realm of soft sciences, authors heavily emphasize the use of framing signals, underscoring the discursive significance of these elements. In contrast, hard science authors gravitate towards transition signals as the most frequently employed function of lexical bundles. These findings carry substantial implications for researchers, highlighting the importance of embracing lexical bundles as a fundamental aspect of scholarly writing within their specific domains.