The Linguistic Features of Research Article Summaries in Science: A Corpus-Based Genre Analysis

Abstract

Research Article Summaries (RASs), a structured abstract, remain underexplored in terms of their linguistic realization. This study investigates the distribution and rhetorical functions of five linguistic features in RASs: sentence types, verb tense and aspect, grammatical voice, first-person plural pronouns, and reporting words. A corpus of 115 RASs published in Science Special Issues between 2017 and 2020 was analyzed through manual, section-based coding of 2,488 sentences containing 63,019 words. The findings reveal that RASs employ linguistic features strategically across their four rhetorical sections. Complex sentences predominate (68.61%), particularly in the Results, whereas simpler structures are more frequent in the Conclusion, supporting clarity and accessibility. Present Simple is the most frequent tense, while Past Simple is strongly associated with the reporting of completed findings and Present Perfect with the continuing relevance of previous research. Active voice predominates (81.13%), indicating a contemporary preference for clarity and authorial agency, while the frequent use of we and complete absence of I demonstrate a controlled form of collective authorial presence. Reporting words also vary rhetorically, with assertive forms such as find, show, and reveal concentrated in the Results and more tentative forms such as suggest used in the Conclusion. These findings demonstrate that grammar in RASs functions as a rhetorical resource rather than a set of isolated structural choices. The study contributes to research on scientific discourse by providing a systematic explanation of RASs and offers practical implications for genre-based, corpus-informed EAP/ESP instruction aimed at supporting non-native English-speaking researchers in international publication.