Unraveling Tortured Phrases in the Scientific Literature through the Lens of Talbots

Abstract Scientific literature has witnessed an alarming upsurge in the utilization of tortured phrases within scholarly articles. These special phrases cast doubt on the credibility of research, irrespective of the underlying rationale. Therefore, we introduced Tal- bots, which facilitates the identification and categorization of tortured sentence clusters. Our investigation thoroughly assessed 8,792 academic publications, including books and articles, characterized by their deployment of odd and incomprehension language, such as information mining instead of data mining. We carefully analyzed a comprehensive dataset of 5,582 tortured phrases and identified 2,235 scientific literature containing them. Subsequently, we proceeded to systematically categorize these sentences of tortured phrases into six distinct groups, including grammatically incorrect structures (GIS), inappropriate word usage (IWU), semantic inconsistencies (SI), complex language constructs (CLC), unnecessary coined terms (UCT), and non-tortured phrases (NTP). This categorization significantly helps to enhance the understanding of the science community on this issue. By shedding light on the prevalence of tortured phrases, our study contributes to enhancing the quality and integrity of content in future scientific publishing.