What Do Author Self-Citations Do? A Citation Context Analysis of their Semantic Functions

Author self-citation is a controversial phenomenon, with some scholars viewing it as a legitimate scholarly practice and others attributing it to self-serving biases (Tagliacozzo, 1977). Citation context analysis offers a complementary perspective by examining the textual environment surrounding a reference to understand how and why a work is cited (Teufel et al., 2006). Using this approach, recent studies have shown that national self-citations—citations to publications originating from the same country—play important roles in the reuse of methods, datasets, and findings (Chen et al., 2022). However, citation context analysis has rarely been applied to author self-citations. Consequently, little is known about their semantic functions—that is, how and for what purposes authors cite their own work—including methodological continuity, data reuse, result extension, comparison, and conceptual linkage, or about the linguistic features through which these functions are expressed. To address this gap, we conduct a large-scale analysis of the semantic patterns of author self-citations. Leveraging the accessibility of full-text articles and the openly available JATS XML format for systematic citation context extraction, we use research articles published in PLOS journals between 2012 and 2025 as our experimental dataset. We develop a reproducible pipeline to identify author self-citations and classify their citation contexts according to the ten-category citation-function schema proposed by Agarwal et al. (2015). Semi-automatic annotation is employed to construct an annotated corpus, which is subsequently analyzed to examine the distribution of semantic functions across article sections, publication years, disciplines, citation frequency, and linguistic characteristics. This study provides a scalable methodology and the first large-scale characterization of the semantic functions of author self-citations, offering new insights into how researchers use self-citations to support cumulative knowledge-building.