Metadiscourse in Reading and Inferential Comprehension: Evidence from L1 Chinese Readers and Implications for L2 Pedagogy

This study examined how metadiscourse marking relates to reading-stage processing and subsequent inferential comprehension among L1 Chinese readers, and considered implications for L2 pedagogy. Forty-two participants completed an online cumulative self-paced reading experiment involving four Chinese texts presented in metadiscourse-present and metadiscourse-reduced versions. Whole-text reading time, target-sentence and spillover reading times, perceived difficulty, inferential-comprehension response time, and accuracy were analyzed. The findings showed different patterns across reading and subsequent inferential-comprehension measures. Neither length-adjusted whole-text reading time nor target-sentence and strict spillover reading times showed reliable condition differences. Perceived difficulty was somewhat lower with metadiscourse, although this result was tentative and model-dependent. The clearest condition difference emerged in the subsequent inferential-comprehension task: participants responded approximately 12% faster following metadiscourse-present texts, whereas accuracy showed no reliable difference and was near ceiling. Because the full text remained available during question answering, the response-time difference may reflect information search, retrieval, or rereading, as well as inferential processing. These findings suggest that metadiscursive cues may help readers access and use discourse-level information in subsequent comprehension tasks. We further propose treating metadiscourse as a pragmatic metaconcept in L2 instruction to help learners systematically understand its functions in discourse organization, stance, and engagement. This may enhance learners’ metapragmatic awareness, enabling them to recognize and use metadiscursive resources more effectively as discursive scaffolding for comprehension, inference, and interaction.