Theoretical perspectives on reading comprehension: A narrative review of classical and contemporary models of reading comprehension

Reading comprehension is a complex cognitive process shaped by linguistic, cognitive, motivational, and contextual factors, yet theoretical accounts of it remain fragmented across separate literatures. Prior reviews have generally treated component-skill models, process-level theories, and multiple-text frameworks in isolation, leaving unclear how these levels of explanation relate to one another and how they hold up under digital and AI-mediated reading conditions. This narrative, theory-focused review addresses that gap by tracing the development of reading comprehension theory from top-down, bottom-up, and interactive models to contemporary integrative frameworks, including the Active View of Reading, the Direct and Indirect Effects Model of reading (DIER), and dynamic systems perspectives. It then synthesizes process-level accounts of meaning construction, such as memory activation, inference generation, coherence building, and situation-model construction, and examines how these operate across single- and multiple-text comprehension, with particular attention to source evaluation and discrepancy management. The review proposes an explicit four-level conceptual framework, organizing component skills, moment-to-moment comprehension processes, multiple-text integration, and medium- and context-related moderators within a single explanatory architecture, and extends this framework to hypertext, multimodal, and AI-mediated reading environments. The synthesis indicates that comprehension is best understood as a dynamic, multi-layered process arising from interactions among lower- and higher-level cognitive operations, prior knowledge, reader goals, and contextual conditions. By connecting classical and contemporary perspectives, the review clarifies points of convergence and unresolved tension among established models and outlines directions for future research on comprehension in increasingly complex information environments.