How Prompting Shapes Meaning: Discursive Reframing and Contextual Narrowing in AI Interpretations of Dede Korkut Narratives

This study examines how generative AI organizes interpretation in Dede Korkut narratives across different prompt conditions. Rather than evaluating the correctness of AI-generated responses, the analysis focuses on how interpretation is shaped through discursive processes within specific prompting frameworks. The dataset consists of two narratives-“Dirse Han Oğlu Boğaç Han” and “Kam Pürenin Oğlu Bamsı Beyrek”-each analyzed using three prompt types: low-guidance, structured analytical, and conceptually framed prompts, with an additional user condition included as a limited robustness check. Across outputs, recurring patterns emerge, including semantic flattening and contextual narrowing. Culturally embedded symbolic elements are translated into generalized analytical categories, and narratives are treated as bounded textual units detached from their oral and performative context. A comparable shift appears in the representation of narrative figures, moving from flexible, typified roles toward more stable and individualized forms. These tendencies remain consistent across prompting conditions and user contexts, indicating that variation in interpretation is shaped less by user differences than by the interaction between prompt design and model tendencies. When disciplinary guidance is not specified, outputs tend to align with generalized sociological frameworks. The analysis suggests that AI-generated interpretations often translate culturally embedded meanings into more uniform explanatory forms, at the expense of contextual specificity. These patterns carry implications for literature and language learning contexts, where AI-supported interpretation increasingly shapes meaning-making practices and points to the need for critical AI literacy.