The Metaphor Integrity Framework: Metaphor Collapse, Symbolic Cognition, and AI-Driven Literalism

Metaphor is not ornamental language but a foundational mechanism of human cognition. Research in cognitive linguistics and philosophy of mind demonstrates that metaphor structures conceptual thought, emotional understanding, identity formation, and social connection. As artificial intelligence systems increasingly mediate communication, interpretation, and meaning-making, metaphor is being systematically flattened into explicit, literal language. This paper introduces the Metaphor Integrity Framework (MIF), a theoretical model that treats metaphor as cognitive infrastructure rather than stylistic expression. The framework identifies metaphor collapse as a form of premature semantic closure and examines how contemporary AI systems accelerate this collapse through optimization for explicitness, speed, and disambiguation. Drawing on cognitive science, symbolic anthropology, media theory, and AI ethics, this work argues that metaphor erosion represents an under-recognized cognitive and cultural risk with consequences for empathy, imagination, identity, and collective meaning-making. The framework provides a foundation for studying symbolic cognition in both historical and AI-mediated contexts without reducing meaning to belief or prediction. Keywords: symbolic cognition, metaphor, AI ethics, cognitive linguistics, meaning-making, interpretability, human–AI interaction.