Every time you ask AI to summarize something, it gives you back something shorter. But it also gives you back something forgettable. The facts are there. The meaning is gone. Cognitive science explains why: humans do not store information as data points. They store it as stories. Strip the story out, and the mind has nothing to hold onto. This paper introduces Narrative Glue Components (NGC), a framework that treats text compression the way MP3 treats audio. MP3 does not discard random sound. It removes only what the human ear cannot perceive. NGC does the same for text: it removes scaffolding while preserving the narrative beats that the human mind uses to encode, retain, and recall information. The framework draws on decades of narrative theory, from Campbell and Propp to McKee and Vonnegut, and grounds every decision in modern cognitive science research on how memory, narrative, and emotion interact. It classifies text into 25 narrative genres, applies approximately 1,140 genre-specific compression laws, and operates through a five-step pipeline that identifies what can be safely removed and what must survive. Across 33 test passages, NGC achieved 50 to 67 percent compression while preserving 100 percent of emotional beats, something even AI systems explicitly told to protect narrative could not match. But NGC is not only a compression tool. Run the pipeline in reverse and it becomes a creation engine, generating narrative-rich content from raw facts. In a world where AI-generated content is increasingly polished but emotionally flat, this framework offers a way to keep information human: structured for how minds actually work, built to be remembered, and designed to preserve the creative capacity that makes us more than machines that process text. This is Version 1.0. The theoretical foundations, the pipeline, the taxonomy, and the evidence are here. What comes next is validation, expansion, and application across education, healthcare, policy, and beyond.
Akhil Bhima (Tue,) studied this question.