Monolith: Visual Structure Grounding for Grammar-Free Language Learning

We present Monolith, a web-based tool that visualizes English sentence structure using color-coded nested blocks without any grammar terminology. Traditional language education relies on metalinguistic labels (e.g., “subject,” “relative pronoun”) to explain structure—a multi-step indirect reference process that conflicts with the cognitive strengths of visual-spatial learners. Drawing on Symbol Emergence Systems theory (Taniguchi, 2016), Monolith replaces this label-mediated approach with direct visual grounding: syntactic roles are mapped to colors, and hierarchical clause nesting is rendered as block containment. AI (Gemini) handles parsing; the learner simply experiences the structure. We argue that this draws inspiration from Symbol Emergence Systems theory to create perceptual scaffolding, and that embracing the non-idempotent nature of natural language—rather than forcing a single “correct” parse—leads to more authentic language understanding.