I’ve been working on a project called *Mimesis Protocol* — a structured semantic representation model for natural language. The basic idea is that an utterance

I’ve been working on a project called *Mimesis Protocol* — a structured semantic representation model for natural language. The basic idea is that an utterance often has a gap between its surface form and its actual function in context. For example, a sentence like “Do you know what time it is?” may look like a question, but pragmatically function as criticism, pressure, or a prompt for reflection. Mimesis tries to represent an utterance not just as text, but as a structured object with separate layers for: - `core`: semantic meaning and emotion - `will`: communicative purpose and influence - `flow`: discourse framing / progression - `exp`: surface realization I’ve put together: - an English/Korean whitepaper - a schema spec - a working interactive demo Still early, but I’m interested in whether something like this is useful as an intermediate representation layer for LLM systems, agent workflows, or discourse analysis. Repo: https://github.com/hangil131/Mimesis-Protocol https://github.com/hangil131/Mimesis-Protocol Demo: https://chatgpt.com/g/g-681be5f72cc4819191c2b12c9d89b336-mimesis-protocol-demo https://chatgpt.com/g/g-681be5f72cc4819191c2b12c9d89b336-mim...