The questions behind it.

Linguistics & sociolinguistics

Language isn't just what grammar books say it is. It's what people actually do with it — slang, code-switching, tone, and everything a sentence means because of who said it, to who, and where. If a system is going to understand meaning, it has to deal with the social side of language, not just the syntax. For me this is the starting point. Everything else builds on it.

Not communicating is still communicating. The second someone is there — seen, or just able to act — everyone around them responds differently. They respond to that person existing, and to what that person could have done but didn't. Silence, absence and holding back all carry meaning, so a model of communication has to deal with what goes unsaid as much as with what's said.

Mathematics of meaning

If meaning has structure, that structure should follow rules. This is about figuring out what those rules are: what it means for two meanings to be close, and which changes keep the sense and which break it. The goal is to be able to show that two sentences mean the same thing, not just say it.

Causality

Correlation is easy to find. Causes are hard to pin down. This is about representing what actually causes what — in the situations language describes, and in a model's own outputs. If a system can't tell cause from coincidence, I wouldn't trust it with decisions.

Dynamic causal graphs

Meaning doesn't sit still. Causes link up, get stronger, fade and rewire as the context changes. This is about networks of cause and effect that change over time — so you can follow not just what caused what, but how those links shift as a conversation or situation plays out.

Social cause and effect

Ideas move through groups the way weather moves through air. This is about how beliefs spread, change and settle in groups of people — the cause-and-effect side of influence, persuasion and change. If language tech is going to sit between people, it needs some model of what communication does to groups.

Social quantum effects

Some group behaviour looks a lot like measurement in physics: an opinion can stay unsettled until someone asks, and asking changes the answer. This tests that comparison carefully — where the actual math of quantum theory fits social behaviour, and where it's only a metaphor.

Semantic information theory

Classical information theory counts bits. It doesn't care what the bits mean. This is about the missing piece: how much a message changes what the person receiving it understands, and how much information is in a metaphor, a hint, or a pause. Until meaning can be measured, it's hard to know what you're actually optimizing.

Nosology of human behavior

A dimensional model of human behaviour I've been developing, in the psychoanalytic tradition that goes back to Freud and Jung. I built it for people first — to describe and make sense of patterns in how people think, feel and act. It's here because any model of communication has to account for the people doing the communicating, and this work may turn out to be useful for the project.