The place where Primitive has to explain itself.

We develop and test ideas about representation, reasoning, relationships, information and emergence. Some begin as philosophy. Some become formal models. A smaller number become software. A smaller number again survive benchmarking.

This page exists so those stages never get confused with one another. Ideas should be ambitious. Evidence should be conservative.


Every research artifact on this site carries an epistemic status. It's not decoration — it's the contract.

BELIEF

A philosophical or foundational position. It may motivate research. It is not presented as demonstrated.

OPEN QUESTION

We don't currently know. It's recorded because the answer could change the framework.

HYPOTHESIS

Specific enough that an argument or experiment could support or contradict it.

FORMALISED

Precise enough — mathematically, computationally, procedurally — to inspect and attack. Formalisation is not correctness.

EXPERIMENTAL

Currently being tested. Treat results as provisional.

REPLICATED

Survived predefined repeat testing under stated conditions. Not universal validity.

INDEPENDENTLY REPLICATED

Reproduced by someone who doesn't work here. The standard we want more of.

FALSIFIED

The prediction didn't survive. It stays in the archive.

SUPERSEDED

Replaced by a newer formulation. The old one stays reachable.

Primitive works across four connected layers.

01

Philosophical

What do we assume exists? What is state? What is difference? What draws a boundary? What is an observer? At what point does a pattern deserve to be treated as an object?

02

Representational

How do we encode those ideas? Coordinates, states, graphs, fields, relations, context, uncertainty, temporal position, scale, observer-relative boundaries.

03

Transformational

How can a representation change? What operators alter perspective, scale, boundary, time, semantic state, relationship? Which transformations preserve information — and which destroy it?

04

Experimental

Does any of it improve an actual task — retrieval, reasoning, prediction, planning, coordination, compression, decision support? If the answer is no, we need to know.