Notes from the initiative.
Field Notes are working articles from the Geometry Intelligence initiative. They set out positions, boundaries and reasoning at the field level. This section is editable: entries may be revised as the initiative\'s public account develops.
How to read this section.
Notes carry a status marker — Draft, In review, or Published. All entries listed below are currently in preparation. When a note is published, it will carry a stable URL under this section, and any subsequent revision will be recorded in the change register on the Timeline page.
Notes.
Why "Geometry Intelligence" and not "geometric intelligence".
A note on why the initiative uses the noun form of the field name in prose, and why the domain string is not treated as licence to substitute the adjective form.
Evidence, entities, relationships, constraints.
The four organising structures used at the field level, and why each is treated as first-class rather than as a special case of another.
Domain-general in structure, domain-specific in validation.
A statement of the principle that a shared reasoning architecture does not, and must not, absorb the validation rules of the domain to which it is applied.
What provenance is doing.
The role of provenance as a first-class attribute of evidence, rather than as metadata attached after the fact.
The operator gap.
The distance between what a system is capable of and what it is authorised to do — and why closing that gap by hand-waving is a category error.
Sovereign context in practice.
How jurisdictions, authoritative bodies and public registers enter the reasoning frame, and what happens when they are treated as noise.
The evidence economy.
The observation that better-structured evidence often outperforms more computation, and what this implies for how resources should be allocated in an intelligence system.
Structured discovery over free-text similarity.
Why traversal of an explicit structure is a different reasoning surface from ranking by textual similarity, and where the two are complementary.
Machine-readable authority.
The requirement that authoritative sources be parseable directly by a machine, and what has to change on the source side for that to be true.
What this site is not.
A written statement of the boundaries this initiative places on itself — no product page, no live system, no claim of endorsement absent independent evidence.
Corrections and correspondence.
Substantive corrections to a Field Note — factual errors, misattributions, or misstatements of an adjacent field — should be routed through the Briefing form with mandate type Factual correction. Corrections accepted after review are logged in the change register on the Timeline page.