Research · 20 Mar 2026

Thirty-six thousand qualia relations, and what they miss

The count is the easy part. Coverage is the number that decides whether the engine can answer you.

Relations
~36,000
Seeded
WordNet
Verified
Human
Coverage
< 1

Specifics

The numbers, and where they come from

Figures on this page, with the basis of each
QuantityValueBasis
Qualia relations~36,000Counted
WordNet-seeded61%Counted
Human-verified27%Counted
Model-generated, unverified12%Counted
WordNet-seeded61%Human-verified27%Generated, unverified12%
Provenance of the relation set. The last band is the one that limits what we can claim about soundness.

The count

Thirty-six thousand of what

A qualia relation records how one thing stands to another: a book has print, a kettle boils water, a diagnosis has a date. Roughly thirty-six thousand of them are authored, which sounds large and is small.

Twelve per cent are model-generated and not yet human-verified. Soundness with respect to a knowledge base nobody checked is not soundness, so that twelve per cent bounds what the engine can honestly claim.

Coverage

The number that actually matters

Coverage is strictly less than one and stays there. Every relation not in the set is a bridge the engine will flag rather than resolve. Adding relations moves the boundary and never removes it.

Related limit

This is novel bridging counted from the knowledge base side. The failure accounting states it as a limit; this entry states it as a number.