Back to The Search for 100 Thinkers

Chapter 3 of 13 · 2 min read

Two maps, neither a leaderboard

Capabilities describe the work an investigator can do; questions describe what IoV must survive.

The Search for 100 Thinkers · Published 1.0 · 2026-09-22 ·

The word “systems” had been doing too much unpaid labour. We needed to unpack it.

The canonical research ontology contains 22 capabilities and 30 IoV questions. The capabilities distinguish kinds of intellectual work. The questions identify challenges for the proposal. They are related, but they are not interchangeable.

Ability and demand

A capability asks: what can this investigator demonstrably do?

A question asks: what must the proposal be able to withstand?

Two maps meet at inspected evidence
  1. Capability: a method someone can demonstrate
  2. Artifact: inspect the actual work
  3. IoV question: explain the relevant test

A tag such as “measurement” can help us find a person. It cannot establish that the person has demonstrated the particular method our question requires. That needs an inspected piece of work and a reasoned connection.

Five particularly consequential gaps

For the next evidence audit, five gaps matter enough to keep visible:

GapPlain-language test
Q12: measurement invarianceDoes the measure mean the same thing across groups or occasions?
Q21: conversion into capitalWhat mechanism, if any, connects the proposed value to capital?
Q24: adversarial metricsWhat changes when people have an incentive to game the measure?
Q28: structural and dimensional validationDo the model's relationships and units hold together?
C22: implementation feedbackCan implementation reveal errors and change the design?

These are editorial explanations of selected entries, not a replacement ontology. Having all 22 and 30 columns in a matrix means the questions have a place to go. It does not mean we have answered them.

A beautifully labelled empty drawer is still an empty drawer. Useful, certainly. Full, no.

~

End of signal.