Shared Wisdom Loop

How does a group become wiser without forcing everyone to think alike?

Use the Shared Wisdom Loop to turn evidence into know-how, know-how into independent capability, and experience back into better shared judgment.

purpose → question → gather → test → integrate → argue
→ scaffold → practise → transfer → observe → evolve

The result is a hive mind only in the useful sense: many distinct participants can retrieve, challenge, apply, and improve shared knowledge. It is not consensus, obedience, or one averaged voice. Provenance, uncertainty, protected human judgment, and useful dissent must survive.

Why Learning Comes First

Learning is the meta-capability behind every other capability. It determines how quickly a person or agent can acquire a model, test it, act without support, transfer it to a new context, and improve the lesson for the next participant.

Knowledge has value when it improves a worthwhile judgment or action. More documents, links, retrievals, or fluent answers do not prove wisdom. The loop closes when a beneficiary experiences a better outcome and the evidence improves the next decision.

The Three Loops

Inner loop: think honestly

Use the Integrity Stack before accepting or sharing a knowledge unit.

Gate	Question	Required result
Ontology	What exists?	Entity, claim, boundary, state, and relationships
Epistemology	How do we know?	Source, evidence, confidence, uncertainty, and falsifier
Taxonomy	Where does it belong?	Domain, reader job, lifecycle, canonical owner, and consumers
Nomenclature	What is it called?	Stable term, useful aliases, and retrieval cue
Axiology	Why does it matter?	Beneficiary, value, protected constraint, and decision improved

Stop at the first unresolved gate. Do not fill uncertainty with polished prose.

Outer loop: communicate for judgment

Build an argument another participant can inspect:

valuable question → bounded claim → reasons → evidence
→ strongest counter-case → limits → practical consequence

Use emotion to help the learner recognise why the question matters. Never use it to bypass judgment. A sound argument makes its premises, evidence, uncertainty, and consequences easier to challenge. Rhetoric should improve judgment, not make weak reasoning persuasive.

Learning loop: transfer capability

Use the learner's Zone of Proximal Development (ZPD) to choose a challenge just beyond what they can do alone. A More Knowledgeable Other (MKO)—a person, agent, example, tool, or resource—then provides the smallest useful scaffold.

The MKO does not own the learner's judgment. Support adapts to evidence and fades as capability grows. Transfer is stronger when the learner can explain, apply, challenge, and adapt the knowledge without the original support.

Run the Loop

1. Declare value

Name the beneficiary, valuable change, current independent state, protected constraints, and the decision the knowledge should improve. If the value is unclear, do not gather more information yet.

2. Ask one decision question

Turn the topic into a question that could change a belief, action, or control. State what evidence would change the current answer.

3. Gather bounded evidence

Seek primary, practical, contrary, and historical evidence in proportion to the decision. Preserve sources and exact observations. Separate observation, interpretation, hypothesis, working proof, and mature evidence.

4. Test through the Integrity Stack

Define the claim, test how it is known, place it with one canonical owner, give it a stable name, and state why it deserves future attention. Preserve unresolved dissent beside the claim it limits.

5. Integrate for retrieval

Update the canonical owner instead of creating parallel truth. Use OKF-compatible Markdown as the portable floor:

• type identifies the artifact's job;
• title and description make its value legible;
• tags and retrieval cues support discovery;
• provenance, confidence, freshness, and review conditions bound trust;
• typed relationships connect the knowledge to evidence, dependencies, risks, and applications.

OKF makes knowledge portable. It does not decide whether the knowledge is true, valuable, or safe.

6. Construct the teaching bridge

Meet the learner at their current model. Create an honest emotional connection, explain the value, show a credible contrast, teach the smallest useful model, and give one reachable action. Include a counter-case or failure mode so the learner can test rather than merely repeat the lesson.

7. Scaffold practice

Choose one ZPD challenge. Supply only the support needed for the next attempt. Observe where the learner succeeds, hesitates, or depends on the scaffold. Adjust the challenge from evidence, not assumed ability.

8. Fade and test transfer

Remove support deliberately. Ask the learner to perform independently, explain the reasoning, find an error, and apply the capability in a materially different context. Assisted output is evidence of support, not proof of learning.

9. Return evidence

Record what the learner or agent could do before, with support, without support, and in the second context. Capture the beneficiary consequence and any counterevidence. Return the result to the canonical owner.

10. Evolve the shared model

Retain, narrow, revise, or retire the claim, scaffold, relationship, or standard. State the next question and the evidence needed to claim improvement. Do not let popularity, repetition, or fluent retrieval overrule evidence or named human authority.

Minimum Transfer Packet

Use this small packet when knowledge must cross between people or agents:

Purpose and beneficiary:
Decision question:
Claim and confidence:
Evidence and strongest counter-case:
Canonical owner and retrieval trigger:
Relationships:
Independent state:
Proximal move:
MKO and contingent scaffold:
Fading plan:
Independent and second-context proof:
Evidence-return path:
Review or kill condition:

Proof

The loop is working when:

• a cold participant retrieves the rightful knowledge without reconstructing its history;
• they can distinguish evidence, inference, uncertainty, and dissent;
• the argument changes a considered judgment rather than winning passive agreement;
• support helps them cross one reachable capability gap and then fades;
• they perform independently and transfer the capability to another context;
• observed use changes the canonical lesson, relationship, question, or standard.

Do not substitute content volume, search ranking, traffic, citations, assisted completion, or consensus for these outcomes.

Failure Modes

• Collection without value: information grows without a decision it should improve.
• Semantic theatre: metadata and links multiply without improving retrieval or judgment.
• Argument as manipulation: emotion or rhetoric conceals weak evidence or commercial intent.
• MKO authority inflation: the helper's confidence replaces the learner's judgment.
• Permanent scaffold: support creates dependence instead of independent capability.
• Consensus collapse: disagreement is averaged away to make the shared model look settled.
• Parallel truth: a new note or page competes with the canonical owner.
• One-way teaching: use produces evidence that never returns to improve the lesson.

Practice

Choose one important piece of know-how your group repeatedly explains. Complete the Minimum Transfer Packet, teach one learner at their current edge, fade one scaffold, and test the capability in a second context. Change the shared model only after comparing the evidence with the starting state.

Context

• depends-on Learning — defines independent performance, fading, and second-context transfer.
• depends-on Knowledge Schema — governs the Integrity Stack, canonical ownership, and semantic structure.
• pairs-with Thinking — supplies multiple reasoning modes for constructing and challenging the inner model.
• risk-governed-by Truth Recognition — keeps confidence, claims, and rhetoric bounded by evidence.
• applies-to Research — turns a decision question into source-traceable evidence.
• applies-to Note Taking — captures knowledge units without erasing provenance or human synthesis.
• risk-governed-by Wisdom Transfer Standard — prevents readable artifacts from being mistaken for transferred capability.
• pairs-with Open Knowledge Format — provides the portable metadata floor for human and agent retrieval.
• pairs-with Inner/Outer Knowledge Extension Model — keeps public teaching and private operating evidence connected without sharing authority.

Questions

• Which participant should become more capable, and what can they do independently now?
• What valuable judgment should this knowledge improve?
• Which dissent or uncertainty must survive transfer?
• What support should fade next?
• What evidence from use should change the shared model?

Teach-back: explain how honest thinking, inspectable argument, semantic retention, and fading support combine to produce shared wisdom.

Changes my mind: comparable evidence that retrieval or assisted performance alone predicts independent judgment and second-context transfer would narrow this method.

Next question: which repeated explanation should become the first measured Shared Wisdom Loop?