Start with one useful learning question.
Run a first learning cycle.
Choose a decision worth improving
Name an owner, affected people, and a question worth answering.
Establish the starting point
An unknown baseline is a valid first finding; do not claim improvement yet.
Plan change and human practice
State the candidate, comparison, criteria, participant conditions, and permitted effects.
Run within agreed boundaries
Collect comparable evidence; do not quietly change criteria to pass.
Review evidence, then decide
Keep failed, supported, and insufficient findings distinct.
Verify change and revisit lesson
Observe what actually changed and when the lesson should be reconsidered.
Choose a reading path.
Teams
Start with the first cycle and worked example. Ask whose judgment and workload might change.
Implementers & reviewers
Review the specification, record boundaries, threat model, pilots, and partial conformance matrix alongside the example.
Use the original record structure.
These four source-grounded templates can be combined into one learning case. They do not mandate four separate forms. Downloads are labeled website adaptations with a pinned source header.
<!-- Website download adapted from TUN Systemic Evolution, revision a0d68bc061c7ed150a25a4cb1d13cd2613e1d351. Original: https://github.com/kochrisdev/TUN-Systemic-Evolution/blob/a0d68bc061c7ed150a25a4cb1d13cd2613e1d351/templates/hypothesis-record.md --> # Hypothesis record Status: blank template. Replace brackets with actual information; mark unknowns explicitly. - Record ID / revision: [identifier] - Learning owner and accountable human: [names or authorized roles] - Management initiative / outcome / LearningRecord references: [exact revisions] - Signal and source: [what happened, when, and where the evidence is held] - Observed fact: [what the source establishes] - Interpretation and alternatives: [possible explanation and competing explanations] - Assumption under test: [specific belief] - Hypothesis: For [population/context], changing [mechanism] from [baseline] to [candidate] should improve [outcome] while preserving [constraints]. - What would disconfirm it: [observable conditions] - Decision this evidence will inform: [adopt, revise, retire, or investigate what] - Existing supporting and contrary evidence: [references and limitations] - Proposed human practice or feedback intervention: [what people will do differently, or why none is needed] - Human development question: [skill, judgment, workload, agency, or justified non-applicability] - Applicability and transfer limits: [where this lesson may and may not be reused] - Next experiment or inquiry: [reference, owner, and trigger] - Source access and reuse restrictions: [who may inspect the evidence; permitted uses] - Review or expiry condition: [date or material change] An observation or proposed hypothesis is not an established learning claim.
Find the source you need.
Commitments, not certification.
These are the verified manifesto headings. They are distinct from the twelve draft TSEV requirements and are not a maturity score or conformance claim.
A short working vocabulary.
Hypothesis
A testable explanation of what a bounded change should improve and preserve.
Baseline
The stated starting point for comparison, even if initially unknown.
Candidate
A particular versioned change proposed for evaluation.
Evidence
Source-linked observations with limits and permitted uses.
Evaluation
A scoped judgment against declared criteria; not an adoption decision.
Adoption
An accountable decision about what should change, distinct from activation.
Activation
A separately authorized and observed change to current state.
Lineage
The trace from assumption and evidence to candidate, decision, and effect.
Drift
A change in conditions that may weaken a lesson's applicability.
Recovery
A checked response to adverse effects; restoration cannot undo every consequence.