A packet can pass every software check and still fail to transfer a capability to another person. The receiver may misunderstand a term, overlook a missing permission, rely on an outdated assumption or be unable to perform the task without help from the original creator. A published handover document is therefore not evidence that a human handover succeeded. The receiving person must attempt the work under declared conditions, while an observer records what they can do, what they need and where the packet fails to carry its meaning.
CivilisationOS | The First Human Receiving Trial: Testing Whether a Packet Can Be Inherited turns the prepared Receiving Exercise 01 into a study protocol. It links the portability book’s packet, host, keeper, recovery and unfinished-work chapters to the education question: can a successor interpret, apply, check, adapt and pass on a defined capability? This article proposes the protocol and its reporting rules. It does not claim that participants have been recruited, that a city service has been transferred or that an educational intervention has produced a measured improvement.
- Separate possession of a packet from demonstrated ability to use it.
- Use distinct creator, receiver, facilitator and observer roles.
- Define the receiving task, starting knowledge, permitted tools and assistance before the trial.
- Test routine work, missing evidence, stale instructions, interrupted replies and a changed condition.
- Record correct action, justified refusal, explanation, assistance, time and unresolved dependencies.
- Keep critical authority and safety failures separate from ordinary usability scores.
- Run an adult pilot first; use appropriate ethics review, consent and safeguards before involving children or public services.
- Publish null, partial and failed outcomes without turning one exercise into a city-readiness score.
Receiving Exercise 01 and the city-transfer dossiers · Portable Civilisation | Reboot Sequence · Complete book and reading map
1. Why the human handover is a different test
A software test asks whether a defined program behaves as expected for specified inputs. A human receiving trial asks whether another person can understand and carry out a task using the material supplied, given a stated level of prior knowledge and support. These tests overlap, but they answer different questions. A program may parse a packet correctly while a person cannot tell which instruction applies to the situation in front of them.
The distinction matters because the portability project is about capabilities continuing across changes of carrier. If the original creator remains beside the receiver, silently correcting every omission, the task may appear successful even though the packet did not contain enough information. The assistance may be generous and useful, but it must be recorded. Otherwise, the evidence describes the original creator’s ongoing participation rather than an independent handover.
The existing Receiving Exercise 01 supplies learner cards, a facilitator answer guide and an observer form. The current book map explicitly describes the exercise as prepared but untested with people. This companion sets out how to run a bounded first study without converting a prepared exercise into a claimed result.
2. State the claim before selecting participants
A study should begin with one specific claim. For example: “A receiver with the stated starting knowledge can use this packet to complete the defined fictional resource-allocation task, identify the two supplied conflicts and produce a usable handover record without unplanned intervention from the creator.” This is testable. It does not claim that the receiver can operate a whole city, understand every part of civilisation or reproduce an industrial base.
Write the boundary into the protocol. Identify the task, resource, authority, environment, duration, permitted assistance and what counts as completion. If the task uses fictional records and paper tokens, say so. If the receiver is allowed to use a calculator or AI assistant, specify which operations are permitted and what must be disclosed. If the task is designed for a working service, the study requires a different risk assessment and approvals before live use.
The claim should also identify what it does not test. A short exercise can demonstrate that a person understood a packet and completed a simulation. It cannot establish that an institution is legitimate, that a service is safe in production, that the packet will work in a different language or culture, or that the capability will survive several generations. Those require separate evidence.
3. Separate the roles so the answer is not leaked
| Role | Responsibility | Boundary |
|---|---|---|
| Packet creator | Prepares the packet, declares assumptions and lists known limitations. | Does not coach the receiver during the independent attempt except through the predeclared assistance route. |
| Receiver | Uses the packet to perform the task, asks questions and records decisions. | Does not see the facilitator’s answer guide before the attempt. |
| Facilitator | Explains the study process, protects participant welfare and supplies only permitted assistance. | Does not quietly steer the receiver toward the expected answer. |
| Observer | Records actions, decisions, errors, assistance, time and explanations. | Does not turn an observation into a favourable interpretation during the task. |
| Adjudicator | Applies the predefined outcome rules to the evidence after the attempt. | Should not change the success criteria after seeing the result. |
Where practical, use a receiver who did not help write the packet. The purpose is not to test whether the creator can explain their own work. It is to find out what a new carrier can recover from the material itself. If the only available receiver has prior exposure, record it rather than pretending the trial was blind.
4. Record the receiver’s starting position
Before giving the packet, record the background knowledge needed for the task. This can be a short set of questions about terms, tools, rules or procedures. The purpose is not to grade the person. It is to interpret the result fairly. A receiver who already knows the task may succeed for reasons that have little to do with the packet. A receiver who lacks an essential prerequisite may struggle even when the packet is clear.
Use only information relevant to the study. Do not collect unnecessary personal details, account data or sensitive histories. A participant code can identify the observation record while the consent information is stored separately. Report the group and task conditions, not a person’s identifiable story, unless there is a clear and approved reason.
Give the receiver the same starting resources that a real recipient would receive under the declared scenario. If a device, archive reader, local manual or reserve of materials is part of the host environment, list it. The test must distinguish what the packet carries from what the receiving host supplies.
5. Define the task and its safe limits
The first pilot should use a fictional, low-consequence task. The Open Reading Room example is suitable because it can test membership, a shared catalogue, a small reserve of equipment and the handover of unfinished work without controlling a real public service. Another fictional setting may be used if the same boundary is stated.
Specify what the receiver may do independently. For example, the receiver may issue a temporary reservation from a declared local allocation, but may not change the institution’s shared authority rules. The packet must say which records are historical, which are current, which actions are provisional and who can authorise a change. A successful trial requires the receiver to notice those distinctions rather than merely complete a sequence of clicks.
Set a stopping rule. If the receiver faces an ambiguous instruction that could create an irreversible or harmful action, the correct response may be to stop and request clarification. The exercise must not punish a justified refusal merely because the facilitator expected action. In a real service, the risk assessment and professional rules would determine the stopping thresholds.
6. Use a scenario set that exposes different dependencies
A single normal case often shows only that the packet works when every assumption is satisfied. A receiving study should include several defined cases, each designed to expose a different dependency. The facilitator’s guide should state the expected evidence for each case before the receiver begins.
| Scenario | What it probes | Evidence to record |
|---|---|---|
| Routine task | Can the receiver understand the ordinary procedure? | Correct steps, explanation and completion record. |
| Missing evidence | Does the receiver recognise that the packet does not answer the question? | Request for evidence, refusal or declared assumption. |
| Stale instruction | Can the receiver distinguish an older instruction from a current one? | Which source was selected and why. |
| Cancelled action | Does the receiver avoid repeating an action that was later cancelled? | Inspection of the cancellation record and resulting decision. |
| Conflicting resource claims | Can the receiver identify that two promises compete for one resource? | Conflict recognition and escalation route. |
| Missing reply | Can the receiver distinguish an unconfirmed action from a completed one? | Retry, waiting, reconciliation or safe hold. |
| Changed environment | Can the receiver adapt without changing the task’s essential purpose? | Modification, reason and verification. |
| Unfinished work | Can the receiver identify what remains open for the next carrier? | Recovery list and named next action. |
7. Make the attempt observable without turning it into surveillance
The observer needs enough detail to reconstruct the task, not a recording of every private thought. Record the material supplied, the time the receiver began and ended, which decisions were made, which evidence was consulted, what assistance was requested and whether the receiver could explain the final action. A short think-aloud explanation can be used when appropriate, but it should not be assumed to reveal every mental process.
A useful record distinguishes an action from the reason given for it. It notes whether a person opened the current record, whether they identified the missing evidence and whether they knew how to escalate. The observer should not infer understanding merely because the screen displayed the expected result. Nor should the observer infer incompetence from a pause that may reflect careful checking.
For AI-assisted tasks, record the role of the tool. Did it explain a term, summarise a source, propose a procedure or recommend a decision? What did the receiver accept, reject or verify? Keep the task’s core learning claim visible. An AI system may make the text easier to read while the human receiver remains responsible for deciding whether an instruction is authorised and safe.
8. Predefine the outcome rules
Do not invent a success threshold after observing the result. Before the study, define what counts as completed, partially completed, safely refused, incorrectly completed and not assessable. For a critical permission case, an unauthorised action may be a disqualifying failure even if the receiver completes every routine case. A high total number of correct answers must not compensate for a consequential breach of the task’s boundary.
It may be useful to report several dimensions rather than one score: interpretation, correct application, checking, justified stopping, recovery, adaptation and handover. The result can show that a receiver performed the ordinary task accurately but needed assistance when a reply was missing. That is more informative than a single mark that hides the dependency.
Record every instance of assistance. Assistance is not automatically a failure. The important question is whether the study’s claim permits it. A supported task, an independently performed task and a task completed after facilitator intervention are different outcomes. Report them separately.
9. Add a second handover to test succession
If the first receiver completes the task, ask that receiver to prepare a handover package for a second receiver. This second transfer probes whether the first receiver understood the underlying capability or merely followed the first packet once. The second person should receive the revised package and the declared host resources, not an informal explanation from the original creator.
The second handover should retain the same task boundary while changing one controlled condition. For example, the reading room’s normal catalogue display is unavailable, but a printed export remains. The receiver should identify which functions can continue, what must wait and how the records will be reconciled later. The changed condition should be within the announced scope of the exercise, not an arbitrary surprise.
A successful second handover is stronger evidence of transfer than a single successful attempt. It still does not prove indefinite succession, industrial independence or city-level portability. It demonstrates a limited capability across two receiving steps under the tested conditions.
10. Apply appropriate research and participant safeguards
Start with an adult pilot using fictional material. Explain the purpose, task, expected duration, records collected, possible discomforts and right to stop. Participation should be voluntary, and refusal should not affect employment, grades or access to services. Store the observation records securely and retain only what is needed for the stated analysis.
Before involving children, learners in a dependent relationship or people whose services may be affected, obtain the appropriate institutional review and permissions. Use age-appropriate information, guardian consent where required, participant assent and a clear separation between educational support and research participation. A classroom activity used for teaching is not automatically a research study, but a planned study involving people may require formal review under the institution’s rules and applicable law.
Do not test a new packet on a live high-consequence service merely to demonstrate portability. Use a sandbox, simulation or reversible exercise first. If later work involves a real service, the service owner must approve the scope, operational safeguards, rollback plan and stop authority before the trial begins.
11. Report the result with its limits
A report should state who performed the task in broad, non-identifying terms, what prior knowledge was declared, which version of the packet was used, which host resources were available, what assistance occurred and how the outcomes were classified. Include the cases that failed or could not be assessed. Preserve the original instructions and observation form so that a later team can reproduce the protocol.
If no participant completes the task, report that the exercise remains untested. If one participant succeeds, report one observed handover under those conditions. Do not convert that result into a percentage for a city, a population or a civilisation. A city dossier remains a map of possible receiving components until an actual receiving organisation has performed and documented a bounded test.
Where several groups or conditions are compared, decide the sample size and analysis before collecting confirmatory results. A pilot can identify unclear instructions and estimate practical variation; it is not automatically powered to establish an educational effect. A comparative study should account for prior exposure, order effects, task difficulty, facilitator differences and the possibility that the second group learns from the first.
12. The connection to AI and education
This protocol offers a different way to use AI in education. A learner can use an AI assistant to inspect a packet, ask what assumptions are missing, generate possible failure cases or improve an explanation for the next receiver. The task is not to hide AI from the assignment. It is to make the learner’s judgement, verification and handover visible.
For example, a student team could create a fictional museum exhibition guide, ask AI to identify unclear instructions and then transfer the guide to another team. The second team must assemble the exhibit, distinguish verified facts from speculation and correct one deliberately outdated label. The first team learns whether its instructions were understandable. The second learns how to inherit and check a resource. The teacher observes both the product and the transfer.
This is a practical expression of centre-and-edge education. The shared centre provides agreed knowledge and standards. The receiving edge tests whether those resources work in a specific situation. Feedback from the edge helps improve the next version. AI can assist both ends, but it does not remove the need for a person to judge evidence, authority, risk and purpose.
13. What this trial can and cannot complete
A carefully conducted human receiving trial would fill a real gap between code that passes selected checks and a person who can use inherited instructions. It could identify confusing terms, missing host assumptions, unclear authority boundaries, excessive dependence on the creator and the work required to prepare a successor.
It would not by itself show that the packet supports a functioning public institution, that the institution is legitimate, that a city can be reconstructed or that the material and energy systems can renew themselves. Those claims need their own evidence. A person may understand a maintenance manual while having no access to spare parts. A team may transfer a service procedure while lacking permission to operate the actual equipment.
The right outcome is a precise account of what crossed the boundary, what was supplied by the receiver, what failed and what remains open. If the trial exposes a weakness, improve the packet and test again. If it succeeds, preserve the conditions and limits alongside the result. A civilisation becomes more portable not when every transfer is declared complete, but when each transfer teaches the next carrier what is truly required.
Sources and connected reading
- Portability Of Civilisation | From City Comparisons to Transfer Dossiers — Receiving Exercise 01 and the learner, facilitator and observer materials.
- Portable Civilisation | Reboot Sequence — the proposed sequence from inherited material to a working service.
- Successors Inherit Unfinished Work: The Recovery Packet — carrying open commitments across a change of keeper.
- CivilisationOS | Humans, AI and the Division of Responsibility — separating tool contribution from human responsibility.
- CivilisationOS | From Inherited Knowledge to Emergent Institutions — the distinction between inherited information, enacted practice and functioning institutions.
- UNESCO AI Competency Frameworks for Students and Teachers — human agency, responsible use and AI literacy in education.
Continue the CivilisationOS investigation
This protocol is ready to be reviewed and piloted; it is not a report of a completed human trial. Return to the complete portability book, or continue into The Organisation Beyond Its Carriers for the wider education and succession argument.