Two Sheets Both Record the Same Grain Owed
Returning to the deferral negotiation of Chapter 2: Lin He and Zhou Cen asked Xu Wen to preserve their agreement. Xu Wen did not create this request—the delivery of the hull plank and the commitments of both sides had occurred first—what he did was keep the relevant content in a position where it could be seen again. The first sheet of paper recorded the subjects, the specification of the grain, the original date, and the part already completed; the revision was written on another sheet.
A few days later, Xu Wen gave the first sheet to the toolmaker who came to inquire. He thought the original date could still help the other judge, and forgot to provide the revision at the same time. The toolmaker therefore concluded that Zhou Cen was long overdue. Lin He said the two sides had already changed the agreement, but the toolmaker pointed at the paper and answered: "Is this not what you jointly confirmed?" A genuine record, detached from state and version, supported an inaccurate present judgment.
This is still a thought experiment in the fictional Tingqiao. The problem is not that paper is inherently unreliable, nor that Xu Wen must have had ill intent. A record can accurately preserve the past and at the same time lose, when used, its sufficient explanation of the present. The original agreement, the later revision, and the actual performance are each true; how they connect determines what kind of request a sheet of paper can support.
This chapter must answer why joint records can acquire authority, and how they keep their character as a limited tool of cooperation. Once a record becomes the basis on which others judge standing and act, an error affects more than memory. It may change who obtains material, who is permitted to wait, and who must first prove that they have not broken trust.
Records First Reduce Repeated Proof
Without joint records, each time Lin He raises the remaining request she must retell it from the delivery of the plank. Each time Zhou Cen delivers part of the grain, he must show that this part is not some other exchange. The longer the time and the more participants, the more easily repeated narration omits conditions, and the harder it is for those concerned to know which entry the other is speaking of. Paper gives subjects and requests a relatively stable correspondence.
Records also support handover of work. If Xu Wen leaves for a time, another person can know the original request, the partial performance, and where the revision stands, without depending entirely on Xu Wen's personal memory. A relation moves from the tacit understanding of acquaintances into checkable joint representation, increasing the capacity to cope with changes of personnel. Part of the reason for the authority of records arises here: it helps the community continue to recognize unfinished matters under limited memory.
RC understands language and writing as representation that supports cooperation and at the same time causes loss of experience. Bookkeeping in particular requires this two-sided recognition. It omits most of the original conversation, letting important conditions be reused; the omission may also exclude the tacit background of the time. There is no need to copy every conversation into the ledger in full, but one must know what judgment the ledger is meant to support, and which omissions would distort that judgment.
"Two bags of grain owed" is suitable for reminding that a request remains; it is not necessarily enough to determine maturity, specification, and responsibility. "Already deferred" can preserve a change, but does not state who agreed, to what day, or whether the original request has changed. The more concise the expression, the more its use must be made explicit. Simplicity is not an error; it is when a simple field is used for adjudication beyond its scope that an interpretive gap may turn into power.
Joint records are therefore not a copy of joint experience. Lin He remembers how the waiting affected her work, Zhou Cen remembers how the harvest and transport changed, and Xu Wen preserves what they confirmed. The three positions are not the same; the record can connect them and should not cancel any position's standing to keep providing new material.
What Kind of Identification an Entry Requires
Xu Wen reorganizes the record: he first gives this hull-plank request an identifiable entry, then places the original agreement, the partial performance, and the revision in the same place. An identifier does not prove that the entry's content is correct; what it solves is not mixing different requests into one entry, and not letting the same request be counted twice across scattered copies.
Who the subject is, what the object is, where the obligation arises from, which parts are completed, and what term is currently in force each need to be retained. If only the balance is kept, it may later be impossible to know which delivery or modification formed the remaining quantity. If only the process is kept without confirming the current state, each reader may have to re-adjudicate for themselves every time. A record must both retain the course of origin and make the currently usable conditions recognizable.
For multiple grain deliveries on the same day, it must also be stated which delivery corresponds to which request. The bag of grain Zhou Cen brings may complete part of this hull-plank request, or may be used to buy a new object. The same quantity of an object does not equal the same obligation. Confirmation of correspondence should occur at the time of delivery, and the keeper of accounts cannot afterward choose for himself, in order to preserve some balance, which entry to assign it to.
Dates likewise have different objects. The date the agreement was made, the expected date of performance, the formal maturity, the date of actual delivery, and the date of correction each support different judgments. Treating the time of actual entry as the time of performance can make a timely performer look overdue because of a registration delay; treating the expected date as an agreed maturity may also enlarge the original request. An accurate number cannot make up for a wrong object.
At this stage, Tingqiao keeps only requests for specific objects, and does not convert all claims in advance into bridge-yuan balances. The common unit has not yet formed, and grain and transport cannot be added together merely because they are written on the same sheet. When universal payment appears later, records will gain a capacity for common measurement, but the recognition of subject, time, object, and state still cannot be replaced by a single total.
Why Verification Is Not a Stamp of Approval
Xu Wen asks both sides to confirm the entry. Confirmation provides a ground: at that point in time, both sides accept that the content matches their agreement. But a name appearing on paper also requires knowing the object of confirmation. If they confirmed only that the hull plank was delivered, and did not read the term added later, the same confirmation cannot automatically cover the term.
To make confirmation usable, Xu Wen reads out the main conditions item by item and explains the later changes separately. Lin He can point out that the waiting arrangement is missing; Zhou Cen can state that the part already delivered has not been counted. This action does not require both sides to master all documentary methods; it makes the content directly related to their obligations understandable. The authority of symbols is supported by a checkable relation and cannot be established by form alone.
Some worry that allowing the parties to raise objections repeatedly will make the account permanently unstable. The worry has reason: if either side can withdraw already-confirmed content without material, a request can hardly support action. But stability can rely on the object of the objection and the rules of handling. Newly occurring performance should be updated promptly, errors in old fields should have their grounds checked, and repeated disputes without new material may temporarily keep the existing state.
Maintaining a state does not equal declaring it forever correct. It states that the current grounds suffice to support limited use, and that it may still be reopened if relevant material appears. RC's processual completeness appears here as the continuous existence of an entrance for judgment, rather than doubting everything anew at every use. That the ledger is usable precisely requires this cooperation between temporary stability and correctability.
Joint confirmation also cannot directly bind those who did not take part. The toolmaker may learn of Lin He's and Zhou Cen's agreement from it, but whether to accept its request still requires forming his own judgment. That Xu Wen issues a certificate does not mean he is entitled to require the toolmaker to provide tools. The evidentiary force of a record and the recipient's obligation to act cannot be merged into one.
Who Can Explain the Blank
After the reorganization, blanks remain on the ledger. How much of the request the transport help during the deferral actually offsets was never determined: the two sides originally said only "help carry on" and did not fix a settlement in grain. If Xu Wen converts it on his own for the neatness of the account, he will write a consensus not yet formed as a completed certainty.
Blanks need classification. Some are copying omissions that can be restored from the original material; some are matters the two sides genuinely did not agree on and that require separate negotiation; some concern actual performance and await observation. The three kinds of blank cannot be handled with a single default value. If, in the absence of a term, the earliest possible date is automatically adopted, and in the absence of a substitution ratio the ratio preferred by the keeper of accounts is automatically adopted, the table may quietly add obligations.
The institutional meaning of the power to interpret appears in just such details. The keeper of accounts need not directly command anyone; so long as his filling-in is treated by others as a reliable request, Lin He and Zhou Cen must act within the filled-in framework. Xu Wen's technical choice about "what is more convenient to write" may change who delivers, who waits, and who is entitled to consider themselves finished.
RC's power hierarchy concerns how evaluative conclusions propagate downward. Joint records are one path of transmission: an interpretation enters an entry, the entry enters the toolmaker's judgment, and the judgment in turn affects the resources Zhou Cen obtains afterward. As the scope of transmission expands, the keeper's interpretation is no longer merely helping two people remember. He can influence entrances to action that were not originally part of the conversation.
Records should therefore distinguish jointly confirmed content, one-sided statements, provisional interpretations, and undetermined parts. This distinction may make the account less neat, but it protects its accurate use. For interpretations that genuinely require provisional adoption, the grounds, authority, and scope should be stated, and a convenient temporary handling should not be quietly upgraded into an indisputable fact.
Authority Comes from Continued Use, and May Harden in It
Xu Wen has kept a large number of entries, and members find it more convenient to check with him than to ask each person. More and more transactions consult Xu Wen's paper first, and then decide whether to continue. The authority of records does not require a grand appointment; it can form in continued use that reduces the cost of verification. The gains of professional division of labor push the scope of delegation to expand.
Expansion is sometimes necessary. If every toolmaker re-investigates Zhou Cen's harvest and grain delivery, much duplicated resource is consumed. Joint records let part of the verification be reused. But reuse should know the scope of the original observation: that Xu Wen checked a certain grain delivery does not equal having verified all future harvests; that a settled request supports its history of performance does not guarantee the future will certainly be reliable.
The danger is that an entry turns from a specific state into a fixed identity. The toolmaker says, "Zhou Cen has a record of being overdue, so he cannot be trusted," and then also understands Zhou Cen's offering of new material as evasion; the original entry begins to exceed its object. One erroneous transmission may close the entrance that was meant for updating, and the party can only explain himself again within the conclusion of "untrustworthy."
This is precisely the mechanism by which evaluation alienation may enter records: the record no longer accepts correction by consequences and new material, and instead lets all material prove the existing identity in reverse. The more people adopt the same conclusion, the more it looks like a consensus across subjects; in fact they may only be repeating one outdated version. The number of times something is jointly used does not equal the number of added independent observations.
Professional delegation and alienation must still be kept apart. If Xu Wen makes state and version explicit, allows supplementary delivery material, and corrects transmission beyond its scope, records can spare weaker subjects from having to prove themselves anew each time. Centralized bookkeeping may even protect those who command fewer resources. One cannot deny this function merely because authority is concentrated; one should see whether the concentration preserves a path of return.
Correction Must Handle the Transmission That Has Already Occurred
Xu Wen admits that he forgot to attach the revision and corrects the current entry. If the toolmaker still keeps the old copy, the original judgment may continue to affect transactions. That the source record is now correct does not mean the transmitted conclusion has been updated. Correction requires identifying who used the old material, which judgments were affected by it, and whether the relevant entrances can receive a sufficiently timely response.
In the thought experiment, Xu Wen provides the toolmaker with the original agreement, the revision, and the current state of performance, noting what the earlier copy lacked. He need not make Zhou Cen's whole life public, nor disclose irrelevant private matters to prove his own honesty. Correction should center on the actual error and the scope of transmission, both supplying what the judgment requires and limiting the unlimited expansion of the authority of records.
The toolmaker may judge anew, and is not required by the correction to accept the deferred request. What correction restores is the opportunity for correct material to enter his decision, not a guarantee that Zhou Cen obtains the expected result. If the toolmaker genuinely cannot wait, he may still refuse; if his refusal rests mainly on the old interpretation of overdue, the new material should be given a real possibility of changing the judgment.
If the old material has already caused Zhou Cen to lose an opportunity, a later correction cannot retrieve that opportunity as it was. One should distinguish correcting the current record, reducing continued impact, and handling the loss that has already occurred. Calling all three "the problem is solved" lets the keeper of accounts end the whole responsibility with one technical action. Which consequences can be made up and which can only be acknowledged still require concrete discussion by scope.
Correction must also preserve a trace. Deleting the old version outright may make it impossible for others to know what they once acted on, and also make Xu Wen's error no longer checkable. Keeping the old record together with the reason for correction lets later judgments review the process; at the same time the old version should not continue to be treated as the present state without explanation. Historical visibility and current usability are not the same marking.
A Person Under Correction Cannot Wait Only Inside the Procedure
Lin He waits for the toolmaker to look at the material again, and her tool repair is thereby postponed. The toolmaker is unwilling to bear the interval of Zhou Cen's future grain delivery, and the misjudgment caused by the old copy makes this verification even slower. If verification takes a long time, even an ultimately accurate correction may leave the actual entrance meaningless. The reliability of records therefore also includes the timing of handling. Being correctable and being promptly correctable leave subjects in different conditions.
But demanding that all disputes be checked immediately may also crush Xu Wen and increase further errors. Tingqiao needs to arrange the order of verification by consequence: errors directly affecting near-term basic deliveries should be handled early, complex disputes with slower consequences may adopt an explicit provisional state, and other requests proceed by an explainable procedure. The ordering itself is also evaluation, and those affected should know its grounds and be able to present material of urgency.
The scope of transmission may be limited for a disputed state. For example, only this request is marked "revision material pending verification," without freezing all of Zhou Cen's transactions. Whether a certain resource needs temporary restriction depends on the loss an error or duplicate request may cause, and "not yet determined" cannot be directly expanded into general disqualification. The unknown has its own scope and does not automatically support the broadest restriction.
Lin He also cannot lose an already-confirmed request during correction. That she provides a statement of revision does not mean this grain can stop being recognized. Provisional handling should retain the part that already has grounds, separating the genuine dispute from other definite states. Otherwise any tiny objection may leave the whole obligation suspended, in turn weakening the requester's actual capacity.
These conditions show that empowering a subject is not merely granting a standing to complain. Support in organizing material, a reasonable time limit, provisional states, and necessary transition determine whether the standing can be used. If a member must stop all work in order to explain an error again and again, he may be able to do no more than give up correcting it. If the procedure is open but the cost of using it is too high, the authority of records will still concentrate among those familiar with the procedure and able to afford the wait.
Who Supports the Labor of Bookkeeping
As entries increase, the time Xu Wen spends on verification and correction also increases. If the community only praises his accuracy without supporting it with resources, he may shorten checks, postpone updates, or rely on the oral confirmation of acquaintances. If the cost of maintaining records is ignored, it may be transferred in the process into other people's risk.
Bookkeeping does not run on paper by itself. It requires preservation, recognition, version handling, continuity, and error verification. Who bears this labor and how it is supported are also part of economic cooperation. Providing records free of charge may be a contribution someone is willing to make, but it does not mean the community can expand the scope of the contribution without limit, without asking about the actual conditions it bears.
If Liang Xu later is willing to support Xu Wen's maintenance, records may obtain more stable resources, but support and authority should still be separated. Providing maintenance resources does not automatically confer the right to rewrite others' entries. If Xu Wen can preserve his support only by giving priority to Liang Xu's requests, resource dependence may enter a record standard that claimed to be uniform.
The community should therefore state the source of support and the boundary of delegation, and keep checkable material about the relations involved. This does not require every keeper of accounts to be entirely without interest, but requires that the interest involved not make the record accept the observations of only one position. Alternative preservation, handover of work, and checking from different positions can reduce the risk of bookkeeping depending entirely on one person.
Maintenance itself also needs a sense of scale. If every small request demands a long investigation, the cost of verification may exceed the cooperation it can support. Different depths of checking may be used according to purpose, but it should be explicit which record suffices for which decision. A simple reminder should not be taken to decide an irreversible disposition, and detailed verification should not become a threshold that prevents all routine delivery.
The Boundary of Provisional Adoption
After the material is completed, the toolmaker asks whether they must still wait for every detail to be fully consistent. Xu Wen says that at present it suffices to confirm that the original date underwent joint revision and that part of the grain was indeed delivered, but that the offset scope of the transport help has not yet formed an explicit agreement. The three judgments should be used separately: the first two need not be abandoned because the last is undetermined, nor may a ratio be filled in for the last because the first two are definite.
This is a scope of stopping, not a stopping of continued observation. Present action requires certain conditions, and when material suffices to support them, they may be adopted provisionally. Undetermined parts should remain marked, and their checking continued when they affect some new decision. If every record had to prove all past and future, no record could enter work; if once used it no longer allowed updating, the record would again lose processual completeness.
Some questions are irrelevant to the present decision. For example, the toolmaker need only know whether this request genuinely exists; the details of another of Zhou Cen's carriage jobs need not necessarily enter. If the toolmaker hopes to accept future grain on this basis, the date of performance and the remaining quantity cannot be omitted. Whether a record is sufficient must attach to the purpose, and "already reviewed" cannot become a uniform pass valid across every scale.
When a conflict cannot be resolved, it should be stated which material differs, what provisional judgment is currently adopted, who will continue checking, and what change would make the provisional judgment no longer usable. A provisional state cannot continue indefinitely with no one taking it up, nor does every conflict require immediately adopting the maximum restriction. Sustainable records maintain action in this intermediate range: neither pretending to have exhausted everything, nor leaving everything suspended because nothing is exhausted.
Only in this way does the ledger's authority come from an explicit scope of usability, rather than from a "final" label that asks people to stop thinking. When each use knows what it depends on and each deviation knows where to return, records can support the larger-scale cooperation that comes later.
Joint Confirmation Is Not Yet Joint Payment
After the reorganization, Lin He's and Zhou Cen's request can be identified more reliably, and the toolmaker can understand the difference between the original copy and the current state. Tingqiao has obtained a set of records more reusable than private memory, but joint records are not yet joint payment. The toolmaker must still decide whether to accept deferred performance; the record cannot bear the waiting for him on its own.
A record of a claim can prove standing to request; a record of a delivery can prove that a certain performance occurred; a history can support a limited judgment of reliability. For these proofs to participate directly in wider payment, there still need to be a scope of acceptance, rules of transfer, limits on repeated use, and handling of failure. Different roles do not naturally merge merely because they are written in the same ledger.
Beginning from the two sheets of the record, this chapter has explained where its authority comes from: records reduce repeated proof, support handover, and give obligations not yet completed a stable position. It has also explained how authority expands: interpreting blanks, determining states, and transmitting entries can influence many later entrances to resources. Both functions come from the same cooperative capacity, and one cannot acknowledge only the benefit or only the risk.
For a record to keep its character as a limited tool, it requires a scope of confirmation, a current version, a checkable course of origin, constrained interpretation, and a real path of correction. It should not fix a person's identity with one erroneous field, nor end all responsibility for transmission because the source record has been changed. True stability is not that the old paper never moves, but that new material from different positions can still change its reasonable use.
Xu Wen places the original entry and the revision in the same place, and explains to the toolmaker what can currently be confirmed. Zhou Cen does not thereby obtain everyone's unconditional trust, nor does Lin He thereby obtain grain produced ahead of time. But the two sides' request no longer depends only on who remembers more clearly, and the next cooperation has a point of support that can be checked.
Tingqiao can now ask a further question: if a person does not want to accept private deferrals one by one, can an exchange be completed through a more widely accepted arrangement of payment? Joint records will become one of its necessary conditions, though not the only one. The mechanism by which payment exceeds acquaintances must continue to hold beyond records.