Whose Success Is Reaching an Agreement?
At ten on Saturday morning, Tang Ke wrote "try to reach an agreement" in the goal field of the D17 action question card. To her, this was clear enough: the client was willing to talk again the following week, so of course the repair station hoped to win the job. Ye Cheng asked whether it would still count as success if the new job delayed the six slots already promised to other work.
Tang Ke added, "Reach an agreement without affecting other jobs." Rather than adding still more conditions to the sentence, Gu Ning separated what an agreement contained. Winning a new commission, fulfilling it as promised, maintaining a communicative relationship, learning about a new kind of need, and preserving the capacity to adjust are all possible desired outcomes. They do not always arrive together.
Chapter 1 separated probability, action, consequence, and deadline. This chapter addresses goals. If actors do not state what must be protected before they act, they can easily select some benefit that happened afterward and declare the original choice successful. They can just as easily stare only at one result that did not happen and erase every other value.
This remains the fictional Chengwan case. At ten on Saturday, no new message has arrived from the client, no Version 2 has formed, and no work bay has been reserved or material purchased. We organize D17's goals only; discussion of a goal must not make the action or result happen in advance.
A New Commission Is Only One Goal
The repair station may make obtaining a suitable new commission a goal. It could bring income, continuity of work, and a new opportunity for cooperation. This is a real goal and need not be hidden behind "learning about the need" merely to sound cautious.
But obtaining a commission is not an unconditional good. If the price cannot cover direct inputs, the delivery period conflicts with existing commitments, or the station lacks a required capability, obtaining it does not automatically improve the station's position. The action object should be a commission that can be fulfilled, rather than acceptance of any version of the terms.
"Can be fulfilled" cannot be defined by Gu Ning's confidence alone. It must correspond to scope, labor, materials, quality requirements, delivery window, and responsibility. Because Version 2 does not yet exist, the station can only define an interface for later judgment, not declare that future conditions will necessarily be feasible.
Making the acquisition goal clear does not calculate income in advance. It prevents Tang Ke from treating every review of the terms as an obstacle to success merely because an affirmative reply has finally arrived.
Fulfilling Existing Commitments Is Not an Ancillary Wish
Six of next week's ten labor slots are already promised to other jobs. Fulfilling them is not background outside D17 that can be ignored at will. It is an existing goal of the station. When a new opportunity enters, the clients behind the old work cannot become loss bearers with no place in the record.
If reserving two slots for R17 leaves the existing six untouched, the goals may proceed together. If old slots must be reassigned, the station must state who will be delayed, whether agreement has been obtained, and what consequences the change will create. It cannot record income on the new commission's page while leaving delay in a separate table no one reads.
Existing commitments do not absolutely prohibit every adjustment. Parties can sometimes renegotiate, and rescheduling after conditions change can be reasonable. But renegotiation requires material and consent. "Better overall" cannot accept the change automatically for the affected party.
Gu Ning therefore listed "fulfill the existing six-slot commitments" as a separate goal and put "no reassignment without corresponding agreement" among the candidate constraints. A goal states what is to be maintained; a constraint states which actions cannot presently be adopted. The two are related but distinct.
Income and Working Capital Occupy Different Places
Tang Ke said that income from a new job would replenish working capital. Gu Ning noted that they did not yet know when income would arrive, what had to be paid first, or whether rework would be required. A final increase in the ledger and the ability to keep paying throughout the process are two temporal relations.
A job may ultimately earn positive income while requiring more than the four deployable units at the start. If payment arrives after existing bills fall due, final profit cannot travel backward through time to fund present operations. Conversely, a lower-income job with clear advance payment and little occupation may make continued operation easier.
This chapter invents no price or payment method for a future version. It first gives "increase net contribution" and "maintain continuity of working capital" separate positions among the goals. When concrete terms arrive, a plan can be checked at both positions.
The distinction also prevents double counting. If net contribution already deducts a financing cost and the consequence table records it again as a separate loss, the cost is counted twice. If only the final net amount is examined, the moment at which the station crosses its floor may disappear. Goal names must leave clear interfaces for later calculations.
Continuing a Relationship Cannot Replace Suitable Terms
Because the client proposed talking the following week, Tang Ke wanted to keep responding rather than let the failure of Version 1 become a permanent break. Continuity of the relationship can be a goal: it preserves an entrance for future negotiation and lets both parties know which conditions failed to align.
Maintaining a relationship does not mean accepting every demand. If the station promises a period it cannot fulfill merely to appear responsive, the immediate exchange may go smoothly while later default damages the relationship. Clear boundaries can support continuity better than immediate agreement.
Nor can a relationship be represented only by the number of replies. Many messages may increase contact without increasing understanding, and an absent immediate reply does not prove a broken relationship. If "send three messages this week" becomes the relationship goal, means and outcome have already been mixed.
Gu Ning wrote the goal as "preserve an entrance for honest negotiation" and noted that current visible indicators can reflect it only partially. Whether it is achieved must later be judged by whether both parties can exchange real conditions, revise versions, and accept refusal, rather than by whether words keep appearing.
Learning Cannot Be an Unlimited Safety Net
Ye Cheng thought that organizing requirements and testing a sample might produce learning even if the station did not win the job. Such value can exist. The station might better understand how timing affects acceptance or create a check useful in future work.
But "we bought experience" readily becomes a universal explanation after an adverse result. If no one stated beforehand which unknown was to be tested, and no material was preserved during the action, claiming learning after failure leaves no learning product that can be checked.
A learning goal should be a concrete question. Can one clarification obtain the scope, delivery, and receipt conditions for Version 2? Can a small test distinguish material options? The more specific the question, the easier it is to judge whether the commitment actually reduced the relevant unknown.
Learning is not a costless bonus. Obtaining material may consume labor, testing, or communication, and the result may apply only to this version of R17. Later chapters compare information gained with resources committed. Here we require only that learning not be filled automatically by any experience whatsoever.
Preserving Capacity Can Itself Be a Goal
The four schedulable labor slots and four deployable units allow the station to respond to changes in the following week. Locking all of them into a version that has not formed would make adjustment difficult if other jobs or R17's own conditions changed.
Preserving some capacity is therefore not passive inactivity. It maintains an ability to act in the future and leaves choices available when new material arrives. This goal can conflict with winning a new commission immediately: earlier deep commitment may secure an opportunity while reducing room to switch.
Capacity matters only if the subject can actually use it. An empty bay without the necessary people or materials is not complete choice. A balance that already bears an implicit payment is not free margin. Gu Ning is using the four slots and four units explicitly defined by the case; later changes must be checked again.
Nor is more preservation always better. If capacity is never used, the station may lose a suitable job. Sustainable action must establish conditions between use and preservation instead of turning preservation into a slogan that overrides every goal.
Do Not Put Goals and Means on the Same Level
Sending a clarifying message is an action, not a final goal. Reserving a bay is also a means unless the station truly desires occupation of the bay for its own sake. When actions are written in the goal field, completed execution is mistaken for achieved purpose.
Tang Ke's sending a message on time shows that the action was executed. Whether the client supplies complete conditions, whether the relationship remains open, and whether a feasible version forms still require separate observation. Execution indicators matter, but they cannot sign for goal outcomes.
Likewise, "complete one sample test" may be a stage product. If the test was supposed to answer whether a material is suitable but no adjudication condition was set, completing the action may yield none of the needed information. Busyness can be real without automatically becoming learning.
Gu Ning created a separate column for means. Every means must point to a goal and state which goal it may harm. A means can then change without rewriting its goal, and a changed goal reveals whether an old action remains useful.
An Indicator Observes Only Part of a Goal
A goal often cannot be read at a single instant. The station may use receipt of a complete version to observe progress in negotiation, labor variance to observe fulfillment of existing work, and working capital balance to observe part of its continuing viability. These are indicators, not the whole of the goals themselves.
Treating a reply as a good relationship may miss that the reply is a refusal. Treating full bay utilization as operating success may miss rework and recovery capacity. Treating the absence of complaints as complete fulfillment may mean only that feedback has not yet arrived.
An indicator needs an observation window and a stated use. Validity for one goal does not make it portable to another. Part 1 showed that receipt of a checklist cannot replace complete confirmation; Part 2 carries the same discipline into action outcomes.
This does not demand perfect measurement for every goal. Limited indicators may be used if the record preserves what they cover, what they omit, and what new material would change the judgment. The lack of a perfect indicator does not permit choosing whichever number is most convenient.
A Constraint Is Not a Goal with a Negative Weight
Suppose the station requires at least eight units of working capital for existing payments, basic operations, and unexpected rework. This is a present survival constraint. It is not merely a heavily weighted goal called "retain funds"; every plan must first keep the available balance above this line.
If a hard constraint is only a weighted goal, sufficiently large benefits elsewhere may let a formula cross it. A high total score does not change the station's inability to bear the result. Feasibility must be checked against constraints before goals are compared among feasible plans.
A constraint also needs grounds and revision conditions. Eight units is an explicit assumption of this case, not a universal constant. The floor may be recalculated after existing bills are paid or reliable backup funding arrives; new rework may increase the amount that must be retained.
The constraint cannot be quietly lowered merely to make a favored plan feasible. It may be reassessed openly, with the new material and bearers of the consequences stated. A revisable boundary and a boundary that moves with desire are two different processes.
Multiple Goals Do Not Imply Equal Weight
Listing five goals does not give each a one-fifth weight. A count of goals supplies no weights, just as an unknown in Part 1 supplied no automatic equal probability. Income, fulfillment, relationship, learning, and capacity can have different priorities at different stages.
Some goals form a hierarchy. The station may first preserve its capacity to fulfill irrevocable existing commitments, then compare how feasible plans add new contribution. Other goals permit tradeoffs, such as communication time against additional information.
Gu Ning did not immediately calculate a total score because the goals also use different units. Units of working capital, labor slots, relationship states, and learning questions cannot be added without conversion. Giving each a score out of ten may look uniform, but the scoring scale still comes from an additional judgment.
Explicit weights or priority rules can be used later if their sources and sensitivities are disclosed. This chapter first refuses to let the symmetry of a table manufacture equal weights, leaving the substantive disagreement visible for discussion.
Goal Conflicts Must Appear Before the Outcome
Suppose a future version can increase income but requires a bay already committed to other work. If the conflict is discussed only after acceptance, the station has already committed to one goal and leaves the others to bear accomplished pressure.
Writing the conflict beforehand can prompt other paths: reducing scope, changing the period, confirming in stages, or refusing unsuitable terms. None has yet occurred in the main case. They show why a goal table must precede commitment.
A stated conflict does not guarantee a solution satisfying both sides. The station may have to surrender an opportunity or renegotiate an existing arrangement. A clear table cannot create resources, but it can keep losses from disappearing inside an ambiguous overall success.
If one goal is ultimately given priority, the record should preserve who decided, who bears the result, and when the priority will be reopened. An adverse outcome can then test whether the original priority remains reasonable instead of pretending no conflict existed.
Different Time Scales Produce Different Successes
A clarification may win no commission today but form a clear version next week. Early purchasing may accelerate the immediate response while reducing working capital for the rest of the week. Without a time scale, short- and long-term results can impersonate one another.
"Maintain the relationship" especially requires a time boundary. One courteous reply shows only this exchange, not long-term cooperation. One refusal of unsuitable terms need not damage a long-term relationship. Later material must establish concrete consequences; the word "long-term" cannot enlarge them in advance.
Gu Ning gave the initial D17 goal table two observation points: conditions obtained before Monday's scheduling decision, and resources and fulfillment at the end of the following working week. The longer relationship retains only an entrance for continuation and is not adjudicated in advance here.
Time scale also changes assessment. Looking only for a reply by Monday favors frequent contact. Looking only for income at week's end may hide a breach of the working capital floor during the process. Multiple checks let action receive its fuller consequences.
Those Who Bear Consequences Belong in the Goal Table
Income from a new job belongs to the station. Tang Ke bears most communication labor, Ye Cheng organizes materials, and scheduling changes affect clients with existing jobs. A goal table that records only what the decision-maker wants omits some subjects who bear consequences.
This does not grant every related party equal authority over every detail. It makes affected relations visible and requires authority and responsibility to be stated. Changes to existing commitments should be negotiated through the corresponding relationship; small internal communications can follow existing roles.
Gu Ning, as recorder, cannot let "preserving complete data" override every operational goal. Records serve review of action, but the cost of recording must remain proportionate. If filling every cell exhausts communication time, the table harms the work it was meant to protect.
The goal table is therefore an interface of responsibility rather than a declaration of values. It tells the three people where a plan produces results for different subjects and where later feedback must be obtained.
A Counterexample Tests Whether Goals Can Be Revisited
Consider an independent thought experiment. A team writes only "project success" beforehand. No product is delivered, but a set of internal documents is produced. The person in charge later says the goal was always learning; another says it was income. Both find support in the ambiguous phrase.
If timely delivery was stated as the primary goal and learning as secondary, the documents may count as something gained but cannot rewrite non-delivery as overall success. If the project was expressly designed to explore an unknown, the result should be assessed by whether material answering the question was obtained.
This comparison does not require goals to remain fixed forever. New material during execution may justify changing them, provided the time and grounds of the change and the treatment of old commitments are preserved. Goals can be updated; this does not mean deciding after the outcome what the past was pursuing.
Chengwan remains at a point where outcomes are unknown, making this the right time to record its goals. If D17 later produces unexpected value, that value can be recorded separately. It cannot flow backward and fill the present table with an answer that is always correct.
RC's Sustainable Decision Is Not a Total Success Score
RC understands sustainable decision-making as keeping choice present and using margin in time, space, and assessment to avoid decisive failure. In this case, that first requires identifying whose choice, the actual resources that support it, and the kind of failure that would remove the capacity to continue.
This internal position does not automatically rank the five goals or prove that preserving four units is the correct amount. Domain conditions still come from the station's commitments and resources. The theory explains why continuity of action must be examined; the case material determines how it can presently be examined.
Sustainability cannot become a total score under which "as long as we can continue, we succeeded." A plan may stay above the survival floor while repeatedly wasting resources. Another may achieve its goal while imposing an unnegotiated loss on those who bear it. Applicability still requires concrete comparison.
The multi-goal table lets RC judgment receive feedback. A result can change the assessment of income, fulfillment, learning, or capacity separately, without a smooth story swallowing the differences. The table does not make every goal equally important; it makes the choice of importance visible.
A Changed Goal Also Needs New Grounds
Goals can reasonably change during action. If the client wants only a scope description, the station may turn the original commission discussion into a paid assessment. If an existing job suddenly requires rework, preserving capacity may rise from an auxiliary goal to a present priority.
The change must correspond to new material and a time. After discovering that income is inadequate, the station cannot say it was always acting only to learn; after receiving unexpected income, it cannot say fulfillment never mattered. A new goal constrains subsequent action from the time of change, while the old goal remains in its earlier version for assessment.
Changing a goal must also deal with existing commitments. If an action obtained another party's agreement under the old goal, changing an internal table cannot make that party bear the new use. Relations that require renegotiation must be reopened explicitly, not covered by the phrase "change of direction."
Version discipline does not prevent adaptation. It distinguishes adaptation, which admits changes in reality into action, from retrospective rewriting, which merely keeps the story permanently correct.
Provisionally Determining the Goal Hierarchy
Gu Ning proposed three levels. The first contains conditions that must be maintained: do not reassign the six committed slots without explanation, do not let protected working capital fall below eight units, and do not promise that a version not yet formed can be fulfilled.
The second contains D17's primary pursuit: obtain and fulfill a new commission when its conditions are suitable, so that added commitment corresponds to a realizable return. "Suitable conditions" still awaits a later version and cannot be approved in advance by hope.
The third contains auxiliary goals: preserve an entrance for honest negotiation, obtain information about a defined unknown, and maintain capacity to adjust when new material arrives. Auxiliary does not mean unimportant. It means these goals alone cannot make a plan that crosses the first level adoptable.
If the primary goal fails, auxiliary goals may still be assessed separately. If the primary goal succeeds, the first level must still be checked. The three levels prevent one benefit from covering every consequence and prevent failure to reach an agreement from denying all learning.
Saving the Goal Table at 10:40 on Saturday
At 10:40 on Saturday, Gu Ning attached the D17 goal hierarchy to the action question card. It contains no total probability score and invents no price, process, or delivery date for a job that has not formed.
Level 1 preserves existing fulfillment and continuity of working capital. Level 2 pursues a suitably conditioned new commission and a contribution that can be realized. Level 3 preserves negotiation, learning, and capacity to adjust. Each goal has its own observation point, possible indicators, and affected subjects.
Clarifying, reserving a bay, and buying materials remain in the means column. They do not become accomplished goals merely because an action is easy to execute. The old 0.5 retains its historical identity; by this Saturday moment there is still no new client material or resource commitment.
This chapter reaches the following judgment: an action may pursue several outcomes at once, but goals, constraints, means, and indicators cannot be blended into one label of success. The next chapter connects the goal table to a consequence table, distinguishing income, direct consumption, opportunity occupation, rework, and recovery so that loss no longer appears only as "we did not reach an agreement."