7.1When does an open record make peace credible?
The question cannot be answered from these records in the sense of outcomes. What the readings support is a narrower claim about design. Kant forbids a hidden reservation, allows a declared postponement bounded by its purpose, and makes publicity a test that the first fails and the second can pass, while denying that passing it is enough.
The two records build a reserve that fits the second category. In both, the judgments withhold material: regional figures in the first, the evidence itself in the second. In both, the withholding is published, with its reason, its loss, the party who bears the loss, a review that bounds it, and a rule that the review's date does not decide. In both, the judgments preserve the objection that the reserve may conceal: "under the pretext of protecting personal information", in the first; "the committee controls the verifiers and the issues", in the second.
On this reading, an open record contributes to the credibility of a peace when it does three things: it renounces, rather than hides, what it holds back; it publishes the holding back in a form that would defeat a hidden purpose; and it bounds the holding back by a review that must be argued, not merely reached. These are conditions of design. Whether they are sufficient is the question Kant's own limit on publicity leaves open, and the records preserve that question rather than closing it.
7.2The two sides of the debate, read again
Fortna's note lists verification among the elements of a stronger agreement. The first record shows that verification produces a second decision: what to make of its results. The choice of unit has costs that the note does not mention. This does not contradict the note; it adds a step between verification and the durability the note links to it.
Walter's note places the danger at the disarmament point and its remedy in guarantees and power-sharing. Neither record contains a disarmament point. The first, though, is about an arms reduction, and its third stake is the trust of participating institutions in the next round of verification, which the third premise says depends on "the predictability of the publication method". That is a commitment concern: participants want to know in advance how what they reveal will be used. The open reserve answers it by publishing the method and its conditions for change before the deadline, a commitment of the kind Schelling's note describes as binding its maker, made public here by the record. The record cannot show whether it would bind.
Fearon's note names both mechanisms. Kant's first article, as read in Section 4, joins them: the secret reservation is private information whose harm is a failure of commitment. The open reserve is the converse. It keeps material private, but it makes the keeping public, and so turns what could have been a hidden reservation into a stated commitment about when and how the material will be released or kept.
7.3Kant's limit, and the keeper of the reserve
The most important finding of the reading is negative. Kant's test of publicity cannot by itself distinguish an honest reserve from a pretext. He says so: "we cannot conclude that the converse holds, and that all maxims which can bear publicity are therefore just. For anyone who has a decided supremacy has no need to make any secret about his maxims." Both records contain a keeper of the reserve: the verification data committee, which holds the regional figures, and the public verification committee, which approves access to the evidence. Both keepers publish their maxims. Neither publication proves good faith.
The records respond in the only way a document can: by binding the keeper. They require that the continuation of the reserve be argued afresh at a fixed point; that the reasons for not reflecting a party's explanation be recorded; that errors be sorted by source, including the keeper's own handling; that agreement among verifiers not substitute for evidence. Each is a written plan, and Clarke's note concerns plans that could not work as written but persuade others that the risk is under control. The records cannot show which of these would be kept.
7.4What the evidence can and cannot carry
The evidence is two records out of fifty, each written by AI about a fictional polity, each recording a decision before its outcome. Several limits follow.
No peace outcome is recorded. Neither record says whether a treaty zone remained at peace, whether participants returned for the next verification, or whether the shelter dispute was resolved. The article's question can be answered from them only as a question about what such a record is designed to say.
The judgments do not part. In both records, all thirteen judgments choose the same option; in the parent study's ledger, both records show one distinct option among thirteen. Across the collection, judgments chose more than one option in 13 of 50 cases. The records state that all the judgments use the same input and model; unanimity says something about how they were produced, not about the merits of the option.
The question form guarantees part of what the article reads. Every Meridian question carries a decider, a deadline and conditions for reopening, because the form asks for them; the ledger finds conditions for revisiting in all 50 cases. The presence of a schedule is not a finding; its content, such as the refusal to let a date decide, is what the article reads.
The ledger was not built to measure openness. In plain words, its two readers disagreed about the decider in both records (one called the committee a team or organisation, the other an institution or state), and about who bears the cost of waiting in both. Across the collection, the readers agreed on the decider in 28 of 50 cases (56%), the lowest agreement of any cell. The program that flags reopening conditions stated with a number, a threshold or a time flagged one of the three decision-brief conditions in the first record and none in the second, although several of the unflagged conditions are scheduled weekly or daily; the flag is a coarse instrument for the question this article asks. The appendix gives the cells in the ledger's own terms.
Where the article goes beyond a reading note, as in placing Kant on both sides of the information-commitment debate, the step is its own. The records' English is a translation, with the defect noted in Section 6.6.
7.5Other readings
Other readings remain open. First, a reader closer to Walter's note could hold that the open reserve is beside the point: what makes a settlement hold is a guarantor, and no amount of careful publication substitutes for one. Second, a reader closer to Jervis's note could hold that even a perfect open reserve will be read through prior beliefs, so that civil society groups will see concealment and participating institutions will see prudence whatever the record says. The records preserve both in their counterpositions without settling them. Third, a reader of Kant could hold that the first article's strictness rules out any reserve in a peace document, open or not. The reading offered here rests on Kant's own distinction between strict articles and those whose enforcement may be put off, and on the observation that the reserve in these records concerns publication of verification results, not the terms of a peace. A reader who rejects that distinction will read the records differently.