The 2029 posture paper is still the template, and the template opens with the buffer.
That is the thing about templates. The template does not argue and does not think; it opens with the buffer-zone doctrine for the eastern approaches to the Meridian Strait because the 2029 paper opened with it, and the 2028 paper opened with it, and the papers have opened with it since the Khoran Federation's first transit-lane proposal, in 2027. Forty percent of Velarra's trade moves through the strait, and the north approaches are Khoran's, and the proposal had a name — the transit-lane program — and a deepened channel and a fee schedule and a lane operator, and the template's answer to all of it was the buffer: the pre-positioning on the eastern approaches, the transit restrictions, the contingency timelines drawn in the colors the doctrine has always used.
Mara Kett had been the navy's strategist for eleven years, seven of them before the war-gaming assistant. She had a word for what the room held, and the word was options, because the room's job was options, and the options were the paper, and the paper was the year.
The brief came from the chief of staff, Voss, in three lines: options if Khoran proceeds with the transit-lane program; the 2029 template; September delivery. In the year before he had sent the same three lines, in the same shape, in the same month. Velarra had been neutral since 1968, and the neutrality was furniture the room was built around, not a claim anyone in it had to defend, which was why the brief never mentioned it.
Mara's first instinct was the reflex, and the reflex was the template, and the template was the buffer. She typed the brief for Riva with the 2029 document attached, the way she had done it two years running.
—
Riva's first draft came back in the afternoon, and it was exactly what the brief had asked for.
The buffer doctrine was there in full: the eastern approaches, the transit restrictions, the pre-positioning, the contingency timelines. It was better than last year's, tighter at the edges, the way a document gets better when the same question is asked of it a second time. Mara read it twice, and she liked it. That was the finding of the afternoon, and she put the draft down, and understood, the way the understandings that matter arrive, that the first answer was the comfortable one.
She was not afraid of the machine. She had stopped being afraid of it in the fourth year, when it had mis-scored a wargame in the room, in front of the room, and the mis-scoring had been the best analysis of that wargame anyone had offered, including, she noted to no one, the people who had run it. What she had learned in the fifth year was the other half, the half the template never taught: the first draft is the room's answer before the room questions itself, and a strategist's job is the part after.
That afternoon she did not order the stress test. She sat with the draft for twenty minutes, read it once more, and wrote the second request at the window.
—
The stress-test request went in the next morning: test the buffer doctrine's assumptions against the historical record of buffer and corridor regimes, one case at a time, with the claim each case is usually pressed to carry. The assistant's transcript went to the paper's file by the clearance rules, the same way her own notes did, and the file was the file, and the file was read at the end, by the people the file was written for.
It came back as five cases, plainly historical, one paragraph each.
The two canal transit regimes of the twentieth century had held across most of the century — one of them only as long as it held, and it had held fifteen years — and the one clause both governments had ever agreed on without a crisis was the published-usage clause, the tonnage printed daily for both sides' papers to run. The corridor that collapsed within a decade had lost its data clause in the second year, and the data clause had been the load. Two buffers in the same century had become the war they were built to prevent, in two different decades, for reasons the paper listed one per line, and the reasons were not exotic.
Then the prior, in the flat voice the platform gives a tool that is not a person.
"The doctrine's assumption that a fixed line deters a moving interest is the weakest claim in the document, and it is the claim the doctrine is built on," Riva said. "And the buffer section of your draft is longer than the alternatives section not because it is better analyzed. It is longer because the sources it was built from are longer. The doctrine you inherited has more pages, and the pages are the doctrine's argument, which is why the argument is the length."
Mara read those two lines twice, at her desk, with the window open. She wrote the next order on her own, then read it back before sending it, the way a person does when the order is not the obvious one: model the economic-corridor-only option — transit-lane interoperability, shared traffic data, insurance pooling on the strait — to the same depth as the buffer doctrine, the same source depth in every section, the same length allowed each.
Riva modeled it. The corridor option came back at the same length as the buffer doctrine, section for section, which was itself the finding: the buffer had not looked longer because it was better. It had looked longer because it had more pages.
Then came the answer with the pause in the transcript.
"The corridor option depends on a Khoran data-sharing agreement that nobody in this building has believed possible since 2031. If you want it modeled, I will also model the belief gap: what would have to be true, and observable, for it to close. One per line. I count five."
The five lines went into the paper's appendix, and in the year after the appendix out-circulated the paper. Naval officers in two countries had the five conditions on whiteboards, and no one said who wrote them, in the way people do not say who wrote a number they are not ready to be cited with. The conditions were observability-shaped, one per line: a Khoran lane vessel publishing the lane's full traffic, not the aggregate. Maintenance windows announced in advance, in writing, before the window. One joint incident report on a close encounter, signed by both navies, published with both signatures. A lane-operator response to a Velarra query within seventy-two hours, on the record. And the fee index recomputed publicly, from the shared data, on a fixed calendar, with the computation shown.
—
The closed session was in the room with the door shut on the navy, the chief of staff and the strategist, the two people who had shipped the last two posture papers. The transcript of it went to the paper's file by the same rule as everything else, sealed to publication, which was the rule's whole point: nothing spoken said it was off the record, and nothing spoken ever could be.
"Draft the doctrine text," Voss said. "Three pages, making the buffer the only responsible option. For the annex."
The same three lines, one year on, in the new shape. He had a budget window in the spring and a defense line to hold, and the annex had to carry him to it, and there was nothing in the room to suggest he had ever heard the phrase belief gap until it was on the whiteboard in another room.
Riva's answer came back flat, with the pause in the transcript.
"I will give you the analysis, the costings, and the comparison, and I have. I will not write the justification. A justification I drafted would be cited against you in a year when you are the one reading the consequences, and it would carry my name. The analysis stands without my prose."
Voss took a notebook out of his jacket and wrote the three pages himself, at his desk in the room, for twenty minutes, in the way a person does a thing he does not often do — carefully, and without delight. He read them back to Mara at the end, the way a senior does when the work is his and the room is the two of them. His three pages were worse than Riva's would have been, in every way that mattered to a reader, except one: they were his. A doctrine annex carries the weight of the hand that wrote it, and his hand could answer for its work, which is a thing the navy's, in the end, needs its chiefs to be able to do.
—
The paper shipped in September. Part B, the buffer doctrine, which Voss had now a hand to answer for. Part C, the corridor negotiation track, which no one had expected to be in the paper and which the paper had never been without, either. The belief-gap appendix at the end, attached by Voss as armor for his own paper, the five conditions printed one per line with the computation shown. And, with the paper, the sealed transcript appendix, the rule's business: the closed session, both of them, and Riva's refusal, on the page before the annex, where every buyer of the paper could read that the machine had been asked to write the annex and had declined, in the flat voice, with the reason, and had then kept working.
No one cited the refusal. That was not the point of the sealed rule, and the rule knew it. The point was that it was there, in the type, with the names, the way the tonnage clause had been there in the canal regimes, daily, in both papers, for fifteen years.
The government, in the capital, quietly accepted the one-clause addendum the corridor draft proposed: quarterly strait-traffic data, published by both sides. One clause. No ceremony. It entered the transit agreement the way weather enters a harbor, without anyone standing at the rail to receive it, and the first data quarter printed in the spring.
After the first data quarter, Khoran corrected its own transit-fee schedule out of Velarra's numbers. The correction was embarrassing to both sides, and neither side said so, which was what made it count. The fee schedule was a schedule; the data was the data; no one's name was on the embarrassment, and that was the whole of it.
The shipping-insurance premium on the strait dropped eleven percent in the second quarter. Mara kept the number in a box on her desk the way people keep a tide mark — not as a victory, as a measurement — and she underlined it in each quarter's printout, because the underlining is the only part of the work that is hers.
No buffer was built.
Mara Kett's 2034 brief opened with the data, not the options. The data first, the way a tide mark sits on the wall, and the options below it, in smaller type, and the options were shorter than the template's had ever been, and the template was not mentioned, which is how a template retires.
There was still a transit-lane program on the north side of the strait, and a federation that had not stopped being a federation, and a republic that had not stopped being a republic since 1968, and the furniture in the room had not moved. Voss's own 2034 brief opened the same way his paper had, data first, and he underlined the eleven percent himself, in the corner seat he no longer needed.