In 1921 Theodor Kaluza showed that a fifth dimension, with nothing varying along it, turns Einstein’s equations into Einstein’s equations plus Maxwell’s. Oskar Klein added in 1926 that if that dimension is a small circle, electric charge comes in whole units. The corpus had a dimension ladder in Book 4 and nowhere to attach it to this history. Ten chapters now do that. Page 2 lists them.
The link to the series’ own geometry is one calculation, done by hand and labelled as untested in Lean. Write the fifth coordinate as y and form η = dy + κA. Then η∧(dη)² = κ² dy∧F∧F, so η is a contact form exactly where F∧F ≠ 0, which for a Maxwell field means E·B ≠ 0, and its Reeb field is ∂/∂y.
The other half of the week was subtraction. Page 4 lists the corrections: a Book 4 chapter that called a rank-2 system a contact manifold, a Galilean check that fails, a theorem whose hypothesis is its conclusion, and a count of “0 sorry” that hid seven results closed vacuously. It also records where the Banach theorem actually lives.
Page 5 turns from mathematics to money: a working analysis of how a town near the city might bring visitors to eat, shop, stay and come back, using one town’s attractions and public finances.
None of the Lean was rebuilt: there is no Lean toolchain on the machine used. The Reading Room, Page 6, carries Part 8 of the serial, set from its source.
Klein’s 1926 circle comes out at 8.4×10−31 cm, which is his 0.8×10−30 cm. That is the circumference; the radius is 1.34×10−31 cm, about 83 Planck lengths. A later convention gives about 23. The ratio between the two is exactly √(4π) ≈ 3.545, a choice of how charge is normalised and not a disagreement about the physics.
The chapters are committed but, at the time of writing, not yet pushed to the live site, so this issue names them without linking.
Each chapter was drafted from named sources, and the last one lists what each claim rests on. The set is in English and in a machine-assisted Portuguese draft.
| Chapter | Covers |
|---|---|
| Five Dimensions, 1921 | Kaluza’s metric and the cylinder condition; charge as momentum in the fifth direction. |
| The Circle You Cannot See, 1926 | Klein’s compact circle, charge quantisation and the size of the loop. |
| Reducing Five to Four: One Worked Case | R(5) = R(4) − ¼κ²F², checked symbolically for flat 4D with A = Bx dy. |
| Einstein Returns, 1938–1943 | Einstein and Bergmann relax the cylinder condition. |
| The Circle Fibre | Hopf, the monopole, and the third rung of the ladder. |
| Eleven Dimensions and the String Detour | Eleven-dimensional supergravity, the five string theories, and Witten 1995. |
| Thirty-Seven Dimensions That Are Not Space | The recent experiment, on the arXiv abstract and secondary coverage. |
| A Contact Condition for the Fifth Dimension | The calculation on Page 1. |
| The Ladder Meets Kaluza: Rungs 3 to 6 | Contact in 3 and 5, symplectic in 4 and 6. |
| Sources and Status Ledger | What each claim rests on, and what was not opened. |
Two limits stated in the chapters themselves. The article behind the “37 dimensions” experiment could not be fetched, so that chapter does not rest on it. And the reduction check covers one case, not the general theorem.
A contact manifold has odd dimension 2n+1, and its symplectisation has even dimension 2n+2. So the ladder alternates: contact in 3, symplectic in 4, contact in 5, symplectic in 6. Kaluza’s fifth dimension sits on a contact rung.
That has a consequence for Book 4, which had treated a 5D object with two forms as a contact manifold. It is not one: a rank-2 system of that kind is a 3-plane field in five dimensions, a Pfaffian system. The genuine 5D contact space in that setting is J1(ℝ²,ℝ), whose Darboux form dz − p1dx1 − p2dx2 equals dy + x1dp1 + x2dp2, with F = dx1∧dp1 + dx2∧dp2.
That theorem replaced a Book 4 statement that did not hold. A second finding sits in the next chapter of Book 4: the Galilean prototype (y1+vt, p1+v) does not preserve the form, since its pullback is α1 + v(dt − dx), and boosts commute. Both are marked in place with a dated box.
Not yet audited: the count of 16 invariants in that chapter (6·5/2 + 1), and whether its sixth slot is Galilean time or the symplectisation coordinate.
| Where | What changed |
|---|---|
| Book 4, chapter 5 | “Contact manifold” wording and its Theorem 5.1 replaced, as above. |
| Book 4, chapter 6 | Galilean prototype fails; marked in place. T1 correction box. |
| Book 4, recursion and hub pages | T1 note on five pages. |
| Book 5, Sorrys chapter | Nine-row table: how each “closed sorry” was closed. |
| Book 5, front page, Constants, Preface | “0 sorry” and the “banach_fixed_point” tag qualified. |
| Book 5, AXLE, G⁶, Seed, Student Edition | Correction boxes, one each. |
The result called T1. In AXLE_v6.lean (version 6.2) the theorem reads, in effect, that if the return state differs from the base point then the return state differs from the base point: the hypothesis is the conclusion. Earlier versions had a sorry. The file’s own header says that “0 sorry” does not mean every theorem is substantive.
The nine former sorrys. Read against the file’s own final status: seven were closed vacuously or by assuming the conclusion (a definition set to 0, a conclusion that is just True, a hypothesis that can never hold, a premise equal to the claim). One is real but trivial arithmetic. One was replaced by a different true statement, about the trace of a matrix rather than the trace of its sixth power.
What the repair is not. It does not show the underlying claims false. It shows that the Lean did not prove them.
Where the Banach theorem lives. A Book 5 tag names banach_fixed_point, and no file has a theorem of that name. The nearest is ExistenceWellPosedness.lean (3 July 2026, never compiled): a toy model on the real line in which the compression step is proved to contract, the other three operators are assumed non-expansive, and the fixed-point step ends in a labelled sorry. A count of “1,080 theorems” could not be reproduced: counting theorem and lemma lines gives 1,397 across 132 files, with versions overlapping. No Lean was rebuilt.
The Population Employment Gap, September. The BLS report of 2 October puts the gap (population minus employment, over population) at 40.79%, down from 40.91% in August; the unemployment rate rose a tenth to 4.2%. The fall came from fewer people outside the labour force (38.36% to 38.21% of the population), not from fewer unemployed (2.55% to 2.58%). One month, first release. These figures were read from the release’s Table A-1 and the August column matches the August values already published; the script that checks the series against FRED has not been re-run, because FRED could not be reached, and September’s U-6, 7.6% against 7.7% in August, is from BLS Table A-15. Working paper WP-128 carries the update.
The loop. A visitor who arrives for one thing and then eats, shops, stays and finds a second thing is worth far more than one who parks and leaves. Historic Hudson Valley’s figures for Kykuit visitors in 2025, as reported by The Hudson Independent: over half dined out, 40% visited another cultural attraction, 26% stayed overnight and nearly 20% shopped locally.
The gap this season. Kykuit tours, about 25,000 people a year, are paused for the 2026 season while the Pocantico Center is renovated. Those visitors are up for grabs, and the campus stays reachable.
| Place to point at now | What it offers |
|---|---|
| Rockefeller State Park Preserve | 1,632 acres, more than 45 miles of trails, free; reportedly about 200,000 visits a year |
| Pocantico Center campus, Pocantico Hills | Kykuit tours paused for 2026; a new parking lot is meant to give access to the campus and the David Rockefeller Creative Arts Center |
| Union Church of Pocantico Hills | Stained glass by Chagall and Matisse; self-guided, May to December |
| Philipsburg Manor, Sleepy Hollow | Manor and mill tours, May to November; Halloween events in Sleepy Hollow |
| Jacob Burns Film Center, Pleasantville | Opened in 2001 in a 1925 theater; three screens |
Eat here now. For where to eat, Westchester Magazine keeps a current restaurant guide: westchestermagazine.com/food/westchester-restaurants. This desk could not open the page, so it names no restaurant and makes no claim about what the guide lists.
Ideas, untested. A one-page “day in Mount Pleasant” map that pairs a trail, a window or a film with a meal. Film-and-dinner nights with Pleasantville restaurants. A “filmed here” trail built from the town’s own film permits. A campaign aimed at the Kykuit gap. None has been tested, and each should be counted before it is repeated.
Why start here, and what it will not do. Marketing costs little next to a $30 million issue. On today’s numbers it will not pay for water: the 2026 budget carries $75,000 of hotel tax, and a $30 million issue at 4% over 20 years costs about $2.2 million a year (an illustration, not a quote). But it builds the base, produces the counts nobody has, and gives lenders something to read besides the 2025 haloacetic acid violation and a budget with no water capital plan.
Read in the source. The Historic Hudson Valley site list, the 2026 adopted budget, and the Comptroller’s fiscal data (three deficit years, $12.1 million in total).
Second-hand. The 25,000 tour figure, the visitor survey and the pause, all from one undated local article that quotes no statement from Historic Hudson Valley. The 200,000 Preserve visits come from a third-party site. The county’s $2.16 billion of 2023 visitor spending comes from a state tourism body.
Not found. The number of restaurants, shops or rooms in the town; visit counts for most sites; which sites are open today; the hotel tax rate.
Not financial advice. Based on public documents read in October 2026.
It was Bertrand, in the end, who gave the thing its name — not out of malice, Fermat thought later, but because sailors were a people who needed every misfortune to have a name, a shape, something that could be blamed and thereby, however uselessly, avoided next time.
"There's a woman aboard," Bertrand said to the cook, low, at the rail, in a voice that carried exactly as far as he meant it to and no farther. "Has been the whole voyage, dressed as that clerk. I'd stake my hands on it."
He did not say it as an accusation. He said it the way a man reports weather — a fact now settled into the ship's understanding of itself, passed from Bertrand to the cook to the boatswain to every hand aboard within two days, and once it had passed, it did not need to be spoken of again. It only needed to be known, the way sailors had always known that a woman on a ship was a debt the sea would eventually collect, in its own time, at its own price.
They did not confront Fermat. They did not confront Marie. They simply began, in the small unconscious ways sailors have always managed a feared thing, to keep their distance, to touch iron when she passed, to mutter under their breath at a wave that broke wrong or a wind that turned sour, building — plank by superstitious plank — a case against her that required no evidence at all, only the accumulating coincidence of an ordinary voyage's ordinary bad days.
By month five, when the storm came up fast and ugly off a coast none of them had charted well, there was not a man aboard who needed to be told what it meant.
The reef took the hull before anyone had time to do more than shout it. Fermat felt the whole ship shudder the way a body shudders at a wound it hasn't yet understood is fatal, and then the deck was tilting, and the sea was coming up through the boards in a way the sea was never meant to.
He saw the crew move before he'd fully understood they were moving away — not toward the rigging, not toward the pumps, not toward any of the ordinary desperate tasks of men trying to save a ship, but straight to the longboat, hauling it over with a speed and a coordination that told him, even through his own rising panic, that this had been decided before the reef, not after it. They had been waiting, without quite admitting it to themselves, for the sea to do what they'd already agreed it would eventually do.
"Rope!" he shouted, and got no answer, and understood the answer was the silence itself.
Bertrand was the last one into the boat. He looked back once — not at Fermat, and not with any triumph in it, only the flat, closed look of a man who had already decided the matter was out of his hands and put there deliberately, days ago, in his own mind, so that this moment would cost him nothing when it came.
Then the oars bit the water, and the longboat pulled away into the storm, and did not look back again.
It was in that moment — the ship breaking under him, the trunk of silver sliding past in the flooding hold catching what little light there was and throwing it back in a thousand small, useless brilliances, his own hand closing on the smaller bag of coin because the trunk was already gone, already too heavy, already someone else's now — that Fermat felt something enormous surface beside him in the water, and did not, at first, understand it was mercy.
The beast was vast and grey and utterly unhurried, moving through the wreck-churned water the way nothing else in that storm was moving, without panic, without hurry, as though the sea's violence were simply weather it had decided not to take personally. He grabbed for its thick, coarse hide because there was nothing else left in the world to grab for, and felt Marie's hand close over his wrist a moment later, hauling herself toward the same broad back, and neither of them, in that first drowning minute, thought to wonder why an animal built for river mud and mangrove shallows was here at all, in open water, exactly when they needed it.
They only knew that the longboat was gone, small and receding into the rain, its oars pulling away from two people the crew had already decided, weeks before the reef ever found them, deserved whatever the sea saw fit to do.
The buffalo did not deserve any of their fear. It simply swam, steady and enormous and entirely unbothered, toward a shore none of them could yet see, carrying the curse and the calculation both, equally, without troubling itself over which was which.
Part 8 of 18 · Fiction, drawn from the documented world; the ledger below says which is which. Set from its source file, as written.
Rights notice. Pierre et Mademoiselle is not covered by this issue’s CC BY 4.0 licence. The serial and the work it belongs to — a limited series in development — are © 2026 Pablo Nogueira Grossi / G6 LLC, all rights reserved. No licence to reproduce, adapt, translate or perform it is granted here.
Documented, as the series ledger (FACT_AND_FICTION.md) records it; not independently re-checked here. The Marajó buffalo legend: around 1890 a cargo ship bound for French Guiana, carrying water buffalo, was wrecked on the sandbanks near Cabo Maguari, and the surviving animals swam ashore and populated the island. Importations followed in 1902 and 1906. The route La Rochelle, Tortuga, the Guiana coast, Maranhão was a functioning French Atlantic corridor in these years.
Invented, per the same ledger: the three ships of buffalo in 1666, which it calls a relocation and not a fabrication, since the real animals arrive by shipwreck 224 years later; and the voyage itself, Bertrand, the crew, the reef and the abandonment. The ledger flags that this Bertrand shares a name with Bertrand d’Ogeron, the governor of Tortuga; they are not the same person.
Each issue, when it goes out. No list bought, harvested, or lifted from a conference roster. Unsubscribe by replying with the word stop.
tools/corpus_count.py · tools/self_reference.py — entity-aware counting, and the vicious-circle test. Both standard library only.sorry, standard axioms only.