Where the Timelines Agree

You cannot feel the fork and you cannot steer between histories. What survives is not the route that scores highest — it is the one its neighbors agree with, and someone has already walked it.

September 2, 2026

A dark landscape crossed by countless faint paths with translucent figures walking each of them, and one solid figure standing on the single lit path that runs to the sunrise

I. The Fork You Did Not Notice

The Highest Peak left you one mercy. You were standing on a mountain range at night; you could not see the range and did not know which summit was highest, but the slope under your two feet was real information, honestly delivered. Up, right here was at least something.

This essay takes that ground away, and it does it with a sentence you have said a hundred times without meaning much by it: things could have gone otherwise.

Take it literally, for the length of an essay. Not as a wistful remark about a decision you regret — as a description of how the world is furnished. Where things could have gone two ways, let them have gone both, each way running on into a whole world of its own, complete and consistent and populated, and you standing inside exactly one of them.

You don't have to buy that as physics, and very little of it is physics. The argument below runs on the weak version and gets everything it needs from it: that things could have gone otherwise, and that nothing available to you from in here settles which way they went.

That is a fork — not a decision you are facing, but a place where the world already went both ways. You are standing at one now. And there is nothing it feels like.

The argument is three lines long.

At a fork, every branch inherits the same past. Your memory is a function of that past. Therefore no memory, no measurement and no sensation taken at the fork can distinguish which side of it you are on.

That is all. There is nothing to feel, and there was never going to be. The continuity of your life is not evidence that only one life happened; continuity is what every branch has. You didn't notice this morning. Nobody has ever noticed.

And the plain version is worse than the formal one, because you already know it. The forks that decided everything in your life were unmarked at the time. The train you ran for. The room you happened to sit down in. The message you drafted at one in the morning and did not send. None of them announced itself. Sliding Doors built a whole film on one — a woman makes or misses a closing Tube door, and the two lives run on side by side.1 What is worth noticing is the technique rather than the premise: the film has to cut between the two, because cutting is the only way anybody could ever watch both. Helen cannot. Nobody in either life can. You have never once had the experience of choosing between timelines. You have had the experience of an ordinary Tuesday, and the ordinary Tuesday is the mechanism.

The mathematics here is modest. In any system whose parts feed back into each other, two histories that begin almost identically pull apart — and they pull apart fast enough that there is a wall past which prediction simply stops. A weather forecast is good for days and worthless for months, and no amount of care or computing power moves that wall very far.2

Now set that against the epistemics the companion ended on, because it changes them.

It said guidance can only ever take one form — the next step, no explanation — and found the tradition's own statement of it: Thy word is a lamp unto my feet, and a light unto my path.3 A lamp for the feet, not a floodlight over the country. True, and now worse: the country is dark, and it is one of a great many, and you are being handed from one to the next all day without an announcement.

Which makes one detail in an old verse look less like poetry and more like specification. Isaiah does not say the voice arrives on schedule, or in the morning, or when you ask; he says: thine ears shall hear a word behind thee, saying, This is the way, walk ye in it, when ye turn to the right hand, and when ye turn to the left.4

The voice speaks at the fork. It is the only place a guide could usefully speak. It is also, precisely, the place you are not looking.

II. The Index That Is Not a Direction

The picture §I asked you to hold needs one turn of precision, because the rest of the essay leans on it and it is easy to smuggle.

A set of histories, each of them a whole world running its own course, and each one carrying a label. That is the whole construction: not a new dimension you could travel, but a way of saying which.

Borges got there first and drew it as a garden. In his story a Chinese governor abandons office to write a novel and build a labyrinth, and everybody assumes these are two projects; they are one. The novel is the labyrinth, and its trick is that at every fork, the character takes all of them — so the book contradicts itself on every page, and the labyrinth turns out to be made not of hedges but of time.

Two things about that object, before a reader with any physics gets there first.

One: I am not claiming this is established physics, and I am not going to pretend it is. The forking, if it is anywhere in the science, is in quantum mechanics — in decoherence, in Everett's reading of it — or else it is simply the plain fact that things could have gone otherwise, which needs no physics at all.5 The label is my bookkeeping, and I'd rather say so than be caught at it.

Two, and this one is a result rather than a preference: the label is an index, not a direction. It's tempting to promote it — to make which history a real dimension you could move along, the way you move along a road. Try it and physics stops working. Give the world a second dimension of time and you get loops in which an effect arrives before its cause, and you lose the one thing that makes physics physics: the ability to say what happens next from what is happening now. Nothing follows from anything. That is not a strange universe. It is no universe.6

And the fifth dimension physics actually has is a different one: Kaluza and Klein got gravity and electromagnetism out of a single five-dimensional theory in the 1920s, and their extra dimension is a circle rolled up smaller than an atom.7 The trade already spent it on something else.

So there is no steering. There is no toward in the index. You can only walk the history you are in. I will come back to this in §VI, where it does more work than any pious sentence could.

The argument proves something narrower than the sentence it is usually asked to carry. Navigation needs local structure — something at the place where you are standing that counts as toward, a gradient you could read and act on. The index has none. No history in it is nearer or further than any other; there is nothing to read and nothing to follow. So there is no route out of this history into a preferable one, and no method that could find one.

That is a claim about routes. It is not a claim about what fixes a person to a history in the first place, and it is not the claim that nothing whatsoever could ever put you in a different one. Those are different questions, and real ones, and this essay does not touch them.

When I say which history you are in, I'm treating that as a fact about the world and not about you — as though two people could stand in the same history, read it in completely opposite ways, and still be standing in the same one. That may be wrong. It may be that how a life is read is part of what makes it that life. I am assuming the two come apart, and I have not shown it.

And one last thing about Borges. Everybody remembers the forking. The sentence he actually wrote was "a growing, dizzying net of divergent, convergent and parallel times."8 Convergent. Paths that part and come back.

Everything so far has been about the parting — the unmarked fork, the label, the history you cannot steer out of. Nobody draws the other half of the picture: two histories that separated, ran a long time apart, and arrive at the same afternoon.

Whether that arrival is a homecoming or an annihilation is not a question about where either of them has been. It is a question about what the two of them are carrying when they get there.

III. All the Paths, and Why You See One

Here is the movement the whole essay is built to reach, and it begins with a correction.

The companion leaned on Fermat's principle: light takes what physicists call the extremal path — the route fixed by the whole journey rather than groped out a step at a time. It is a beautiful fact and it is the shallow version of the truth. Light does not choose a path. It takes all of them. And the one you see is the one where the others agree.

Feynman drew this for people with no mathematics at all, and the picture is worth more than the formalism. Give every possible route from here to there a little arrow, spinning as it travels, its final direction set by how long the trip took. To find out what the light actually does, lay every arrow end to end and see what is left over. Near the extremal path, neighboring routes take almost exactly the same time — so their arrows point almost the same way, and they add. Everywhere else, neighboring routes disagree, the arrows scatter in every direction, and they cancel to nothing.9

What survives is not a decision. It is an agreement.

And this is what the phrase every physicist says without thinking actually means. "The variation is zero" is not mystical and — the companion was right about this and it matters here — it is not a minimum. It means: along this path, the neighboring paths agree with it. That is the definition. The extremal route is the robust one, the one whose alternatives do not argue with it.

Before I collect on that, I have to hand over the crack in my own foundation, because a reader who knows will find it and I would rather be the one who says it.

Nobody has ever made that sum mathematically legitimate. Adding up infinitely many arrows is the kind of operation that needs a careful definition before it means anything, and for a world running in ordinary forward time, that definition does not exist. What mathematicians do instead is tilt the problem sideways into a related one where the sum is well defined, do the work there, and tilt the answer back — a maneuver that succeeds spectacularly, underwrites an enormous amount of twentieth-century physics, and is not, in general, a proof.10 I am building on it anyway, and I would rather tell you the floor is thin than let you find out by standing on it.

Now the sentence — and a boundary first, because without the boundary the sentence is a cheat.

Nothing in any of this says that the path light takes is a good path. The agreement condition picks out what happens; it has no opinion whatever about what ought to. Let the word ideal walk across that gap on the strength of a resemblance and I have won by sleight, which is worse than losing the sentence. So here is exactly what is borrowed and exactly what is not.

What is borrowed is a form of criterion. Physics turns out to possess a way of specifying a whole path without scoring any point along it — not how good is it here, but does this route agree with its own alternatives — and a test of that shape can pick one route out of more routes than could ever be counted, without once consulting a moment. That is the entire loan. It tells you what kind of thing the LORD's will could be. It tells you nothing about what They will, and it never could; that question was never going to be answered by a sum.

The companion called the goal the highest peak — a height, over a landscape, at a place. That was a number attached to a point, and it was always very slightly the wrong object, because the thing being ranked is not a place. It is a whole history. With the boundary in hand, here is the better object:

The ideal timeline is not the history that scores highest. It is the history along which the neighboring histories agree. Every other route is one whose own alternatives cancel it out.

That is a claim about God, in a grammar physics happened to build first. The loan is the shape of the sentence. The content is three thousand years old.

And it explains, in passing, why you only ever see one world. The cancellation is merciless. At our scale, in bodies this size, and with the grain of the quantum world as fine as it is next to anything we are made of, a world of countless branches looks from the inside like a world with exactly one past — and that appearance is not a lie being told to you. It is what overwhelming disagreement looks like from inside.11 The singleness of your life is not a fact about the universe's inventory. It is a fact about your address in it.

IV. Why It Was Never a Rulebook

Two things fall out of that sentence, and the rest of the essay lives on both.

The first is the companion's thesis, said once more and better. A rule is a verdict on an act: a number attached to a point. A condition of agreement is a property of a whole route, and there is no operation you can perform at a point that evaluates it. That is why the LORD's will could never have been a rulebook, and it is not a devotional claim — it is a statement about what kind of object is being described. A rulebook is what a route looks like to someone who can only ever stand in one place.

The second is that local self-interest is not merely selfish. It is wrong about the answer.

When a problem couples people together, everyone improving their own position lands the whole group somewhere provably no better than the best available, and routinely worse — including, routinely, worse for each person taken alone. The famous case is Braess's paradox: add a road to a network and, with every driver taking whatever is fastest for himself, everybody's journey gets longer. Not a fable with a moral attached. A theorem, with traffic engineers' names on it.12

What closes that gap is never that somebody gets cleverer. Across the whole study of how many local actors solve a problem none of them can see on their own, the answer comes back the same shape every time: they tell each other what they can see. The connection between them is not a nicety in the system. It is the term that fixes it.13 I am aware this is an analogy and I am naming it as one in the same breath — nobody is claiming that love is a routing algorithm. The claim is the smaller one and it is enough: local self-interest and the global good provably come apart on problems like ours, and what closes the distance is the traffic between neighbors.

And the coupling is not optional, which is the beat that keeps this from being advice. Everything you do enters the future of everything it can reach — a region that widens with every hour and eventually contains everyone. This is The Highest Peak's storm falling on the sailors, and I will not re-argue it.14 You are already coupled to every person in your future. Love is not the decision to be connected. It is what you do about a connection you never had the option to refuse.

The tradition says this, and — this is the part that stopped me — it says it as a weighting. Thou shalt love thy neighbour as thyself.15 Not instead of. Not more than. As. That is an instruction about how to weight the terms, and it is precisely the instruction a distributed problem requires. Bear ye one another's burdens is the same sentence in the imperative.16

One limit, held from the companion and not negotiable: this pins the weights, not the content of the good. I am still not going to tell you what the objective is. I don't know, and an essay that claimed to would have installed the very thing its last movement takes down.17

V. The Route Runs Through the Bad Ones

The claim of this movement is the one everything so far has been clearing ground for: God is asking us to love toward the ideal timeline, and the route passes through timelines that do not look good.

It gets two supports, because it deserves better than assertion.

The first is arithmetic. There is a founding theorem about best routes, and it says this: cut the best route anywhere you like, and what remains is still the best way to the end from wherever the cut left you. Whatever has already happened, from here on the best way is still the best way.18 It's a lovely result, and it is exactly half of what people think it says. The same is not true of the beginning, and it cannot be. The first stretch of the best available route is routinely not the best-looking first stretch — if it were, taking the best-looking step every time would solve everything, and nothing would ever be hard.

The companion said this about a valley between two mountains: the base of the highest peak is reached through the low ground, and there is a stretch in the middle where you are lower than you have ever been and the summit is not in view. Said about whole histories, the same claim is colder and larger. The ideal timeline runs through stretches that any honest accounting of the moment would score as losses. Not as a mystery. As a consequence of what agreement over a whole is.

The second support is the one this essay is here to add. It is not that your sense of things is unreliable. It is that the criterion has no local form at all.

Go back to the arrows. There is no answer to the question how is this route doing, here? The question is not hard; it is malformed. An arrow at a point has a direction and nothing else — no score, no verdict, no partial credit — and what decides whether the route survives is what every other route is doing, all the way along. You are not looking at a bad reading. You are looking at a place where the instrument does not read.

And the tradition has a whole day set aside for this.

Between the Friday and the Sunday there is a Saturday. On that day it is over. The stone is in place, the followers have scattered, the movement is finished, and no honest observer could call it anything but a defeat. The church did not delete that day. It kept it, and put it in the calendar, and on it does almost nothing — no Eucharist, a stripped altar, a held silence.19 An entire liturgical day given to the condition of being exactly, precisely on the route, at a moment that reads as total loss.

No one has ever been more on the path than the man in that tomb.

Now the limit, and it is the whole difference between this essay and a cruelty.

This is not the promise that it will look good later. That is the tidy theodicy — the neat reason, the redemptive arc handed to someone who did not ask for one — and the companion declined it on principle and so do I. The claim is smaller and harder to hold: how a moment reads is not evidence about the route. It isn't evidence against it either. A reading is not a route. Neither reading.

Which is also how this squares with what I have already published, and I'd rather square it in the open than hope nobody checks. Flow argued that the decisionless state is neither good nor bad in itself: neither how it feels nor how it turns out is a readout of whether you are where you should be. This essay must therefore refuse the inference in both directions, and does. A terrible stretch does not prove you strayed. A beautiful one does not prove you were led.

VI. Three Ways This Becomes Cruel

Every one of these follows easily from what I have written, and each is worse than the error it corrects. The companion had one of these to refuse. This one has three, because a set of histories is a much better hiding place than a mountain.

(a) "If every outcome happens somewhere, then nothing matters — and somewhere, your child lived."

It is refused three times.

You are not everywhere. The person grieving is on this branch, in this history, and the loss is a fact of the only life they are in. "Somewhere else it went differently" is a sentence about a place they will never be, and it does not reach them, and it was never going to.

The existence of an alternative has never once diminished an actual. However many there are, a count of histories is not a moral theory. This is not a subtle point. It is the difference between arithmetic and a person.

And this essay's own mathematics forbids it. The index is not a direction. There is nowhere you can be sent and nothing that can be shipped back. Whatever else is true of the other histories, they are not addresses and they are not consolations. Nothing is going to arrive from one, and nothing you do will reach one. That histories can come back together — which is where §VII ends — is not a fact about anywhere you could go.

Then the plainest sentence in the piece: never say this to a grieving person. It is not a comfort. It is a way of not being in the room. The instruction for that room has been available for two thousand years and has no cosmology in it whatsoever — weep with them that weep.20

And the companion's refusal carries forward, because it is now load-bearing twice: deviation causes suffering does not entail suffering implies deviation. That inference is what Job's friends make for thirty chapters and it is the one thing in the book God says is wrong; it is what the disciples ask about the blind man, and the answer is one word.21

(b) "Then shift to a better timeline."

There is a whole genre of this — jump timelines, choose your reality, align to the version of your life where it worked out — and the wish underneath it isn't contemptible. It is right. The wish is that the world is not a cage, and the world is not a cage.

But the method is precisely the error the companion named: the small self appointing itself the navigator, now with a whole set of histories to do it in, which makes it grander and no truer.

What fails is navigation — a self setting a course toward a history it would prefer, by technique. Something real stands next to that, and the refusal is not pointed at it; §VII takes up what may be left when the navigator is gone.

Is there a technique in it? If there is a method — a practice, a sequence, a thing to do on purpose to get somewhere better — the navigator is back wearing different clothes, and everything below applies to it.

And on steering, the mathematics answers before the theology needs to. Make the index a direction you could travel and prediction collapses: the price of a world you could navigate sideways is a world in which nothing follows from anything. What that rules out is the index having any local structure — and navigation is nothing but local structure used twice, a toward you can read where you stand and a heading you can hold once you have read it. There is no toward. Nothing in the index is nearer or further than anything else.

A jump is not ruled out by anything I have said. A journey is. And every offer in the genre above is a journey — a course set, a heading held, a method for getting closer. That is what fails, and it fails for the flattest possible reason: there is nothing to get closer to, in a direction that does not exist, by a gradient that is not there.

And failing is the gentler half of it. A course held toward a heading that is not there does not leave you where you started. It runs you aground. The life goes into the steering — reading the day for omens, forcing alignments, grading every ordinary Tuesday against a place you invented and finding it wanting — and the grading is not a way of getting somewhere. It is a way of being unable to be anywhere. I am not claiming a mechanism for that and I don't need one; anybody who has watched someone do it for a decade has the evidence.

The only motion you have ever been able to make was the next step, and it was always in this history.

(c) "Then my choices are theater, since every version gets taken anyway."

The companion answered the determinist form of this and I will not re-argue it. The branching form is answered faster. That a version of you chose otherwise is bookkeeping about a set; it isn't a fact about your will. You are the one here. Freedom, as the novelist the companion leaned on put it, is perfectly real in the context of a consciousness that runs in order — and ours does, all the way down, without exception, forever.22

VII. Set It Down and Walk

The Highest Peak ended by refusing to install anything in the space it had cleared. It broke the right image on the way out — the kingdom of heaven is not a mountaintop — which was the right image to break precisely because it had taught the reader to want one.

I have to break my own, and it's this: the kingdom of heaven is not a coordinate.

Not a better branch. Not somewhere you could get to if you found the door, or the technique, or the frequency. The whole apparatus was a finger and not the moon — useful exactly as far as it got us here, and no further. Do not install the multiverse where the rulebook used to be. I have spent an essay building a structure I now have to set down, and I'd rather set it down myself than watch somebody move into it.

What survives is small enough to carry.

You cannot see the fork. You cannot steer between histories. You cannot grade the day you are having against a route you will never see whole. What you have is one step, and the people you are already coupled to, and a criterion that was never going to be a rulebook because it was never a verdict on a moment.

The ideal timeline is the one along which the neighboring histories agree. Lives that love each other are lives that agree — not in opinion; in phase, and they reinforce where they would otherwise cancel. The step from histories to lives is an analogy and I am naming it as one: two lives are not two histories. So loving toward the ideal timeline turns out not to be a technique for selecting branches at all. It is the only local act with global reach, and it's available to you this afternoon, in this history, to the person in front of you.

One more thing, and then I am done with the mathematics for good.

Go back to the arrows a last time. Every single one of them is a route that left the same place and is on its way back to the same place — that is what the sum is. The parting was never the interesting part. The arriving is. And when two of them arrive together, nothing decides what happens except the angle between them: line them up and you get something brighter than either one alone; set them against each other and they put each other out.

So the merge was never a thing to hope for against the odds. The merge is the whole subject. It is what the sum is a sum over. What is never guaranteed is which of the two kinds it turns out to be.

And every account the tradition gives of reconciliation is an account of a merge that came out the first way.

Which leaves the last thing, and it is not an instrument and not a theorem.

There is a route that the whole journey picks out, that no step of it could be graded on, and that nobody standing at a point could ever identify. He said He was that. Not a guide to it, not a map of it — I am the way. And He walked it straight into the worst-looking stretch in the whole book and came out through the far side, and it did not change what He says. Two words, at whatever volume the day requires, never once as a verdict on the one hearing them.

There is a version of shift to a better timeline that runs the other way from the genre, and with the method gone there is nothing left in the room for it to be confused with. I am not refusing it. It is not a self steering; it is a self letting go — the suspicion that what holds you where you are is your own insistence on a consistent account of how you got here, and that the insistence is doing more of the work than the world is. That is a claim about what fastens a person to a history rather than about how to move between them, and it stands nearer to what the tradition says about losing a life in order to find one than to anything in the genre above.23 I don't know whether it is true.

But whatever that motion is, it could only ever be come to. Never administered, never prescribed, never handed to one person by another as advice. The voice in §I speaks when ye turn; it does not turn you. Anything you could hand to somebody in grief is not that motion — it is §VI's first cruelty, wearing better clothes.

And if anything ever takes a person out of the history they are in and into a better one — nothing in this essay establishes that, and I am not going to pretend otherwise on the last page — it could only have one shape. Not a heading held. An emptying, a guide accepted, and the arriving given rather than got. Which is why steering runs a life aground and surrender does not, and why the one motion that might be real is the one nobody can perform on purpose. You cannot aim at grace. That is very close to what the word means.

He is still a few steps ahead. He is still looking back over His shoulder to see whether you are coming.

Footnotes

  1. Sliding Doors (1998), written and directed by Peter Howitt; Gwyneth Paltrow plays Helen Quilley. The film rewinds the moment at the train door and replays it with her aboard, then alternates between the two resulting lives. I am taking its form and not its plot: the intercutting is a viewer's privilege, and the claim here is about what is available to the person inside. The two lives do brush against each other at points; that is a storyteller's license and I am not leaning on it — whether histories rejoin is a question about the sum in §III and the close in §VII, and how a director chose to cut is no evidence either way. ↩

  2. Edward Lorenz, "Deterministic Nonperiodic Flow," J. Atmos. Sci. 20 (1963), 130–141. One correction worth making in the open, because the misattribution is nearly universal: Lorenz's 1963 paper does not establish the two-week figure. It shows that a deterministic system can be non-periodic and sensitively dependent, which gives a bounded horizon without fixing its length. The familiar two-week number comes from later work on real atmospheric models — Charney et al. (1966), resting on an observed error-doubling time of about five days. The body accordingly cites no number, and says bounded horizon, never "chaos means anything can happen." ↩

  3. Psalm 119:105, KJV. The deity pronoun is capitalized here (Thy word); the King James lowercases it. The Highest Peak owns this verse and its argument; I am taking only the shape and pointing. ↩

  4. Isaiah 30:21, KJV. The Highest Peak used this verse for the fact that guidance arrives at all; I am using the clause it did not spend — the timing — and the two uses should be checked against each other so the two essays do not appear to be making the same move twice. ↩

  5. Hugh Everett III, "'Relative State' Formulation of Quantum Mechanics," Rev. Mod. Phys. 29 (1957), 454–462; for the mechanism by which branches stop interfering, Zurek, "Decoherence, einselection, and the quantum origins of the classical," Rev. Mod. Phys. 75 (2003), 715–775, and Joos & Zeh (1985). Note that decoherence suppresses interference between branches for all practical purposes — that is the literature's own hedge, and I am keeping it. ↩

  6. Max Tegmark, "On the dimensionality of spacetime," Class. Quantum Grav. 14 (1997), L69–L75, for the signature argument: with more or less than one time dimension the equations lose the hyperbolicity that lets an observer predict anything. Ultrahyperbolic equations — those with more than one timelike direction — have an ill-posed Cauchy problem in general. Two qualifications a reader who knows the area will raise, and the essay is stronger for raising them first: well-posedness can be recovered on restricted function spaces (Craig and Weinstein, Proc. R. Soc. A 465 (2009)), and Bars' two-time program reduces to one time by gauge symmetry. The claim in the body is "no predictivity," not "impossible." ↩

  7. Kaluza (1921) and Klein (1926): five-dimensional vacuum general relativity yields four-dimensional general relativity plus Maxwell's equations plus an additional scalar field — the scalar being, historically, an embarrassment, and later a feature. ↩

  8. Jorge Luis Borges, "El jardín de senderos que se bifurcan" (1941), collected in Ficciones (1944). The line is Stephen Albert's, and I quote it in Donald A. Yates's translation (1958), as collected in Labyrinths (ed. Yates and James E. Irby, 1962): "He believed in an infinite series of times, in a growing, dizzying net of divergent, convergent and parallel times." The translation is named because §II leans on the single word convergent, and a reader entitled to check should know which English text is being checked. ↩

  9. Richard Feynman, QED: The Strange Theory of Light and Matter (1985) — the rotating arrows are Feynman's own device, built for exactly this audience; the formalism is in Feynman & Hibbs, Quantum Mechanics and Path Integrals (1965). The description in the body is my paraphrase of the device, not a quotation of Feynman, which is why no words are put in his mouth. The technical name for why the extremal path survives the sum is the stationary-phase approximation; it is an asymptotic result in the limit of small action, and it should be named as an approximation in this note even though the body does not need the word. ↩

  10. The no-go result is R. H. Cameron, J. Math. and Phys. 39 (1960), 126–140: a Wiener measure with a complex diffusion constant is countably additive only when the imaginary part vanishes, so the real-time Feynman "measure" is at best finitely additive and cannot be rendered as a Lebesgue integral over path space. The rigorous Euclidean construction runs instead through the Wiener measure and the Feynman–Kac formula; see Glimm and Jaffe, Quantum Physics: A Functional Integral Point of View. On calling Fermat's principle "the shallow version": the word is licensed by the fact that ray optics is the short-wavelength limit of the wave account (Born & Wolf, Principles of Optics), so the principle is exact in its own domain and derived in the larger one. Shallow is meant as depth in the theory, not defect in the result — Fermat is not wrong, he is downstream. ↩

  11. See the note on Everett and Zurek above. The hedge stands: interference between branches is suppressed for all practical purposes, not exactly — which is the literature's own formulation, not a softening of it. ↩

  12. Dietrich Braess (1968); accessible modern treatments in Roughgarden, Selfish Routing and the Price of Anarchy (2005). The general framework is the price of anarchy — Koutsoupias & Papadimitriou (1999). The body cites no real-world instance and rests on the theorem alone. The usual examples — Seoul's Cheonggyecheon, Stuttgart, 42nd Street in 1990 — are well documented and I found nothing debunking them, but each is a road closure with several things changing at once, which is a weaker kind of evidence than a proof, and the argument here does not need them. ↩

  13. Distributed and decentralized optimization; Nedić & Ozdaglar, "Distributed Subgradient Methods for Multi-Agent Optimization," IEEE Trans. Automat. Control 54 (2009), and the consensus literature behind it. Stated in the body at its defensible size and labeled an analogy in the same sentence, per the voice's precision rule. ↩

  14. The Highest Peak, §VII: Jonah pays a fare and goes the other way, and the storm falls on the sailors, who did nothing. The wake of a deviation lands on people who never took it. That argument is made there; the geometry here is the same claim in different notation. ↩

  15. Matthew 22:39, KJV; the parallels are Mark 12:31 and Leviticus 19:18. ↩

  16. Galatians 6:2, KJV. ↩

  17. The sibling essay The Distance That Love Needs argues that twoness is the ontology of love — that there must be two for love to happen at all. This movement claims something adjacent and different: that the criterion picking a route is a relation between neighbors, so agreement is not a property any one route has by itself. The essays should be read together and are not making the same argument. ↩

  18. Richard Bellman, Dynamic Programming (Princeton, 1957), p. 83 — the principle of optimality, in his own words: "An optimal policy has the property that whatever the initial state and initial decisions are, the remaining decisions must constitute an optimal policy with regard to the state resulting from the first decision." Note what the sentence quantifies over: the remaining decisions. It is a statement about tails, and there is no corresponding statement about heads anywhere in the theory — not omitted, but unavailable, since a head that looked best at every step would make the whole apparatus unnecessary. ↩

  19. Holy Saturday, and the practice described in the body is the Western one. In the Roman rite there is no Mass on Holy Saturday — the altar stays bare from the stripping on Maundy Thursday until the Easter Vigil, and Communion is given only as Viaticum to the dying; the Episcopal Book of Common Prayer likewise appoints no Eucharist, only a spare liturgy of the word. The Orthodox observance differs and is not what is being described. The descent — 1 Peter 3:19, 4:6, and the creed's descendit ad inferos — is genuinely contested, including in the twentieth century. I need the slice, not the doctrine, and I am not adjudicating it here. ↩

  20. Romans 12:15, KJV — Rejoice with them that do rejoice, and weep with them that weep. ↩

  21. Job 42:7 and John 9:2–3, KJV; both are argued at length in The Highest Peak §VIII(a), which is where this refusal properly lives. The one-word answer in John 9:3 is Neither. ↩

  22. Ted Chiang, "Story of Your Life," collected in Stories of Your Life and Others (Vintage) — quoted and located at p. 118 in The Highest Peak §IX. The body here paraphrases and does not quote, which is the cleaner arrangement: the companion owns the quotation, this essay owns one clause of the idea, and there is no second pagination to keep in step. ↩

  23. Matthew 10:39, KJV — He that findeth his life shall lose it: and he that loseth his life for my sake shall find it. The parallel is Matthew 16:25. The body references and does not quote, because the clause it would want is a compression of the verse rather than the verse. ↩

Also published at substack.

← All writings