> ## Content Index
> Fetch the complete content index at: https://otherwise.ghost.io/llms.txt
> Use this file to discover other available public pages before exploring further.

# The Axe Was Visible. The Explanation Was Not.
- URL: https://otherwise.ghost.io/the-axe-was-visible-the-explanation/
- Published: 2026-10-05T09:03:26.000Z
- Updated: 2026-10-17T13:39:05.000Z
- Description: Flight FZ1073, Invisible Explanations, and the Architecture of an Open Future
- Author: Khaled Salih
- Tags: Newsletter, #Migrated-1792244264359, #Import 2026-10-17 14:38

5 October 2026

---

![](https://storage.ghost.io/c/d6/2c/d62cdd13-a476-41c5-8f2c-29270c99d95d/content/images/2026/10/ef2c9a67-5128-4076-bf99-0c907e900ce9_1536x1024.png)

---

# Note

For the next three days, leading into the third anniversary of October 7, 2023, I will publish three essays that approach the same problem from three different directions.

The first begins somewhere unexpected: inside the cockpit of FlyDubai Flight FZ1073.

There was an axe. There was blood. There was a descending aircraft.

Those things were visible.

But visible events do not explain themselves.

Before an idea becomes a weapon, a law, a war, a massacre, an institution—or an act of courage—it exists in an invisible world of explanations, categories, permissions, conjectures, and choices.

And sometimes the decisive question is not merely what went wrong.

It is whether criticism can still reach the error before the error becomes irreversible.

Tomorrow, I turn directly to October 7.

---

# The Axe Was Visible. The Explanation Was Not.

Flight FZ1073, Invisible Explanations, and the Architecture of an Open Future

  
**4 October 2026**

There was an axe in the cockpit.

There was blood.

There was a wounded captain.

There was an aircraft suddenly descending with passengers aboard.

And there was a reinforced cockpit door separating the people outside from the struggle inside.

All of this became physically visible.

Almost none of it, by itself, explains what happened.

That is where I want to begin.

Not with the axe. Not with Oman. Not with Islam. Not with Israel. Not even with terrorism.

I want to begin with something that could not be photographed from another aircraft or reconstructed from a flight tracker:

**an explanation.**

Before a human hand could take an emergency axe and turn it against another human being, something had already happened in an invisible world. An idea, grievance, moral category, doctrine, fantasy, hatred—or some combination of explanations—had apparently acquired enough authority to move a body.

Something had made an action thinkable that should have remained unthinkable.

Then the invisible entered the physical world:

**Explanation → perception → moral permission → choice → action → physical reality.**

That movement—from invisible explanation to visible consequence—is what interests me about Flydubai Flight FZ1073.

But the movement did not stop there.

Something moved in the opposite direction too.

A wounded captain acted. A closed door opened. Other human beings entered the problem. New knowledge entered the cockpit. The range of possible futures changed. An intended outcome failed to become irreversible.

So the complete story contains another movement:

**Problem → criticism → agency → new knowledge → expanded possibility → correction → different physical reality.**

Between those two movements lies almost everything I have been trying to understand in my work.

How do explanations become worlds?

Where does the error go?

And how do human beings preserve the possibility of creating a different future before a bad explanation becomes physically irreversible?

## What We Know—and What We Do Not

The first obligation of explanation is not to outrun the evidence.

On September 30, 2026, FlyDubai Flight FZ1073 was travelling from Dubai to Tel Aviv when a violent incident occurred inside the cockpit.

On October 3, the Attorney-General of the United Arab Emirates announced that investigators had found that the co-pilot had begun executing what authorities described as a planned terrorist act. According to the official account, he attacked Captain Smit Machchhar with the cockpit’s emergency crash axe and attempted to take control of the aircraft’s flight controls.

The aircraft was ultimately diverted to Tabuk in Saudi Arabia, where it landed safely.

But the investigation is not finished.

UAE authorities say they are still investigating the full circumstances, motives and possible links, as well as physical and digital evidence.

That boundary matters.

The Associated Press and other reporting identify the co-pilot as Hamam al-Hammami, an Omani national, and report that he had previously been removed from flying duties at Oman Air because of concerns about extremist views. AP also reports that investigators have not, at this stage, found links to militant organizations.

Those facts matter too.

But they do **not** yet give us a completed explanation of motive.

Nor do they establish that FlyDubai knew of the earlier concerns.

Nor that Emirati authorities knew.

Nor that Oman failed to communicate information it was required or able to communicate.

Nor that an organization directed the attack.

Those remain open questions.

I emphasize this not as a ritual disclaimer but because it is part of my argument.

**A theory of error-correction that cannot correct its own explanation is already contradicting itself.**

## The Visible World Does Not Explain Itself

Consider the crash axe.

It is heavy. It has an edge. It has physical capabilities.

But it contains no instruction telling a pilot whom to attack.

The aircraft contains no hatred.

The cockpit door has no political convictions.

Gravity has no ideology.

A Boeing does not know whether the passengers behind its cockpit are Jewish, Israeli, Muslim, Christian, Hindu, Kurdish, Arab or anything else.

Matter constrains what can happen.

It does not, by itself, explain why a human being selects one possibility rather than another.

Captain Machchhar’s own recollection captures this distinction with extraordinary clarity.

When the first blow struck the back of his head, his initial explanation was physical: he thought the aircraft had suffered catastrophic damage.

Then the blows continued.

“That’s when I realized that it’s not the aircraft,” he later recalled, “it’s the other guy who was in the cockpit with me.”

That is almost a miniature lesson in explanation.

The same physical experience—the blow, the disorientation, the suddenly unstable aircraft—first generated one conjecture and then, under criticism from new evidence, another.

The visible facts had not told him what they meant.

He had to explain them.

When the explanation changed, the problem changed with it: from *aircraft failure* to *human agency*.

This is precisely why observable events do not explain themselves.

Human beings act upon explanations of the world.

They possess ideas about who they are, who others are, what is permissible, what is forbidden, who is innocent, who is guilty, what counts as humiliation, what counts as justice, what death means, what sacrifice means, what history demands, what God demands, what the nation demands, and what an enemy deserves.

Most of these things have no physical dimensions.

Yet they can reorganize physical reality.

This is what I call the **Invisible Reality Test**:

***What invisible explanation is moving the visible world?***

## Invisible Does Not Mean Unreal

We often use the word *real* too narrowly.

The axe is real because we can touch it.

The wound is real because it bleeds.

The aircraft is real because radar can track it.

But explanations can be causally real too.

Nationalism has no weight.

Neither does jihadism.

Neither does antisemitism.

Neither does liberty.

Neither does human dignity.

Neither does a constitution.

Neither does the proposition that all human beings possess equal moral worth.

Yet these invisible abstractions organize armies, parliaments, schools, borders, laws, prisons, hospitals and cemeteries.

Ideas become institutions.

Institutions become incentives.

Categories become permissions.

Explanations become choices.

Choices become actions.

Actions alter matter.

The human physical world is therefore, in significant part, **invisible explanation made visible**.

That is my first conjecture.

It is also why beginning an explanation only after an axe is swung, a missile launched, a woman murdered, a minority expelled or an aircraft begins to fall often means beginning too late.

The physical event may be the final visible link in a much longer explanatory chain.

## The Axe Cannot Explain the Hand

This also tells us what not to do.

The accused co-pilot was Omani.

That may be descriptively relevant.

But **Omani is not a mechanism**.

He was Muslim.

But **Muslim is not a mechanism**.

If the investigation establishes that an Islamist or jihadist explanation was causally important, that explanation should be exposed to criticism without hesitation.

But *Islamist* cannot simply become a sophisticated-sounding word inserted into the space where an explanation is missing.

We have to ask harder questions.

What precisely did this person believe?

What evidence tells us that?

What moral categories did those beliefs create?

How did they alter his perception of other human beings?

What made an action possible for him that remained impossible for millions of people sharing some of the same demographic categories?

What mechanism connected explanation to action?

And, crucially:

***What evidence would show that our explanation is wrong?***

That question separates explanation from accusation.

## Hard to Vary

David Deutsch gives us a powerful test.

A good explanation should be **hard to vary while still accounting for what it claims to explain**.

Suppose I say:

He acted because he was Omani.

Replace Oman with another country. The possibility of political violence remains. The explanation has barely changed.

Suppose I say:

He acted because he was Muslim.

Hundreds of millions of Muslims do not attempt to seize civilian aircraft. Again, the category is doing very little explanatory work.

Suppose I say:

He opposed Israel.

Millions of people strongly oppose Israel without attacking pilots or attempting to destroy aircraft.

Again, insufficient.

A serious explanation must become more constrained, not less.

Which ideas?

Which grievance?

Which doctrine?

Which conception of the passengers?

Which understanding of killing?

Which moral permission?

Which sequence of radicalization?

Which institutional signals?

Which failed interruptions?

The more precisely we specify the mechanism, the more opportunities reality acquires to tell us that we are wrong.

That is a virtue.

**I do not want an explanation incapable of losing. I want an explanation good enough to be refuted.**

## Protect Persons. Criticize Explanations.

There are two easy mistakes after an event like this.

The first is collective accusation.

An Omani individual acts, therefore Oman becomes implicated.

An Arab acts, therefore Arabs become suspect.

A Muslim acts, therefore Islam itself becomes the defendant.

That is not explanatory depth.

It is an expansion of blame.

But there is an opposite mistake.

Because we rightly fear collective guilt, we can become reluctant to investigate ideas at all.

Perhaps a religious interpretation mattered.

Perhaps a political ideology mattered.

Perhaps antisemitism mattered.

Perhaps jihadist thinking mattered.

Perhaps a particular understanding of Gaza or Israel mattered.

Perhaps personal pathology or some entirely different mechanism mattered more.

We do not yet know enough.

Intellectual openness requires us to permit every relevant explanation to be criticized.

My rule is therefore:

**Protect persons. Criticize explanations.**

Do not attack a population because an idea is bad.

Do not protect an idea because criticizing it might offend a population.

That preserves both human dignity and intellectual seriousness.

**Where Does the Error Go?**

Now another question appears.

Reporting indicates that concerns about al-Hammami’s ideological views existed during his earlier aviation career in Oman.

If that reporting survives scrutiny, we face a problem larger than one individual.

But we must formulate it correctly.

The question is not yet:

*Why did FlyDubai ignore the warning?*

We do not know that it received one.

The better question is:

***Was corrective knowledge created somewhere without acquiring a route to the next decision that needed it?***

That leads directly to the question underneath much of my work:

***Where does the error go?***

Suppose one institution detects something important.

What then?

Can the warning travel?

Who receives it?

Is it intelligible outside the institution that produced it?

Does anyone have authority to act?

Can the relevant decision still be reversed?

Does privacy law prevent transmission?

Does jurisdiction intervene?

Do incompatible bureaucratic categories make the information disappear?

Does somebody possess the fragment without possessing the whole?

Could everyone perform a local function correctly while the larger system still moves toward catastrophe?

We often think institutional failure means nobody knew.

But there is another possibility.

**Somebody knew.**

Knowledge can exist locally while the system behaves globally as though the knowledge does not exist.

That gives me another conjecture:

**A warning that cannot travel is criticism without a correction mechanism.**

## Criticism Must Reach the Error

This is profoundly Popperian.

It is not sufficient that criticism exist.

Criticism must be capable of reaching whatever needs correcting.

And it is profoundly Deutschian.

Knowledge does not automatically propagate merely because somebody possesses it. It requires traditions and mechanisms through which it can be transmitted, criticized, improved and acted upon.

So separate the stages:

**Detection → Transmission → Authority → Reversibility → Correction.**

A system may detect an error and fail to transmit the warning.

It may transmit the warning to somebody without authority.

Someone may possess authority but confront a decision that has already become irreversible.

Knowledge may therefore exist somewhere in a system while correction still fails.

That is not simply ignorance.

It is an **architectural failure of knowledge**.

**The Protected Space**

Now return to the cockpit door.

Why is it there?

Because aviation learned from catastrophe.

The danger could come from the cabin. Someone could force entry into the cockpit and seize the aircraft.

The solution was rational:

**valuable function inside → threat outside → stronger barrier.**

Human institutions repeatedly solve problems this way.

Courts receive independence.

Intelligence agencies receive secrecy.

Political parties develop discipline.

Religions protect traditions.

Universities acquire autonomy.

Computer systems restrict privileges.

Open societies protect institutions from people who would destroy openness.

Protection is not the problem.

Sometimes the door absolutely should be locked.

But every solution changes the problem situation.

No institutional design abolishes error.

It redistributes opportunities for error.

And FZ1073 exposes a possibility every protected architecture eventually faces:

***What if the danger is already inside?***

Then the direction reverses:

**threat inside → possible correction outside.**

A barrier designed to keep danger from entering can now obstruct correction from entering.

The door may be doing precisely what it was designed to do.

That is what makes the problem difficult.

## The Inside–Outside Inversion

Call this the **Inside–Outside Inversion**.

A political party creates discipline to protect itself against enemies. Then its leadership becomes the source of error.

A religion protects inherited knowledge. Then one interpretation becomes immune to criticism.

A state creates secrecy because hostile states genuinely exist. Then secrecy begins protecting incompetence.

A university protects academic freedom against political interference. Then autonomy is invoked to protect an internal orthodoxy.

A movement creates solidarity because its people have genuinely suffered. Then criticism of an idea becomes classified as aggression against the people.

An open society protects tolerance. Then an intolerant movement learns to exploit tolerance against the institutions preserving it.

An AI system receives layers of safeguards. Then the mistaken assumption sits inside the architecture defining what counts as safe.

The structure recurs:

**Problem → solution → protection → changed problem situation → internal error → impaired criticism.**

The mistake is not protection.

The mistake is imagining that protection designed against yesterday’s error has somehow abolished tomorrow’s.

## Protected Error

An ordinary error can be criticized.

A **Protected Error** has acquired something additional:

**an explanation for why criticism should not reach it.**

The leader says criticism helps the enemy.

The religious authority says criticism attacks the sacred.

The movement says criticism harms the vulnerable.

The bureaucracy says criticism threatens security.

The institution says outsiders cannot understand.

The ideology says dissent itself proves contamination.

Now the original error has recruited a second explanation to defend it:

**Error → criticism → protective explanation → blocked criticism → surviving error.**

And when failure eventually occurs, another explanation can arise to protect the system from learning from failure:

**Protected Model → Blocked Criticism → Failure → Post-Failure Explanation → Protection of the New Explanation → Second-Order Closure.**

Now the important question is no longer merely:

*What is the error?*

It is:

***Where is the architecture sending it?***

Toward criticism?

Or toward protection?

## The Door Opened

Then something happened aboard FZ1073 that deserves more attention than any metaphor I can place upon it.

The cockpit door opened.

Captain Machchhar’s account makes the mechanism unusually revealing. People outside were attempting to enter using the cockpit access code, he said, while the co-pilot was preventing the door from opening. From outside, potential correction existed but could not get through. The decisive mechanism remained inside: a knob that had to be physically reached and turned.

Machchhar remembered lying on the cockpit floor, hearing people screaming and feeling the aircraft descending.

He described having “one last fight” left.

He reached the door and opened it.

That detail matters enormously.

The problem was not an absence of potential correctors.

They were already there.

The problem was **correction access**.

Knowledge, courage and physical capacity existed on the other side of the barrier, but the architecture temporarily prevented them from reaching the error.

This was not merely symbolic.

It changed the causal structure of the problem.

Before the door opened, knowledge and agency were divided.

Inside were the controls, the attacker and a badly injured captain.

Outside were passengers, crew and other trained people aboard the aircraft.

The barrier separated not merely two physical spaces.

At that moment it separated **a problem from potential correctors**.

When Machchhar opened the door, the distribution of knowledge changed.

People who had previously been unable to act upon the cockpit problem could enter it.

This is one of the deepest characteristics of open systems.

They do not assume that the official holder of authority will always possess all the knowledge necessary to solve the problem.

They preserve the possibility that correction can come from somewhere unexpected.

**The Unexpected Corrector**

Popper’s open society is often described politically: elections, pluralism, tolerance, the rule of law.

All matter.

But beneath them lies an epistemological proposition:

**We do not know in advance where the next correction will come from.**

The junior scientist may notice what the professor missed.

The dissident may understand what the minister cannot.

The outsider may see an assumption invisible to insiders.

The child may ask the question adults have learned not to ask.

And on FZ1073, an Israeli plumber carried a piece of useful knowledge into a cockpit.

Yaniv Hayoun was not an airline pilot. According to his account, after entering the cockpit and helping overpower the co-pilot, he found himself near the controls while the aircraft was descending. He says he pulled back on them until the aircraft began to level, before a trained pilot took over.

When Netanyahu asked how he knew what to do, Hayoun’s answer was almost comically mundane:

**“I watch Air Crash Investigation!”**

I keep returning to that answer because it contains something much larger than its simplicity.

Useful knowledge had been created in one context, transmitted through television, retained by a plumber, and suddenly became relevant inside an aircraft cockpit.

Nobody designed FZ1073’s emergency architecture around Hayoun.

Nobody assigned him this epistemic role.

The knowledge existed somewhere in society.

Once the protected space became accessible, that knowledge could reach the problem.

This is what Deutsch calls the **reach of explanations**: knowledge created to solve or explain one class of problems can unexpectedly become useful somewhere else.

An open civilization therefore does not need to know in advance who will possess the next useful piece of knowledge.

It needs an architecture in which useful knowledge can reach the problem when it matters.

The open society preserves **unexpected correctors**.

## Choice-Space

Opening the door changed something else.

It changed the **Choice-Space**.

At one moment the available futures must have appeared horribly narrow.

An attacker in the cockpit.

A seriously injured captain.

A descending aircraft.

Passengers separated from the controls.

The perceived menu was closing.

But:

**Objective Choice-Space ≠ Perceived Choice-Space.**

The possibilities currently visible to us are not necessarily the limits of what can be created.

This is where Deutsch becomes indispensable.

Human beings are not merely selectors from menus.

We are creators of explanations.

And new explanations can create new possibilities.

Sometimes the correct response to two terrible doors is not to choose one.

It is to create a third.

Machchhar did something analogous.

He did not solve the entire problem himself.

**He changed the problem so that other people could help solve it.**

He changed who could participate.

He altered the architecture within which choices could be made.

The door opened.

New agents entered.

New knowledge entered.

And a future unavailable moments earlier became possible.

## Invisible to Visible—and Back Again

Now look at the event as two opposing causal movements.

The first moved from invisible explanation toward physical consequence:

**Explanation → moral categorization → perceived possibility → decision → action → axe → wounded captain → descending aircraft → approaching irreversibility.**

The second moved from criticism toward a different physical future:

**Problem recognition → agency → opened barrier → distributed knowledge → intervention → recovered control → landing → investigation → learning.**

Ideas become physical.

But criticism becomes physical too.

A bad explanation can move a hand toward an axe.

A different explanation can move a wounded man toward a door.

And knowledge acquired for no imaginable emergency can suddenly become relevant to one.

Hayoun’s account completes the reversal. Something learned from a television program became remembered knowledge; remembered knowledge became action; action entered a cockpit and became part of a changed physical trajectory.

The invisible became visible again—but in the opposite direction.

Human beings do not merely inhabit matter.

We inhabit problems, conjectures, explanations, criticisms and possibilities that continuously reorganize matter.

That is why fatalism is such a poor explanation of human affairs.

## The Deeper Conjecture

I can now state my central conjecture more precisely.

**Much of the human physical world is invisible explanation made visible.**

Error becomes especially dangerous when an explanation acquires enough power to reorganize physical reality while simultaneously blocking the criticism capable of correcting it.

Progress occurs when criticism generates new explanations—and therefore new possibilities—quickly enough to create a different physical future before the old explanation becomes irreversible.

That proposition reaches far beyond one aircraft.

It applies to ideologies, religions, states, institutions, wars, technologies, artificial intelligence—and open societies themselves.

The fundamental problem is not that human beings make errors.

We will.

The problem begins when errors acquire mechanisms capable of turning themselves into reality faster than criticism can interrupt them.

## Two Architectures of the Future

My boldest conjecture about FZ1073 is therefore this:

For a brief period aboard that aircraft, two competing architectures of the future confronted one another.

One moved toward closure:

**explanation → justification → violence → disabled criticism → protected space → narrowing choice-space → irreversibility.**

The other moved toward correction:

**problem → agency → reopened access → distributed knowledge → expanded choice-space → intervention → survival → learning.**

One attempted to reduce possible futures.

The other created one.

One attempted to make an explanation irreversible.

The other interrupted it.

One moved toward closure.

The other preserved an open future.

This is larger than terrorism confronting courage.

For those minutes:

**Protected Error confronted Open Correction.**

The argument had altitude.

It had velocity.

It had blood.

And it had human lives inside it.

## The Backspace Threshold

Not all errors are equally dangerous.

Their danger increases as their consequences become harder to reverse.

A foolish thought in a notebook can be crossed out.

A mistaken article can be corrected.

A bad administrative decision may be reversed.

A bad law is harder.

A war harder still.

A missile already launched radically shortens the correction window.

An aircraft descending toward the ground can reduce it to seconds.

I call the point at which this asymmetry becomes critical the **Backspace Threshold**.

The relevant variables include speed, reach, autonomy, replication, opacity and irreversibility.

Then ask:

***Can our capacity to discover, interrupt and correct the error keep pace with the error’s capacity to become irreversible?***

That is no longer merely an aviation question.

Artificial intelligence makes it urgent.

So do autonomous weapons, biotechnology, nuclear command systems, algorithmic markets and viral information environments.

Technology can move increasingly quickly from invisible instruction to physical consequence.

Can criticism move quickly enough in the other direction?

That may become one of the central questions of technological civilization.

## No Final Institutional Solution

This is why I distrust promises of final arrangements.

There is no perfect cockpit.

No perfect constitution.

No perfect religion.

No perfect democracy.

No perfect security service.

No perfect university.

No perfect algorithm.

No perfect ideology.

There is no human institutional arrangement beyond error.

Every solution alters the problem situation.

Every protection generates new possibilities for failure.

Every generation inherits explanations it did not create.

What matters is whether it also inherits the means to criticize them.

The central question is therefore not:

***How do we build a world without error?***

We cannot.

It is:

***How do we build a world in which error cannot easily become final?***

That is the architecture of an open future.

## And This Explanation Must Remain Open Too

There is one more test.

It applies to me.

Everything I have written here is itself a conjecture.

The official investigation into Flight FZ1073 is continuing.

We still do not know the complete motive.

We do not know the full ideological history of the accused co-pilot.

We do not yet know what information moved—or failed to move—between previous employers, authorities and FlyDubai.

And investigators have not, at this stage, established an organizational connection to a militant group.

Those are not inconvenient gaps to be covered with rhetoric.

They are opportunities for criticism.

So let me state in advance what correction would require.

If investigators ultimately establish that ideology played no meaningful causal role in the attack, then the ideological part of my explanation must change.

If reports concerning earlier extremist concerns prove false or materially misleading, then my warning-transmission conjecture must change.

If evidence shows that the relevant information was transmitted and appropriately acted upon, then the institutional mechanism must be explained differently.

If facts emerge that undermine my account of how correction entered the cockpit, that part too must be rewritten.

I do not regard this as weakness in the argument.

**It is the argument.**

The Open Correction Framework cannot demand corrigibility from governments, religions, universities, movements and artificial intelligence while granting immunity to its own conjectures.

A good explanation should tell us not only what it explains.

It should expose where reality can defeat it.

So this essay has a Backspace Key built into it.

Future evidence is invited to press it.

## The Axe and the Door

I return finally to the two physical objects.

The axe.

And the door.

Both were visible.

Neither explains itself.

Behind the axe was an invisible explanatory world that investigators are still reconstructing.

Behind the door was another invisible world: the recognition that the situation was not yet final; the knowledge that people outside might alter it; the continued agency of a wounded human being; knowledge distributed among people who had never expected to need it; and the possibility that the architecture of the problem could still be changed.

The axe represented an attempt to move from invisible explanation toward irreversible physical reality.

Opening the door allowed agency, criticism and new knowledge to move in the opposite direction.

That is the movement I want us to see:

**Invisible explanation → visible world.**

And then:

**criticism → new explanation → new possibility → different visible world.**

This is why the future cannot be read directly from the present.

Bad explanations are not laws of nature.

Inherited explanations are not destiny.

Protected errors are not necessarily permanent.

They can be criticized.

They can be exposed.

They can be replaced.

Their consequences can sometimes be interrupted.

New knowledge can create possibilities that did not previously exist.

And our explanations of those events must themselves remain open to the same process.

The axe was visible.

The explanation was not.

Then a wounded man opened a door.

**And the future entered.**

## A Note on Correction

This essay reflects the evidence publicly available on **October 4, 2026**. The investigation into Flight FZ1073 is continuing. Some facts reported above derive from official UAE findings; others derive from attributed reporting and participant accounts. The explanatory connections I draw from them are conjectures.

Where subsequent verified evidence changes the factual premises or defeats an explanatory conjecture developed here, I will amend the text rather than protect an obsolete version of the argument.

Substack makes such revision technically easy.

The philosophy makes it intellectually necessary.

**That is not an editorial afterthought.**

**It is the method.**