11 September

Note

For six articles, I have been criticizing the two doors.

Now comes the obligation created by criticism: build something better.

This seventh article is the constructive turn in my series on the Oxford Union debate. The argument has moved from the Binary Trap, through the badly framed Israel–Palestine choice-space, to why winning a debate is not solving a problem, why intelligence is not rationality, and why we must always ask where the error goes.

But criticism alone is insufficient.

What should happen when Proposition and Opposition have both been tested—and neither gives us an adequate explanation?

My answer is the Third Door.

Not compromise. Not centrism. Not “both sides.” And not simply a third predetermined option called C.

The deeper conjecture is this:

The Third Door is not C. It is the freedom to create C.

Human beings do not merely choose among alternatives handed to them. We create knowledge—and new knowledge creates possibilities that did not previously exist.

So perhaps the most important freedom is not the freedom to choose between the doors already open.

It is the freedom to question the wall, press Backspace, reopen the problem, and build a door that nobody could see when the argument began.

Article 7 Build the Third Door

The Third Door Is Not Another Answer. It Is the Freedom to Create One.

By Khaled Salih

11 September 2026

For six articles, I have been criticizing two doors.

Now I have to build something.

That is the obligation criticism creates.

It is relatively easy to say that a question is badly framed. It is harder to propose an architecture in which the question itself can be corrected.

It is relatively easy to show that Proposition and Opposition do not exhaust reality. It is harder to explain what should happen after both have been criticized.

And it is relatively easy to invoke the possibility of a “third door” as a metaphor.

The difficult question is:

What exactly is the third door?

I now think the answer requires an important correction.

The third door is not a third answer.

It is not C.

It is the freedom to create C.

That distinction changes almost everything.

What the First Six Articles Discovered

This series began with an Oxford Union debate about Israel and Palestine.

The motion was:

“This House Believes Israel Never Truly Wanted Peace with Palestine.”

Oxford provided two doors.

Proposition.

Opposition.

Yes.

No.

I began by asking whether those two doors were enough.

In “The Oxford Binary Trap,” I argued that adversarial debate possesses an extraordinary civilizational virtue: it institutionalizes criticism. Someone makes an argument and someone else is permitted—even invited—to attack it.

That must be preserved.

But the architecture contains a limitation. The motion exists before criticism begins. The problem has already been selected, compressed and divided into two positions.

Criticism enters after the choice-space has been constructed.

Then, in “Israel Was Given Two Doors,” I applied Backspace to the motion itself.

What exactly was Israel?

What was Palestine?

What could falsify never?

How could we test truly?

What did peace mean?

Once those words were opened, what had appeared to be one question became many.

Then came Popper.

In “Winning the Debate Is Not Solving the Problem,” I argued that defeating A does not establish B.

A counterexample may destroy a universal proposition.

Good.

An error has died.

But the death of an error is not yet the birth of an explanation.

Then came “Intelligence Is Not Rationality.”

Oxford is full of intelligent people. But intelligence can be given the wrong job. It can search, remember, calculate, rebut, reinterpret and defend. Under the wrong incentives, extraordinary intelligence can become an extraordinary protection system for an assigned conclusion.

Rationality demands something harder:

Can criticism change the answer?

And in “Where Does the Error Go?,” the problem moved from cognition to political architecture.

A bad decision should send criticism toward the decision, its premises, its authors, the institutions that protected it, and the doctrines that generated it.

Instead, politics often sends error elsewhere:

Decision
→ government
→ political movement
→ nation
→ religion
→ civilization
→ people.

Once the error reaches the people, correction becomes extraordinarily difficult.

Six articles have therefore brought me to one constructive problem:

What would an institution look like if it were designed not merely to produce criticism, but to allow criticism to create new choices?

That is the Third Door problem.

Keep the First Two Doors

The first thing I want to preserve is precisely what my argument might appear to threaten.

Keep Proposition.

Keep Opposition.

Keep disagreement.

Keep the adversarial encounter.

Keep the person who says:

You are wrong.

Einat Wilf’s criticism of intellectual withdrawal remains important to me. If an argument is wrong, abandoning the arena does not correct it. People must show up. Facts matter. Preparation matters. Counterarguments matter. Unpopular explanations must be allowed into the room.

Karl Popper gives us good reasons for protecting this architecture.

An idea shielded from criticism can become dangerous.

So the Third Door is not an argument against the first two.

It comes after them.

Let A encounter Not-A.

Let each expose weaknesses in the other.

Let the strongest criticisms do their work.

But then refuse one assumption:

that the intellectual task must end with choosing which side survived.

That is where the third door begins.

The Third Door Is a Meta-Choice

Suppose the available alternatives are:

A

and

Not-A.

The third door is not automatically:

C.

That would merely replace a two-item menu with a three-item menu.

Soon we would be trapped again:

A, B or C?

Choose.

My argument is more radical.

The third door is the institutional permission to say:

A has encountered a fatal criticism.

Not-A also fails to explain the problem.

Something was wrong with the original framing.

We need to conjecture again.

The Third Door therefore protects at least four freedoms:

the freedom to revise the question;

the freedom to reject the available alternatives;

the freedom to create an explanation that did not previously exist;

and the freedom to leave a problem provisionally unresolved rather than force a false choice.

This is why the idea is deeply connected to David Deutsch’s treatment of choices.

Human beings do not merely select among possibilities presented by reality.

We create knowledge.

And knowledge creates possibilities.

The airplane was not waiting on an ancestral menu between walk and run.

Flight became possible because knowledge was created.

Political possibility is different in important respects, but the epistemological principle survives.

Institutions humans once lacked had to be invented.

Constitutional government.

Federal arrangements.

Independent courts.

Peaceful transfers of power.

Religious toleration.

Minority protections.

International institutions.

None arrived because humanity finally selected the correct item from an eternal menu.

Knowledge changed the menu.

That gives us a principle:

Never infer the boundaries of possibility from the boundaries of the presently available choices.

Sometimes the Error Is the Menu

This is where the Backspace Key becomes more important than I originally understood.

We normally imagine correction like this:

Answer
→ criticism
→ corrected answer.

But what if the answer was generated by a bad premise?

Backspace.

What if the premise arose from a misleading category?

Backspace again.

What if the categories were created by a badly formulated problem?

Again.

And what if the problem itself exists only because somebody handed us a menu and persuaded us that reality must fit inside it?

Then Backspace must reach the choice-space itself.

That gives us a deeper sequence:

Answer → Premise → Frame → Problem → Choice-Space.

Every level must remain criticizable.

Otherwise we can possess enormous freedom downstream while protecting the decisive error upstream.

We are free to choose.

But not free to ask who created the choices.

That is not enough.

Oxford Starts Criticism Too Late

This is my central criticism of the Oxford architecture.

Not that it criticizes too much.

It begins criticism too late.

Before Proposition and Opposition enter the chamber, several intellectual operations have already occurred:

Reality has been selected.

A problem has been extracted.

The problem has been framed.

Categories have been chosen.

A proposition has been formulated.

The proposition has been converted into a binary.

Only then are brilliant people invited to criticize.

So imagine extending the institution.

Keep the Oxford debate exactly as vigorous as it is.

But when Proposition and Opposition have finished, do not immediately ask:

Who won?

Ask first:

What did we learn was wrong?

Then:

Which claims from Proposition survived?

Which claims from Opposition survived?

Which claims from the winning side failed?

Which claims from the losing side survived?

What assumptions did both sides share?

Did the motion contain an error?

Has the original problem changed?

What possibilities were excluded before the debate began?

And most importantly:

What can we conjecture now that nobody could have conjectured before the criticism occurred?

That final question changes the purpose of the institution.

The Backspace Debate

I would therefore add a second architecture to adversarial debate.

Call it the Backspace Debate.

Its sequence would be:

Problem
→ Initial Conjectures
→ Adversarial Criticism
→ Error Ledger
→ Choice-Space Audit
→ Hidden-Premise Audit
→ Problem Reformulation
→ Alternative Creation
→ Revised Conjectures
→ Critical Testing
→ Provisional Lock.

The Error Ledger asks what criticism actually destroyed.

Not who scored the rhetorical point.

What explanation failed?

The Choice-Space Audit asks whether the original alternatives were exhaustive or merely inherited.

The Hidden-Premise Audit asks what both sides were taking for granted.

Problem Reformulation permits the question itself to change.

Alternative Creation gives institutional standing to an explanation that belonged to neither team.

Critical Testing immediately exposes that new conjecture to criticism.

And the Provisional Lock says only:

This is the best surviving explanation for now.

Not forever.

The Backspace Key remains attached.

Israel and Palestine Show Why This Matters

Return to the Oxford motion.

The two available positions were roughly:

Israel never truly wanted peace with Palestine.

Or:

that proposition is false.

But after listening to the arguments, another problem becomes visible.

Perhaps wanting peace is not the best explanatory variable.

Perhaps we should instead ask:

Which decisions, institutions, doctrines, incentives and feedback loops have repeatedly narrowed the future choice-space available to Israelis and Palestinians?

Now settlements can be examined as a mechanism.

Violence can be examined as a mechanism.

Security fears can be examined as mechanisms.

Rejectionism can be examined.

Territorial ideology can be examined.

Institutional weakness can be examined.

External intervention can be examined.

Failed negotiations can change later political expectations.

Israeli actions can alter Palestinian choices.

Palestinian actions can alter Israeli choices.

The resulting explanation need not distribute responsibility equally.

This point is indispensable.

Causal interaction is not moral equivalence.

One actor may possess more power.

One decision may cause vastly more harm.

One institution may bear greater responsibility.

A mechanism can be asymmetric.

But once we ask about mechanisms rather than collective souls, criticism acquires an address.

And once criticism has an address, correction becomes conceivable.

That is a third-door move.

Not because we have discovered the final solution to Israel and Palestine.

We have not.

We have created a better problem.

The Third Door Is Not “Both Sides”

This needs to be explicit.

The Third Door is not:

Both sides have a point.

Sometimes one side is simply wrong.

Sometimes one explanation survives and another collapses.

Sometimes evidence overwhelmingly favors one conjecture.

Fallibilism does not require artificial symmetry.

Nor does the Third Door mean compromise.

If someone conjectures that 2 + 2 = 4 and someone else insists that it equals 7, rationality does not require a negotiated settlement of 5.5.

The purpose is not to stand in the middle.

The purpose is to keep the choice-space open to criticism.

Sometimes criticism will tell us A was right.

Sometimes B.

Sometimes both contained errors.

Sometimes the problem was malformed.

And sometimes the knowledge required for a better answer has not yet been created.

The Third Door protects all of those possibilities.

“There Is No Alternative”

This is why one political sentence should always trigger an audit:

There is no alternative.

It may occasionally describe a genuine immediate constraint.

A government facing an imminent attack may have only minutes and a narrow range of feasible actions.

A patient may face a binary medical decision.

A parliament may ultimately have to vote Yes or No.

Choice is real.

Constraints are real.

Decisions cannot always wait.

But:

We currently know of no alternative

is not equivalent to:

No alternative can exist.

The first describes our knowledge.

The second claims knowledge about the boundaries of future knowledge.

That is a much larger claim.

No present generation possesses such knowledge.

No One Owns the Future Choice-Space

This may be the deepest lesson I take from the series.

Political power often consists not merely in deciding which answer wins.

It consists in deciding which answers are permitted to exist.

A religious authority can do this.

A nationalist movement can do it.

A revolutionary ideology can do it.

A state can do it.

An academic discipline can do it.

A debating institution can do it inadvertently.

Even an emancipatory movement can close the future while claiming to liberate it.

The strongest form of protected error may therefore not be protecting an answer.

It may be protecting the menu.

Once the menu becomes sacred, future generations inherit not merely our institutions but our imagination of what is possible.

That is an extraordinary exercise of authority over people who do not yet exist.

They may know things we do not know.

They will encounter problems we cannot predict.

They may create institutions we cannot presently imagine.

They may discover that distinctions we consider fundamental were badly formulated.

We have no right—and, more importantly, no knowledge—with which to close their choice-space in advance.

No present generation owns the future menu.

Brett Hall’s Test

This is where Brett Hall’s Preservation Principle becomes decisive:

Do not destroy the means of error correction.

A political arrangement can be imperfect and still preserve the mechanisms through which it can improve.

Another can promise perfection while destroying the mechanisms by which anyone could discover that it was wrong.

The second is more dangerous.

So when evaluating a political choice, I increasingly ask not only:

Is this the correct decision?

I ask:

If this decision is wrong, can the people living under it discover that?

Can they criticize it?

Can they replace its authors?

Can they revise the law?

Can they alter the institution?

Can they question its premises?

Can they create an alternative?

Can they press Backspace without paying with their liberty, dignity or life?

If not, the choice may be closing the very route through which its errors could later be corrected.

The Third Door Must Also Be Allowed to Die

There is one final safeguard.

The Third Door must not become sacred.

Otherwise I will have committed precisely the error this series criticizes.

“Build the Third Door” could itself become a slogan deployed whenever someone dislikes the available choices.

That would be another protected explanation.

Sometimes two doors really are enough.

Sometimes the existing choice is unavoidable.

Sometimes the newly proposed C is worse than A and B.

Sometimes my diagnosis of the choice-space will be wrong.

So the Third Door Principle must contain its own Backspace Key.

It is a conjecture.

Criticize it.

Try to destroy it.

Improve it.

Replace it if something better emerges.

A method that cannot lose has already become part of the problem.

Build the Third Door

I began this series looking at two physical doors in Oxford.

I now see them differently.

The first door is not the enemy.

Neither is the second.

The danger begins when we mistake the doors that happen to exist for the architecture of reality itself.

That mistake reaches far beyond Oxford.

We inherit political systems, borders, doctrines, identities, institutions, categories, questions and historical explanations. They arrive looking like the landscape.

Then we choose inside them.

But human beings are not merely choosers.

We are explainers.

We criticize.

We discover errors.

We create knowledge.

And knowledge can make possible what yesterday’s menu could not contain.

That is why the Third Door is not C.

It is something more important:

the freedom to create C when A and B have failed.

Keep Proposition.

Keep Opposition.

Keep the argument.

Keep the facts.

Keep the criticism.

But when criticism reveals that the wall itself contains the error, do not spend eternity arguing about which existing door to walk through.

Press Backspace.

Reopen the problem.

Conjecture again.

And preserve the possibility of building a door that nobody could see when the argument began.

The Third Door Principle: When criticism reveals that the available alternatives are inadequate, do not force a choice among them. Reopen the problem, audit the assumptions that generated the menu, and permit the creation of a new conjecture. The Third Door is not another predetermined option. It is the preserved capacity to generate possibilities that did not exist before criticism began.

Two doors may be enough to begin an argument.

They must never be allowed to define the limits of what human beings can discover.