One line from Sunzi's Art of War has been quoted for twenty-three centuries:

"The highest use of arms attacks the opponent's strategy; next, his alliances; next, his troops; the lowest is to assault walled cities."

It is usually read as the summit of war wisdom — the finest generalship dismantles the other side's plan, and storming a city is the last resort. That reading is not wrong. But it mistakes the first rung of a sequence for the last.

Attacking strategy is not the summit. There are four levels above it.

And something matters more than that: war is not on this sequence at all.

1. Five Levels

These five are not five kinds of soldiering. They are five ways a collision can be handled. With each step up, the action called "arms" shrinks and the thing called cultivation takes up more room.

The highest use of arms attacks strategy (shang bing fa mou). Sunzi's position. The collision is handled at the level of planning, and what gets attacked is the opponent's design, his calculation, his disposition. Both sides still recognize each other as subjects; arms exist as a contest of intelligence. This is already a considerable achievement — it lifts collision off the level of brute force.

Arms transcended attack through disputation (chao bing fa bian). Two subjects who already agree that persons are ends interrogate each other, unsettle each other, and make each other disclose. Here "attack" does not mean defeat; it means making disclosure happen. The endpoint of a disputation is not who won, but that both parties have shown themselves more clearly.

The medieval universities turned this into an institution. Aquinas structured the Summa so that each question first states the objections against his own position in their strongest form, and only then answers them. He was not looking for the weakest opponent; he was looking for the strongest. Public disputatio worked the same way — a thesis was attacked in the open, and the respondent had to meet every objection live. The aim of that kind of attack is not to destroy the other's argument. It is to force both positions into clarity.

Arms at the limit attack one's own thought (ji bing fa si). This is where the subject of the sentence changes, and it is the hardest rung on the whole sequence. What gets attacked is not the opponent's thinking but one's own. The question is no longer "how do I beat him" but "why have I not developed far enough that this collision would become meaningless to him?"

The focus moves from across the field back to oneself. The opponent stops being an object to be handled and becomes a mirror for reflection.

Marcus Aurelius wrote the Meditations in a tent on the Danube frontier. He held the strongest military machine in the world and was fighting a border war that had dragged on for years, and what he wrote on the page is almost entirely interrogation of himself — where was I wrong today, why did that provoke me, on what basis do I assume my judgment is sound. The contrast between position and action is exactly what this rung looks like: every instrument of force in hand, and the eyes turned inward.

Arms grown mysterious do not attack; walking away comes first (xuan bing bu fa). Here the act of attacking is cancelled outright. Walking away is not the pragmatism of fleeing a fight you cannot win. It is recognizing that this collision is not a field you should have entered, and returning to your own path.

Four centuries of Mennonite and Amish migration is a long-scale demonstration — from Swiss valleys to Alsace, to Prussia, to the Russian steppe, and on to North and South America. Each time they met conscription, coercion, or a demand they could not accept, their answer was neither to fight nor to submit. It was to move. The cost was enormous: clearing new land, rebuilding a livelihood, learning another language, over and over. But what they were protecting survived. This is not weakness. Four hundred years of not fighting without disappearing takes more resilience than fighting a war does.

The sage has no arms and is hidden within the cosmos (sheng zhe bu bing). Here the subject changes completely. This is not a higher grade of soldiering; it is a person who has stepped out of the sequence of arms altogether. The sage does not need to avoid arms — arms are simply irrelevant to the state he is in. Being hidden is not concealment. It is existing in a way that does not require being seen.

I have no historical example for this rung. That absence follows from its own definition: a mode of existence that does not require being seen leaves no identifiable trace in the record. Any example that can be produced probably belongs to a lower rung. The gap is not missing evidence. It is the nature of the level.

2. What the Sequence Is Counting

On the surface, the sequence counts the stepwise dissolution of the act called "arms." What it actually counts is something else: the structural room cultivation occupies, expanding rung by rung.

At the strategy level, cultivation unfolds through a contest of intelligence — both sides sharpen their judgment against each other. At disputation, it unfolds through mutual disclosure. At self-thought, it unfolds through reflection, and that reflection is itself an act of deepening. At non-attack, it unfolds by walking away, so that growth can continue in the right field. At no-arms, cultivation has filled the subject's entire life and there is no position left for arms to occupy.

There is a second, equally important way to count: the sequence redefines "the other side," one rung at a time.

At strategy and disputation, the other side is an object to be handled — the eyes are on him. From self-thought onward, the other side becomes a mirror: that he wants a collision is his affair; that I have not developed far enough to make the collision pointless to him is mine. Responsibility comes back to my side of the field. At non-attack, the other side is part of a field to be recognized and left. At no-arms, the other side is a partner in mutual sharpening — the question is no longer whom I collide with, but whom I grow alongside.

Both countings describe two faces of the same thing.

3. The End of the Sequence Is Not the End of Collision

One point has to be settled here, or the whole sequence reads as a doctrine of avoidance.

The end of the sequence is not the absence of collision. It is the absence of unnecessary collision.

Real friction between subjects, real grinding against each other, is a necessary condition of cultivation. Without it, a subject shrinks into self-enclosure and growth stalls. The sage hidden within the cosmos has not retreated into an isolated loop of self — he is still in mutual sharpening. Only the form has changed: no longer war, no longer deceit, no longer competition.

What is removed is unnecessary collision, not collision itself.

4. So Where Is War

Now the opening line can be answered.

War is not on these five rungs.

The five are the different ways cultivation unfolds inside a collision. War is what repairs cultivation after it has been broken. It is not some honorable slot on the sequence; it sits outside it, a backstop invoked only because cultivation needs a route back.

This distinction settles the relationship between war and the sequence: war is a protector, not a substitute. It does not replace the upward movement; it gives that movement a chance to resume after an interruption. When a subject's cultivation has been pressed so far that none of the five rungs can even be engaged — because cultivation itself can no longer continue — war appears as the backstop, and its purpose is to restore the conditions under which the climb can start again.

The previous essay noted that war and law hold isomorphic positions in this framework. The isomorphism can now be stated in full: both sit outside the sequence, both are protective mechanisms, and neither is cultivation itself. Law handles ordinary collisions so that everyday conflict does not break cultivation. War handles extreme collisions so that cultivation has a route back once it has already broken.

5. A War Always Begins as Defense

One last thing has to be established, because it fixes the scope of the five essays that follow.

The origin structure of war is asymmetric.

A systemic purpose — national interest, the realization of an ideology, the propagation of some civilizational form that "ought" to prevail — operates by expanding. Its goal is never completed inside the current border; it always requires more: more resources, more compliance, more room to spread. This is not the wickedness of any particular system. It is the internal logic of that structural type. A systemic purpose that stops expanding cannot sustain itself.

When expansion is blocked, it does not actually stop. It goes dormant — continuing ideological mobilization, military buildup, internal suppression — and waits for conditions to return. Its "peace" is a pause in expansion, not the disappearance of the intent.

External expansion and internal suppression are also two faces of one thing. To expand, a system must turn its own people into executing units, because complete subjects might refuse mobilization or question whether the expansion makes sense. So a structure that systematically suppresses its own population will expand outward eventually; and a structure currently expanding outward is necessarily suppressing internally.

On the other side, a subject that genuinely operates by treating persons as ends has no necessity of expansion. Its purpose is that every subject exist as an end — and structurally that cannot be achieved by suppressing others. The opposite: suppressing others violates it directly, since the one suppressed is also a subject. It grows where it stands, and does not need expansion to sustain itself.

Put the two layers together and the real structure of war becomes clear: it always begins with the expanding side attacking, and the defending side's resistance is defense, not attack.

This also fixes the scope of what follows. Collisions between two expansionist systems — which is what most historical wars look like — fall outside the range. Neither side reasons from persons-as-ends, so these standards do not apply to them; those matters are handled by diplomacy, treaty, and the balance of power. Nor do two subjects who already agree that persons are ends need war: their collision is resolved at the level of disputation, because they have no conflict of purpose, only differences in how to realize the same one.

From cultivation's angle, this can be said more sharply. When two expansionist systems collide, cultivation is not present in the field, so we have no wish to discuss it. When two subjects who truly recognize persons as ends collide, cultivation is unfolding healthily, so there is no need to discuss it. Exactly one situation remains — cultivation is present, it faces being swallowed, and it has no instrument of suppression available to it.

That is the only "cannot-not" in this entire series.

6. The Defender's Structural Problem

One thread is left hanging here, and it runs through the five essays ahead.

The defender's resistance is defense — but resistance itself creates a difficulty. Resistance needs organization, mobilization, wartime discipline, centralized decision. Structurally, those requirements look a great deal like the systematized machinery being resisted.

So the defender runs a risk: in the course of resisting one machine, it may build itself into another one just like it.

The four theorems, the three ending forms, and the standard against self-annihilation — none of these exist only to restrain the attacker. They exist just as much to keep the resisting side from sliding into the structure it set out to resist.

Because if that happens, then even a military victory leaves the world with one fewer defender and one more suppressor.

Next

If war is the backstop, what is it aiming at?

The answer is more counterintuitive than it sounds. It is not winning.