Picture a pickpocket. One hand taps you on the shoulder — loud, obvious, the hand you watch. The other hand, the one you forget about, lifts your wallet. Neither hand alone gets the wallet. The tap is useless without the lift; the lift is impossible without the tap. And here is the part the Chinese military tradition saw that the pickpocket already knows in his body: which hand is which depends entirely on where you are looking. The moment you start watching the lifting hand, the tapping hand becomes the thief.
That is zheng and qi — the orthodox and the unorthodox. Zheng (cheng) is the hand you fix on: the expected, the direct, the force that engages. Qi (ch'i) is the hand that decides the outcome: the unexpected, the deviant, the force that arrives from the angle you discounted. Sun-tzu compressed three thousand years of Chinese battlefield practice into a single sentence about these two hands, and the whole tradition this hub maps is a commentary on it.1
The canonical formulation is four characters that Sawyer renders: "one engages with the orthodox and gains victory through the unorthodox."1 You pin the enemy's attention and strength with the zheng — the army he can see, the assault he prepared for. Then you win with the qi — the flanking column, the night crossing, the force that materializes where he allocated nothing.
But the duality is not a tactic. It is a generative logic, and three features make it strange to a Western reader trained on "regular versus irregular" troops.
First: qi is not a kind of unit. A Western army has special forces and line infantry — fixed categories stamped on the soldiers. In the Chinese conception, the same frontal assault is orthodox when expected and becomes unorthodox the instant it is not — the designation lives in the enemy's mind, not in the troops. A force is qi because of where it appears relative to expectation, not because of what it is.1
Second: they produce each other in an endless cycle. Sun-tzu likens it to a ring with no beginning or end: orthodox becomes unorthodox becomes orthodox, like the five notes producing more melodies than anyone can hear, the five colors more shades than anyone can see. The two are not a pair of options you pick between; they are a wheel that, once turning, regenerates surprise out of expectation and expectation out of surprise.1
Third — and this is the inversion the West misses — qi is named first. In every Chinese formulation the unorthodox is the load-bearing term; the orthodox exists to set it up. The deviant is not a clever supplement to the real business of frontal war. The deviant is the business; the frontal engagement is the bait that makes it possible.1
The engine inside zheng/qi is a regress of mutual reading, and it is what makes the duality so hard to defeat. Sawyer traces the commentators chasing it: I plan to flank you (qi). But you anticipate the flank, so now the flank is what you expect — it has become orthodox — and my real unorthodox move is the frontal assault you abandoned. But you anticipate that reversal, so it flips again. This is the "I know that you know that I know" spiral, and it has no fixed floor.1
Two practical consequences fall out of the spiral.
One: the assignment can never be settled in advance. This is the deepest split in the whole tradition (developed across the Questions and Replies — Mutual Transformation page). If the enemy can deduce which of your forces is the decisive one, it has already lost its unorthodox character. So the designation must stay fluid until the moment of contact — "it lies with man, not with Heaven."
Two: there is a fight about allocation, and the fight is itself revealing. The commentators argued over the arithmetic — should you commit three-fifths to the orthodox and hold two-fifths as the unorthodox reserve? Ts'ao Ts'ao, the great early annotator, wanted forces pre-designated before battle; others insisted the split be decided live, in response to the enemy's dispositions.1 The disagreement is not trivial bookkeeping. It is the seam between two philosophies of war: war as plan you execute, versus war as conversation you improvise. Griffith, translating Sun-tzu for the West, reached for the most legible reading — zheng fixes, qi flanks — and Sawyer's point is that this reading, while true, is a flattening: it turns a living wheel into a single static maneuver.1
This page is the grammar the rest of the Chinese-unorthodox-warfare corpus is written in. It feeds:
The cleanest demonstration is the campaign the tradition itself reached for first. Han Hsin needed to cross the Wei River into hostile territory. He did exactly what an enemy expects an army to do: he massed boats, soldiers, and pontoon equipment at the obvious ford — the spot with solid footing, the place any engineer would choose. The Wei commander watched that ford and committed his defense to it.1
That visible build-up was the zheng. It was not a feint in the cheap sense of a trick — it was a genuine, resource-consuming, fully-believable preparation, and that is precisely why it worked. It absorbed the enemy's attention because it deserved the enemy's attention.
The qi was upstream, where Han Hsin's men were lashing earthenware jars to ridgepoles to float a force across at a point no one was defending. The crossing succeeded not because the jars were clever hardware but because the enemy's eyes — and therefore his troops — were three miles downriver, pinned by the orthodox.
Now watch the wheel turn, because this is the part the static "fix-and-flank" reading misses. The ford build-up was orthodox only because the enemy treated it as the main effort. Had the Wei commander ignored the ford and raced upstream, the ford would have become the real crossing — the orthodox would have become the unorthodox without Han Hsin changing a single order. The designation was never in the jars or the pontoons. It was in the Wei commander's allocation of attention. Han Hsin did not out-build him. He out-read him.1
You are negotiating, or competing, or planning anything against an adversary who is watching you back. Stop asking "what is my main move and what is my supporting move?" That question is already a Western flattening. Ask instead the two questions the wheel forces.
First, late at night before the thing happens, you lay your plan on the table and you ask of each element: where is the other side looking? Not "which of my moves is bigger" — which of my moves is expected. The expected one is your orthodox, whatever its size. The one landing where they have allocated nothing is your unorthodox, even if it is small. You discover you have mislabeled them: the big public proposal you were proud of is the tap on the shoulder; the quiet provision in clause nine is the hand on the wallet.
Second — and this is the discipline most people skip — you refuse to commit the labels until contact. You hold the split live. You walk in with the public proposal and with clause nine, and you watch which one they fix on. If they fight clause nine, it has become orthodox and your public proposal is now the live wire — you press it. If they fight the proposal, you let clause nine do the lifting. You have not decided in advance which hand is the thief. You decided in their eyes, at the table, in the moment. That is what "it lies with man, not Heaven" means in a meeting.
The duality fails the moment you treat the labels as fixed properties. The tells:
The evidence is Sawyer's three-thousand-year documentary trace, from the Spring and Autumn precursors through the Han paradigm cases to the PRC's modern extension. The tradition is unusually self-aware: it argued about the duality for two millennia rather than simply asserting it.
The central tension is the allocation debate already named — pre-designate (Ts'ao Ts'ao) versus decide-live (the Questions and Replies position). Sawyer does not resolve it, and it should not be resolved: it marks the real fork between war-as-plan and war-as-improvisation, and different commanders genuinely lived on different sides of it.
A second tension is interpretive: Griffith's Western "fix-and-flank" rendering is useful and true but lossy. The open question is whether the full mutual-transformation reading is a deeper truth or a later philosophical embroidery laid over a simpler battlefield heuristic. Sawyer leans toward the former; a skeptic could read the regress as commentators over-reading a practical rule of thumb.
Only Sawyer is cited directly here, but his account is built by interrogating the Chinese commentators against each other, and the page inherits that argument. The sharpest pairing is Ts'ao Ts'ao versus the Questions and Replies (Li Ching). Both are describing the same wheel — orthodox and unorthodox generating each other without end. But Ts'ao Ts'ao, a field commander running real armies, wants the designation fixed before the fight so the machine can be commanded; Li Ching, theorizing a century-plus of accumulated practice, insists the designation must float free until contact or it stops being unorthodox at all. The tension reveals something neither states outright: the duality is maximally lethal exactly when it is least manageable. The more you pin down which hand is the thief — the more commandable you make it — the more readable you become, and readability is the one thing that kills qi. The pickpocket who tells his apprentice in advance which hand will lift the wallet has, in that instant, made both hands visible.
Plainly: the same two-hand logic shows up wherever one mind has to act in front of another mind that is reading it back — combat, persuasion, and the body's own perception all run it.
The Sharpest Implication If qi is a designation living in the enemy's mind and not a property of your own forces, then you do not control whether your move is unorthodox. Your enemy does. You can build the most elegant flanking column in history and it is orthodox the moment he expects it — and there is nothing in your own camp you can inspect to know whether you have succeeded, because the answer is located inside his head. This is deeply uncomfortable for anyone who plans: it means the decisive quality of your decisive move is never yours to verify in advance. You are not arranging your own pieces. You are gambling on the contents of someone else's attention, and the only test is contact.
Generative Questions