Picture an editor in the Northern Sung, a copy of Sun-tzu open on the table, who has decided the old book is right but no longer sufficient. He does not cross anything out. Instead he writes Sun-tzu's terse definitions at the head of each entry — in quotation marks, so you can always see the original — and then under each one he adds two thousand years of what happened next. Cavalry happened. Mass infantry happened. Sieges, river-dams, ambush-craft refined over forty generations of war. The man is Hsü Tung, the book is the Hu-ling Ching, and the chapter is "Analyzing Terrain." What he produces is not a new theory of ground. It is the old taxonomy turned into a working decision-grid — a field manual where every category of terrain now comes with what to do about it, scored against the accumulated record.1
The rubber-duck version: imagine the original Sun-tzu as a list of terrain names, and the Hu-ling Ching as the same list with an operations column filled in. That operations column is the whole contribution.
"Analyzing Terrain" tackles six fundamental configurations of ground: accessible, suspended, stalemated, constricted, precipitous, and expansive.2 Each entry begins with Sun-tzu's own definition, set off in quotation marks "for convenience," and then Hsü Tung adds the updated explication.3 The chapter's governing instruction is stated up front: "In order to employ the six methods for terrain, one must analyze their configurations and strategic advantages."4 You do not act on terrain. You analyze it first, derive its strategic advantage, then employ the method that matches.
The six run as follows. Accessible terrain is where both you and the enemy can advance freely; here it pays to occupy the heights, keep supply routes open to the rear, and guard the less-visible roads against unseen advances before engaging.5 Suspended terrain is where you can go forth but will find it hard to return; you check first whether the enemy is unprepared, then set ambushes to cut their supply before engaging — and if they are prepared and you sally out, "you will trip yourself up."6 Stalemated terrain favors neither side's advance; if the enemy withdraws here it is bait, so you do not chase but wait for their assault.7 Constricted terrain is a defensible mouth of a valley or the gorge between two steep hills; occupy it first and entrench, and if the enemy holds it fully deployed, wait — but if their deployment is thin, approach by another route.8 Precipitous terrain means holding the heights in security while the enemy is below and endangered; first there, you fight; arriving second, you halt, retreat, and reassess.9 Expansive terrain is when your fortifications sit far from the enemy's; provocations without advance signal ambush, and a withdrawal that doesn't engage cannot be pursued.10 The chapter closes with the line that gives the whole tradition its name: "When the ancients said that advantages of terrain are the real treasures in military tactics, this is what they meant."11
That is the first half. The chapter then narrows to two master categories that the next two Hu-ling Ching chapters take up in full — tenable terrain and fatal terrain — each developed into its own grid of when-to-use and when-not.12
The mechanism is re-glossing. Sun-tzu's definitions are kept fixed as the spine; the operational guidance is the flesh added on top, and the flesh is what changed over the intervening millennia. This is why Hsü Tung leaves the quotation marks in — he wants the reader to see the join, to know which part is the ancient invariant and which part is the new accretion of experience.13 The old book told you what kind of ground you are on. Hsü Tung tells you what that kind of ground now requires, given everything armies have learned to do on it since.
Tenable terrain is the clearest example of the grid in action. It is defined negatively: "the absence of fatal terrain to the front and rear, left and right, so that the supply routes and both advancing and retreating are all unobstructed."14 Sun Pin had added the survival rider — tenable ground is also where troops can safely find grass, firewood, and water.15 But Hsü Tung does not stop at the definition. He attaches a grid. There are six employments of tenable terrain and three non-employments.
The six cases where you can use tenable terrain: a cutoff force has penetrated deep into enemy land; the horses are picked and the men stalwart and drilled; the general's orders are enlightened and strict; your forces are strong and the enemy weak; the general's beneficence and trust have won the officers' following; and the officers and men take pleasure in warfare.16 The three cases where you cannot use it: the troops are thinking of their families; there is nothing in front to entice them so they hope to withdraw; or advancing is disadvantageous and retreating advantageous.17 The chapter ends the section with a rhetorical hook — "Can the advantages and disadvantages of tenable terrain not be analyzed?"18
What is striking is that most of the six employments are not about the ground at all. They are about the army's condition — morale, discipline, training, the general's standing. Sawyer points out that the army's ch'i, though not named by that term, runs through almost all of them, and that the first disfavoring condition (troops thinking of home) is plainly Sun-tzu's "dispersive terrain" smuggled back in.19 The doctrine quietly fuses two readings of "terrain": the physical ground and the psychological state of the men standing on it. To analyze terrain in the Hu-ling Ching sense is to analyze both at once.
Fatal terrain gets the same grid treatment, which the sibling tenable-and-fatal-terrain doctrine develops fully: four circumstances where the death-ground can be deliberately used, three where it must not be.20 The point for this page is structural — Hsü Tung converts every major category into the same four-column shape: definition, employments, non-employments, and the analytic question that forces the commander to score his own situation before choosing.
The Hu-ling Ching terrain doctrine is the hinge between a static taxonomy and an operational intelligence procedure. It takes the terrain-configuration taxonomy — the inherited classification of ground into named types — and adds the missing column: for each type, here is the decision. It gives the tenable-and-fatal distinction its formal grid structure, the employments-and-non-employments lattice. And it sets up Hsü Tung's contrary terrain practices, the same author's next move — where he argues that because this doctrine became so widely studied, a clever commander must sometimes invert it to surprise an enemy who has read the same manual.21
Look closely at the suspended-terrain entry, because it shows the re-glossing doing real intelligence work. Sun-tzu's definition is one clause: ground you can go forth into but find hard to return from.22 A static taxonomy stops there — a warning label. Hsü Tung turns it into a conditional operation. The first move is not tactical, it is epistemic: "first carefully determine whether the enemy is unprepared."23 Only after that intelligence check does the action branch. If they are unprepared, set ambushes to sever their supply lines, then engage advantageously. If they are prepared and you sally forth anyway, "you will trip yourself up."24
The whole entry pivots on a single piece of foreknowledge — the enemy's state of readiness — and the same ground yields opposite verdicts depending on it. This is the difference between a terrain label and a terrain doctrine. The label says "hard to return from, beware." The doctrine says "run the reconnaissance, and let the answer flip the ground from trap to opportunity." Hsü Tung has buried an intelligence requirement inside a geography category. You cannot execute the entry without first knowing something about the enemy that no map can tell you. Terrain analysis, in his hands, is inseparable from enemy analysis.
You hold the manual at the edge of a constricted valley-mouth. You read Sun-tzu's clause first — the gorge between two steep hills — and you place where you are: constricted terrain. Then your eye drops to the operations column. You are first here, so you set out the encampments and improve the deployments and stop worrying about a sudden onslaught, because the doctrine says first-occupier entrenches and waits.25 A scout returns: the enemy is moving up but has not filled the whole defile, their deployment is thin. You read the next line — for a less-than-full deployment, approach by another route — and you send a column wide rather than feeding men into the throat of the gorge.26 You never improvised. Every decision you made was already written under the name of the ground you were standing on, waiting for you to match it.
Now shift to tenable terrain. Your ground is clean — no fatal configuration on any side, supply open, grass and water at hand. The label says "advantageous." But you run the six-employment checklist anyway and you stall on the missing one: your men keep glancing rearward, talking of home. The non-employment list names it exactly — troops thinking of their families — and you read the verdict off the grid: this otherwise-perfect ground cannot be employed today, not because of the dirt but because of the hearts standing on it.27 You hold instead of committing. The doctrine just stopped you from fighting on ground that looked ideal and was not.
You can spot a misapplied Hu-ling Ching doctrine by one tell: the commander reads the terrain name and skips the operations column. He sees "accessible — advantageous" and engages, forgetting that the entry also demands he hold the heights and guard the unseen roads first.28 The second failure is reading the physical column while ignoring the psychological one — declaring ground tenable because the supply routes are open while his men are quietly thinking of withdrawing, which the doctrine flags as a non-employment regardless of how good the dirt is.29 The third, subtlest failure is treating the grid as a lookup table rather than an analysis: the chapter's whole insistence is on analyzing configurations and asking whether their advantages "can not be analyzed" — a commander who treats the entries as fixed answers rather than conditional verdicts has missed that suspended terrain flips on the enemy's readiness and tenable terrain flips on his own troops' morale.30 The doctrine fails the moment it is used as a map-legend instead of a decision procedure.
The most honest tension is preservation versus origination. Hsü Tung explicitly keeps Sun-tzu's definitions in quotation marks, advertising that the spine is not his.31 So how much of the operations column is genuinely two-millennia-later experience, and how much is Hsü Tung systematizing material already floating in the tradition? Sawyer treats the chapter as a real updating of the classical taxonomy, which is supported by the cavalry-and-mass-infantry developments he names as the driver.32 But the Hu-ling Ching as a whole is a Northern Sung compilation, and the line between Hsü Tung's innovation and his inheritance is not always drawable from the text alone — a recurring problem with this whole stratum of Sung military encyclopedism, which Sawyer notes elsewhere preserved passages "hardly a sentence original with the authors."33 A second open question: the six-employment list folds psychological conditions into a terrain chapter without flagging the move. Did Hsü Tung see "terrain" as genuinely including the army's morale-state, or has the category quietly expanded under him to mean "total operational situation"? The text rewards the first reading and never quite states it.
Sawyer's framing is that the Hu-ling Ching "commences with Sun-tzu's concepts and definitions but then provide[s] updated explications based upon two additional millennia of warfare experience."34 That is the convergence story — a clean line of cumulative refinement from the Art of War through Sun Pin to Hsü Tung. The text mostly supports it. But two frictions pull against the tidy line. First, Sawyer's cross-dynasty continuity claim — Warring States doctrine flowing intact into the Sung — is exactly the kind of synthetic through-line that needs independent verification; the quotation-marked spine proves Hsü Tung cited Sun-tzu, not that an unbroken living tradition transmitted the operational content across a thousand years. Second, Sawyer himself, discussing the neighboring Wu-ching Tsung-yao, warns that Sung compilers reworked unattributed material wholesale.35 If that editorial culture is the milieu, then "Hsü Tung's updated explications" may be partly a curation of pre-existing field wisdom rather than his fresh analysis — which would not weaken the doctrine, but would relocate its authorship from one Sung mind to a diffuse tradition. Sawyer does not resolve this, and neither should we.
To confirmation bias as an ancient problem. The genius of the suspended-terrain entry is that it forces a reconnaissance before the verdict — "first carefully determine whether the enemy is unprepared" — instead of letting the commander read his hope into the ground.36 That ordering is a structural defense against confirmation bias: the doctrine refuses to let you classify the ground as opportunity until you have gathered the disconfirming-capable evidence. A commander who wants to attack will see "accessible" and charge; the grid interposes a mandatory check that can return "no." The cross-domain insight neither field generates alone: the Hu-ling Ching is, in effect, a cognitive-debiasing protocol disguised as a geography manual — it hard-codes the sequence (analyze, then employ) precisely because the natural human sequence is the reverse (decide, then rationalize the terrain). The taxonomy's quotation-marked spine even serves the bias-resistance: by keeping Sun-tzu's definition visibly separate from his own gloss, Hsü Tung lets the reader audit where the inherited claim ends and the interpretation begins — a transparency move with no purpose except to let a later reader catch a motivated misreading.
To psychological resilience under survival and humiliation. The doctrine's quiet thesis is that ground and morale are the same object analyzed twice. Six of the tenable-terrain employments are really conditions of the men's state, and the first non-employment — troops thinking of their families — overrides perfect physical ground entirely.37 Resilience under ordeal works the same way in reverse: the "ground" a person stands on during a humiliation is partly external circumstance and partly internal state, and the external can be survivable while the internal collapses, or the reverse. The insight the pairing produces: in both domains, the decisive variable is not the configuration of the situation but whether the interior — the army's ch'i, the person's resolve — can hold it; a position is only as tenable as the morale occupying it. Hsü Tung's refusal to call a position tenable when the men want to go home is the military statement of a resilience truth — that the same circumstance is survivable or fatal depending on the state of whoever is inside it.
Sharpest implication: The Hu-ling Ching doesn't add new terrain — it adds a procedure, converting a static classification of ground into a conditional decision-grid where every entry hides an intelligence requirement and the same ground can flip from trap to opportunity on a single piece of foreknowledge.
Generative questions: