Two cities sit on a plain. One has a notable valley before it and a high mountain rising behind it; its inner ground stands high while the land falls away outside the walls. The other sits in vast marshland with no notable valley and no mounds, its rear sloping down into a valley with no mountains on either flank. To Sun Pin, writing in the middle Warring States period, these are not just two towns. The first is male — a city that cannot be attacked. The second is female — a city that can.1 He never met either one. He's reading their bodies from the shape of the land around them.
That is the whole trick of this page. Sun Pin took the question "which walls do I assault?" and answered it before the army arrives, from topography alone. A city's defensibility, he claimed, is legible in its setting — in marsh and mound and valley and the lie of high ground — the way a doctor reads a patient's constitution from the outside.
Sun Pin's "Male and Female Cities" is a topographic classification scheme that sorts fortified cities into the un-attackable ("male") and the attackable ("female" or "vacuous") by the configuration of the terrain that frames them. Sawyer presents it as the historical counterweight to Sun-tzu's blanket reluctance to assault cities: where Sun-tzu, a generation or two earlier, treated walls as costs to be avoided, Sun Pin's dual categorization reflects "the growth of cities as economic and strategic centers" and the arrival of effective siege weaponry like the catapult.2 Campaign armies could no longer flow unhindered through sparse countryside; they were thwarted by fortified strongholds that now carried real military and economic weight. So Sun Pin built a way to triage them — strategically weaker ones (female) could be attacked, stronger ones (male) avoided.3
One careful qualification runs through Sawyer's reading and should not be lost: Sun Pin never said female cities should always be attacked, nor designated them as more than preferred targets.4 The scheme tells you which walls are worth your effort; it does not order you to throw yourself at every weak one. And the classificatory principles are, Sawyer stresses, "simply topographical" — Sun Pin is reading dirt and water, not garrison rolls or morale.5
The classification, drawn from the recovered tomb text, runs roughly like this. Male (cannot be attacked): a city amid small marshes, lacking high mountains and notable valleys but ringed by moderate mounds on four sides; a city with a notable valley before it and a high mountain behind; a city whose inner terrain is high while the ground falls away outside; a city with moderate mounds within.6 Female / vacuous (can be attacked): a city with a notable valley behind it that lacks high mountains to left and right; a city in vast marshes with no notable valleys or mounds; a city between high mountains but lacking valleys and mounds; a city with a high mountain in front and a notable valley behind, the ground rising before it and descending to the rear.7
The grammar underneath the list is consistent once you look for it. A male city has its defensive features arranged so they protect it: marsh and mounds that break an approach, a mountain at the back that can't be turned, high interior ground that lets defenders look down on attackers while the falling exterior ground exhausts anyone climbing. The configuration does the defending. A female city has the same kinds of features arranged so they betray it — a valley at the rear that opens a back door, an absence of flanking mountains that leaves the sides bare, marsh with no valley to channel an assault, ground that rises before the wall and drops behind it so that taking the front gives you the high ground over the interior.
Think of it as the city's constitution rather than its garrison. A garrison can be reinforced overnight; the lie of the land cannot. Sun Pin's claim is that the permanent fact — the topography — is a more reliable index of attackability than the changeable facts of who holds the wall today and how much grain is inside. The high mountain behind the male city will be there next season; the brave general on its parapet might not. By scoring the constitution, Sun Pin gives a commander a readout that doesn't decay between the scouting and the assault.
This is why the scheme is an intelligence artifact and not a tactics manual. It doesn't tell you how to storm a female city — it tells you which cities are worth the cost-benefit calculation in the first place. It is a filter that runs before the assault math, narrowing a long list of enemy strongholds down to the ones whose terrain hasn't already decided the question.
This page is the topographic leg of the vault's three-part read on urban warfare, and it's the one that converts raw ground into a target list. It feeds directly into City Evaluation and Targeting, whose first filter is "can we even take it?" — Sun Pin's male/female score is one of the cleanest answers to that filter, computed from terrain before any other input arrives. It also sits beside City Defense and Siege Psychology as its mirror image: Sun Pin scores the city from the outside by its land, while the defense page scores it from the inside by its morale and hope of rescue. A male city with a broken garrison and a female city with fanatical defenders are both possible — which is exactly why the topographic read and the psychological read are different pages.
To the terrain corpus, this page contributes the city as a special case of terrain reading. It is the same skill as Sun Pin's Treasures of Terrain — reading the land as an object with intelligence value — aimed at the densest, most human-built piece of ground there is. A walled city is just terrain that fights back, and Sun Pin reads it with the same eye.
Take the cleanest male entry and the cleanest female entry and hold them against each other. The male city: amid small marshes, no high mountains or notable valleys, but "moderate-sized mounds about its four quarters."8 Picture it — a town girdled on every side by low rises, with soft marsh to slow an approach. The mounds aren't dramatic, but they're complete: four quarters covered, no open face. An attacker has no clean line in. The constitution is sealed.
Now the female counterpart: "a notable valley behind it that lacks high mountains to its left and right."9 Picture this one — the rear opens into a valley, and the flanks have no mountains to anchor a defense. The valley behind is the fatal feature. It's a corridor into the city's back, and with no flanking mountains the defenders can't even shield the approach to it. An attacker reads this from a ridge a mile off and knows where the assault goes before he's seen a single defender.
The instructive thing is that both cities might have identical walls and identical garrisons. Sun Pin's classification ignores both. What separates them is purely the arrangement of land — closed mounds on four quarters versus an open valley at the rear. The same skill that reads a battlefield for fatal ground reads a city for its fatal opening, and in the female city, the land has already written the assault plan. That is the analytical payload: defensibility is, to a large degree, decided by geography that no commander chose and none can change.
You ride the perimeter of an enemy stronghold at dawn, staying out of bowshot, and you are doing arithmetic with your eyes. You note the marsh on the western approach — slow ground, a point in the city's favor. You count the mounds: north, east, south — and then a gap to the southwest where the land falls away clean. You ride to the back and find the valley behind the walls, and the absence of any mountain to the left or right that could screen it. Your scribe is writing as you talk. Valley to the rear. No flanking mountains. You've seen enough of Sun Pin's categories to feel the verdict settle: this is a vacuous city, a female city, and the southwest gap and the open valley are where it will fall. You haven't decided to attack it yet — that's the general's cost-benefit call, not yours. But you've handed him the thing he needs first: this wall is one the land has left attackable. Tomorrow you'll ride to the next stronghold and find a high mountain at its back and high ground within, and you'll write a colder word. Male. Leave it.
The first failure is mistaking the topographic read for the whole read. Sun Pin scores the land and the land alone; a commander who treats "female city" as "weak city" forgets that a vacuous setting can hold a resolute garrison and a strong setting can hold a broken one. The diagnostic sign is an attacker who assaults a female city and is shocked by the defense — he read the ground and skipped the people. The land says attackable; it never said easy.
The second failure is treating the female classification as an order rather than a preference. Sawyer is explicit that Sun Pin never designated female cities as more than preferred targets.10 A commander who reflexively attacks every topographically weak city — because the scheme labeled it attackable — has confused a filter with a command, and will spend his army on weak cities that gain him nothing, the very error City Evaluation and Targeting exists to prevent. The classification narrows the list; it does not approve the assault.
The third sign is staleness in the wrong direction: trusting a topographic read while ignoring man-made works that change the equation. A female city that has built new earthworks, set caltrops, and pulled everything inside the walls is no longer the soft target the land alone described. Topography is permanent, but the defenders' preparations are not — and a classification that only sees the permanent will miss the wall that was hardened last week.
The text is a tomb recovery, and that shapes how much weight it can bear. Sawyer notes the "Male and Female Cities" material comes from fragments, and that other situations "similarly categorized for their attack potential" were intermixed with the city list before being deleted from his excerpt.11 So the scheme we read is a curated slice of a messier original. The category labels — male / female / vacuous — are evocative but their precise topographic boundaries (how big a "moderate" mound, how "notable" a valley) are not pinned down in the recovered text, which makes the scheme more of a trained eye's heuristic than a measurable rubric.
There's a live tension with Sun-tzu that Sawyer foregrounds and does not resolve. Sun-tzu's reluctance to assault any fortified city and Sun Pin's willingness to triage and attack the weak ones are usually read as a progression — cities growing more valuable, siege weapons improving.12 But the two doctrines coexisted in the canon, so a commander inherited both: avoid the city / sort the city and take the female one. The vault holds both. Sun Pin doesn't refute Sun-tzu; he adds a filter that makes Sun-tzu's reluctance selective rather than blanket.
Finally, the cross-text continuity — Sun-tzu to Sun Pin to Wei Liao-tzu — is Sawyer's synthetic frame across a span of recovered and transmitted sources, and the precise relationship of the Sun Pin fragments to the Art of War (Sawyer notes they "may have been an integral part" of it) carries the usual manuscript uncertainty. The scheme is real and the categories are in the text; the neat developmental line connecting it backward and forward is the translator-synthesizer's reading.
Sun Pin's scheme converges with the later cloud-ch'i city-reading and with Wei Liao-tzu's defender psychology on one shared conviction: a city's takeability is legible from outside, before the first ladder goes up. Three different reading systems — topographic form, atmospheric omen, rescue-certainty — all promise to score the wall in advance. That shared confidence, that the decisive judgment is made before the siege rather than during it, is the tradition's real signature. The tension is twofold. The male/female categories come from tomb-recovered Sun Pin fragments whose completeness and exact relation to the Art of War stay uncertain; and Sawyer's neat developmental line, linking the scheme backward to Sun-tzu and forward to the encyclopedias, is the synthesizer's reading. The categories are genuinely in the text. The continuous conversation around them is partly built.
Handshake with Physiognomy and Character Reading. Sun Pin reads a city's defensibility from its outward topographic form exactly as the physiognomist reads a person's character from the outward form of the face and body — both claim that a stable inner property (a city's strength, a person's nature) is legible in a fixed outer arrangement. The structural parallel is precise: marsh-and-mound-and-valley is to the city what brow-and-jaw-and-bearing is to the man, an external configuration treated as an index of an internal truth. The tension here is the same one that haunts physiognomy — the gap between form and present state. The face can lie; the man can dissemble against his "type." So can the city: a topographically female city can have hardened itself with new works and a resolute garrison, just as a "weak-faced" man can have steeled himself. Both reading systems are strongest on the permanent and blindest to the recent change. Setting them side by side shows that Sun Pin's scheme is a physiognomy of cities — and inherits physiognomy's exact failure mode, which neither page names on its own: confident reading of a fixed form that the subject has quietly altered.
Handshake with Fatal Terrain and Death Ground. Both pages read land for a verdict that's already half-written before any fighting — but they read from opposite chairs, and the comparison sharpens what "fatal" means. Fatal-terrain doctrine reads ground that will destroy the army standing on it (the gorge, the net, the pit); Sun Pin reads ground that will destroy or save the city built into it. The shared mechanism is topographic determinism — geography deciding outcomes that men then merely execute. The insight the pairing produces: a male city is, in effect, fatal terrain for the attacker — the mountain at its back and the high interior turn the approach into the same kind of death-ground the fatal-terrain pages warn a general to keep at his back and force on the enemy. So Sun Pin's classification can be re-read through the fatal-terrain lens: assaulting a male city is leading your own army onto fatal ground, and the "cannot be attacked" verdict is the fatal-terrain warning wearing a city's name. Neither page alone shows that the same topographic logic governs both where an army dies and which walls it must not climb.
Sharpest implication. A city's defensibility is, to a large degree, a property of the land it sits in — permanent, unchosen, and readable from a distance before any other intelligence arrives. Sun Pin turned "which walls do I assault?" into a topographic classification that can be computed from a ridge a mile off, making geography the first and most durable line in a city's intelligence file.
Generative questions.