In the middle of the Southern Sung, around the turn of the thirteenth century, a man named Shih Tzu-mei sat down to write commentary on the old military books and did not stop until the commentary was long enough to be called lectures. People copied them, reprinted them, carried them to Japan, and studied them for centuries.1 What makes him matter is not a battlefield trick. It is that he is the moment the doctrine of spying stops being a clever chapter in one general's book and becomes a subject — something taught, glossed line by line, handed down with the names of earlier teachers attached. Before Shih Tzu-mei you read Sun-tzu. After him you study "Employing Spies" the way a student studies a canonical text, with the great commentators arrayed in the margins like a faculty.
Shih Tzu-mei wrote the first work to treat the Seven Military Classics "as canonical."2 That word is the whole point. He took a loose tradition of military writing and froze part of it into scripture — a fixed set of authoritative texts that you interpret rather than merely consult. His commentary on Sun-tzu's "Employing Spies" chapter, drawn from his Sun-tzu Chiang-yi, is the espionage doctrine getting its canonical form.
His method is synthesis, and he is open about some of it and quiet about the rest. He weaves in the views of the Ten Commentaries edition of Sun-tzu, the thoughts of Li Ching from Questions and Replies, "sometimes citing the speakers, at other times merely co-opting their phrases."3 And he leans heavily — without saying so — on Li Ch'üan's T'ai-pai Yin-ching and its chapter on roving agents.4 So the page is really about a process: how one Sung scholar braided Sun-tzu, Li Ch'üan, and Li Ching into a single continuous teaching, and in doing so showed "the continuity of military thought" running from the Art of War through the T'ang and into his own age.5
Watch the mechanism on a single sentence. Sun-tzu names five spies — local, internal, turned, expendable, living. Shih Tzu-mei does not just repeat the list. For each type he sets two earlier teachers side by side: Tu Yu's gloss and Chang Chao's. On the local spy, Tu Yu explains you use the enemy's local people to learn his "vacuities and substantialities," and can run them as reconnaissance scouts; Chang Chao adds that you recruit them by "thoughtfully and ceremoniously treating them well," and cites the T'ang taking prisoners to attack Korea.6 On the internal spy, Tu Yu says rely on those who "lost their government positions"; Chang Chao expands the list to resentful officials, "favorites and concubines no longer loved," and men who "love profits."7 The same paired treatment runs through turned, expendable, and living agents — Tu Yu's terse definition, then Chang Chao's worked example.8
That pairing is the engine. Shih Tzu-mei is not generating doctrine; he is curating it — taking the dispersed commentary tradition and rebuilding it as a teachable, line-by-line apparatus where every one of Sun-tzu's claims arrives already annotated by the masters. Then he reaches past Sun-tzu entirely and grafts in Li Ch'üan's three "roving agents," concluding that though Li Ch'üan's categories "are special, the way in which they all learn the enemy's affairs are one."9 The braid is the product. He is showing the reader that the whole corpus says one thing.
The richest piece of canonization in the lectures is economic. Sun-tzu opens "Employing Spies" with a number: mobilize a hundred thousand troops and "seven hundred thousand families" are unable to pursue their farming.10 On its own the figure is just rhetoric. Shih Tzu-mei does what a teacher does — he shows the arithmetic behind it.
The number, he explains, comes from "the ancient well-field system": nine families farm nine plots arranged in a grid, "the middle plot being the duke's share." Eight families work the eight outer plots. If one man among those eight is "impressed into the army, the remaining seven had to supply his clothes and food." So every soldier pulled from the land drags seven families off their own work — and a hundred thousand soldiers therefore idle seven hundred thousand families.11 The ratio is exact: one in, seven out.
This is the case study in what canonization does. Sun-tzu asserted a cost; Shih Tzu-mei grounded it in an institution. He converts a slogan about the price of war into a calculable economic argument — and that argument is the whole justification for spying. If mobilization is this ruinous, then a general who "begrudged… a hundred catties of gold to seek out the enemy's true situation" is committing "the ultimate inhumanity."12 The well-field math is the floor the spy-economics stands on. To teach the doctrine, Shih Tzu-mei had to teach the economics under it, and that is exactly the move that turns a chapter into a classic.
Imagine the page as it was used. A student has Sun-tzu's italicized line in front of him — there are five types of spies — and below it Shih Tzu-mei's voice, unhurried, building the world around the sentence. First the warriors who break enemy lines, then the strategists who know the Tiger Secret Teaching, then the logisticians who move "cows and horses," then the diviners who read the wind. Each gets his due. And then the turn: as for the one "who spy out the enemy's closely held secrets… how can they not be treated as resources?"13 The student feels the spy slotted into a complete order of military talent. The teacher names the five types again, calls their combined use "spiritual methodology," calls them "a ruler's treasures."14 Then Tu Yu speaks. Then Chang Chao answers. The student is not being told a fact; he is being walked through a tradition, each authority entering on cue. That is canonization happening in real time — and it is why these lectures were copied for seven hundred years.
The failure mode is buried in Sawyer's own description and Shih Tzu-mei half-commits it. The lectures incorporate "extensive, though unacknowledged, material from Li Ch'üan's chapter on roving agents."15 When a teacher co-opts a predecessor's phrases without attribution, the canon he builds starts to launder its own sources — and a reader two dynasties later cannot tell Sun-tzu's voice from Li Ch'üan's from Shih Tzu-mei's. The smooth continuity that makes the lectures valuable is the same smoothness that hides the seams. A doctrine that reads as one unbroken teaching may in fact be three teachers stitched together, and the stitching is exactly what canonization is built to make invisible.
The second warning sign is reverence hardening into closure. Once "Employing Spies" is canonical, the gloss stops being a question and becomes an answer. Tu Yu said, Chang Chao comments — and the pairing that was once a live argument between commentators sets, over centuries of reprinting, into a fixed catechism. The lectures preserve the tradition and, in preserving it, risk ending it.
Sawyer's framing leans on continuity: the lectures matter because they "show the continuity of military thought from Sun-tzu's Art of War through Li Ch'üan… and into his own work."16 That is a real and useful observation, but it is also the kind of cross-dynasty continuity claim that flatters the source. Continuity is partly something Shih Tzu-mei manufactured by braiding the texts together; reading the braid as evidence of an underlying unity risks mistaking the editor's craft for the tradition's nature. Sawyer notes the unacknowledged borrowing in the same breath as the continuity, and the honest reading holds both at once — the continuity is real and partly produced. The seam between "the tradition was continuous" and "Shih Tzu-mei made it look continuous" should stay open.
To Sun-tzu's Five Spies. That page holds the original doctrine; this one holds its reception. The structural relationship is text-and-commentary, and the productive tension is what commentary adds and subtracts. Sun-tzu's five-spy typology is spare to the point of cryptic — five names, a few lines each. Shih Tzu-mei's apparatus buries each name under two layers of gloss (Tu Yu, Chang Chao) and worked historical examples. Place the two pages together and an insight neither holds alone appears: the meaning of Sun-tzu's five spies as later China understood them is not in Sun-tzu at all — it is in the commentary. The turned agent "is the foundation of the five agents and the crux of spycraft" is Tu Yu's claim, not Sun-tzu's,17 yet it became the orthodox reading. The doctrine the vault inherits is the glossed doctrine. Knowing where Sun-tzu ends and the commentators begin is the difference between reading the Art of War and reading what the Sung schoolroom made of it.
To Chinese Military Theory — Historical Development. This page is a dated, named instance of the broader process that page tracks: the long arc by which Chinese military thought consolidated, classicized, and got institutionalized as the Seven Military Classics. Shih Tzu-mei is the hinge — "the first work to treat the Seven Military Classics as canonical."18 The handshake produces a sharper claim than either page makes alone: canonization is not a neutral act of preservation but a strategic one. By fixing the texts and pairing every line with authoritative glosses, the Sung made military knowledge teachable, examinable, and transmissible across the empire — turning scattered battlefield wisdom into state-reproducible doctrine. The development of Chinese military theory is, at this hinge, the development of a pedagogy; Shih Tzu-mei is where the corpus stops being read and starts being taught.
Sharpest implication: Doctrine is not made only by the original strategist; it is made again, and decisively, by the commentator who decides which earlier voices get to stand in the margin — and the version posterity inherits is the curated one, with the seams hidden.
Generative questions: