Translating Sun Tzu, or, What is Technical Literature?

Over on birdsite John F. Sullivan noticed something which readers of ancient tactical manuals or surveyor’s manuals or medieval painters’ handbooks and fencing books have also noticed.

Whenever I pick up a new Sun Tzu translation, the very first thing I compare is two verses found in chapters 7 & 11 of the text. Why? It turns out they are exactly identical verses, should be rendered identically, but only rarely are. It gives us an indication of how careful and thorough a translator is. The verse itself is not one of Sun Tzu’s most memorable. It is basically composed of three thoughts—understand your neighboring rulers’ intentions, conduct a detailed assessment of the enemy’s terrain, and employ local guides to assist you in traversing enemy land. It does, though, reinforce two main themes of Sun Tzu’s military thinking—detailed assessments of the enemy situation (primarily terrain) and a preference for deep offensive invasions as the ideal military strategy.

John F. Sullivan @JohnF_Sullivan@twitter.com on15 June 2022 https://nitter.it/JohnF_Sullivan/status/1536928990879182848#m

But in both chapters the verse is basically a non-sequitur. It doesn’t flow naturally with the thoughts directly preceding and following it. This hints at something very few translators discuss— that the text is not very well organized, obviously corrupted at several points, and might not be as logically consistent as we would like to believe. The fact that the extant text repeats an identical verse without explanation is noteworthy, yet many translations never inform the reader about this anomaly. Moreover, many translators seem unaware of the repetition at all, often subtly changing the meaning between verses or in some cases inexplicably coming up with two contradictory interpretations. The quantity of Sun Tzu translations doesn’t necessarily reflect the quality of analysis we have subjected the work to. Unfortunately, this does not just impact third-rate translations, but also those done by accomplished scholars.

Take Michael Nylan’s recent effort. Nylan is an excellent scholar and translator of ancient Chinese texts. In her Introduction she tells us that she and her group worked for three years on the effort and that “no word escaped the group’s notice, and this is how it should be.” But how did she do on the translation consistency test? Below is the verse as found in chapter 7.

“So, if you don’t know the designs of the local lords, you cannot predict how best to engage them. And if you don’t know the local topography- the mountains, forests, ravines, marshes, and wetlands- you won’t know where to march your troops. Nor can you take full advantage of the local terrain, unless you use local guides.”

I will focus on the first line of the verse. In the footnote to this rendering, Nylan tells us that “as it is not clear whether the nobles are enemies or allies, the translation here remains neutral.” Compare this with her translation of the same line found in the chapter 11. Here there is no footnote, but the neutral stance is now dropped. Neighboring states are no longer potential enemies to “engage” with, only potential allies.

“Unless you know their intentions, you cannot enter into alliances with the rulers of neighbouring states in advance. Unless you know the lay of the land – its mountains and forests, its passes and defiles, its wetlands – you cannot deploy an army on it. Unless you can employ local scouts, you cannot take advantage of the terrain.”

The main point is that a reader will remain unaware that these two verses are exactly identical in both sections of the original text. It’s part of a larger problem in which Sun Tzu, in particular, does not merit (receive? – ed.) the level of scrutiny and carefulness found in other works.

Not to continue picking on the Nylan translation–in many ways it is very good–but in chapter 2 Sun Tzu outlines the resources one needs to initiate war, including, in Nylan’s rendition, “ten thousand armored soldiers.” But the original text clearly states that you need TEN ten-thousand (十萬) armored soldiers, which is the Chinese way of computing “100,000.” Nylan’s version leaves the commander with 90,000 fewer troops than originally recommended. Too much effort is spent on pumping out new translations, when what we really need is more careful commentary of the original text, and better contextualization of its meaning compared to other contemporaneous works and what we know about the military history of its era.

John F. Sullivan @JohnF_Sullivan@twitter.com on15 June 2022 https://nitter.it/JohnF_Sullivan/status/1536928990879182848#m

We are used to books which are composed over a few months or a few years by one or a handful of authors and then reproduced as hundreds or thousands of identical copies. We expect a book to be original, and we harshly criticize and sometimes punish people whose books contain other people’s words. But many ancient technical texts are repetitive and contain lists of practical advice which is similar to other lists. Vegetius has his General Rules of War and constant appeals to earlier authorities, Maurice has a whole book of wise sayings for generals which often repeat themselves. What is going on?

It seems that when people want to write about how to do something practical, they often start by writing short recipes or formulas or puzzles. These are useful in a world where writing materials are expensive and most people don’t read for hours every day. They jog the memory of someone who remembers roughly how to do something but not every detail. They do not include definitions because anyone who needs them can ask someone who already knows. These lists of recipes circulate on their own and add elements or lose elements as each owner finds different things useful. These lists are often anonymous, and if they are associated with a name sometimes that name is just a famous person in the field. Medievalists debate whether the French chef named Taillevent wrote a cookbook because the oldest version survives in a manuscript written before he was active – did he edit and expand this list, or was his name just used to make this list sound especially impressive? The closest thing to these lists in our culture is Grandma’s box of recipes in the kitchen.

At some point people who read for pleasure discover these lists. They often do two things: either they compile and organize short lists into a big book, or they expand the lists by rewriting them in less dry and concise language, defining terms, and adding flowery prefaces and conclusions. The second of these kinds of text or treatise is more like a book as we understand it. But if you know what to look for, you can see traces of the anonymous lists poking out like bumps in the floor of a renovated building. Sullivan suspects that the text we think of as The Art of War was put together by editing and expanding previous lists or short treatises, and that sometimes something from the sources ended up in two places or a disagreements between the sources were not resolved. Other scholars suspect that the European book on painting by ‘Theophilius Presbyter’ was written this way, because its wordy preface is followed by much more concise language and then by things which might not be based on experience at all just on written authorities.

If we read a text like The Art of War and expect to see the work of a single genius with a unique vision, we can have trouble seeing how much of it was borrowed from earlier teachers and earlier workers. The polished treatises survive because they are lovingly protected in libraries (or buried in tombs!) by rich people, not kept in a dirty kitchen and workshop and recopied whenever they become dingy by working people. Scholars and antiquarians want the authentic original version, but most workers trying to solve practical problems just want the version which is most relevant to their challenges today. We have to work back from the polished treatises to see the type of texts which their authors might have drawn upon but which do not survive. We also have to recognize that while a printed edition presents us one canonical text, every manuscript of a practical work like Vegetius can be different because readers wanted useful knowledge not a perfect copy. The discovery of a bamboo manuscript of Sun Tzu in a tomb gave both opportunities and problems for scholars who wanted to know what The Art of War looked like before it became a printed text for the imperial examinations.

Scholars such as Karel van der Toorn see signs of this method of composition in many parts of the Hebrew Bible. Existing material could be placed side by side, interwoven, or expanded at the beginning and ends. In ancient Near Eastern scribal culture, this was not seen as wrong, but as a way of respecting powerful texts and skilfully creating something that was both old and new.

My writing is not supported by the King of Wu just by generous supporters. All donations and shares are appreciated!

Further Reading: Sullivan recommends Mark Edward Lewis, Writing and Authority in Early China (State University of New York Press, 1999). To Lewis, “Warring States texts worked like a loose-leaf binder into which one inserted essays or notes by different hands, and added, removed, or rearranged the materials to suit the evolving interests of the compiler.” Sullivan has an academia dot edu page

Edit 2023-06-27: see also Christopher C. Rand, Military Thought in Early China (SUNY Press, 2017)

Edit 2023-12: Ralph D. Sawyer and Mei-chün Lee Sawyer (Seven Military Classics 1993, Ancient Chinese Warfare 2011) seem like another important pair of names. Ralph D. Sawyer has a YouTube channel on early Chinese warfare https://www.youtube.com/@sinowarfare/videos

(scheduled 22 December 2022)