History Without Philology

Joseph Levenson and Franz Schurmann. China: An Interpretive History / From the Beginnings to the Fall of Han. California 1969

In the divorce between philology and history, which in the US peaked in the year 1964, Joe Levenson's was one of the more acrimonious voices on the antiphilological side. It seems reasonable to inquire how a self-styled "historian" gets along without philology, in interpreting something as remote and text-dependent as classical China. Levenson's survey of that subject, published only a few years after the philology debate, gives us a glimpse of the answer. This estimate of that work was posted to WSW in 2003.



Along with others engaged in the Sinological palace revolution of 1964, Levenson ridiculed the philologists. He found them outmoded. He judged that the future lay with those like himself, who were based, not on a linguistically grounded understanding of China, but on the discipline (never defined) of "history." He wanted the philologists destroyed. And destroyed they were, at least in the United States.

In a key document of that revolution, the August 1964 issue of the Journal of Asian Studies (v23 #4), anthropologist G William Skinner spoke of traditional Sinology as already dead:

In recent years the cry has gone up: Sinology is dead; long live Chinese studies! And in this apothegm, by contrast with its prototype, a fundamental change is implied. Whereas old-time Sinology was given shape by its tools, so that Sinological skills defined the field and became an end in themselves, Chinese studies is shaped by its subject matter and Sinological skills are but means to analytic ends. Whereas traditional Sinology fostered uncritical immersion in a single civilization, modern Chinese studies brings at least that degree of impartial detachment which the comparative method implies. (p517)

This, of course, is nonsense. The tools of philology were invented in order to solve perceived problems with the source texts for history, not "for their own sake." In that same JAS issue, Levenson recorded his own resentment of the Sinologists, and spoke of the "historians of China" as the right sort of people for the job:

Twenty-five years ago, traditional Sinology was supreme in Chinese studies, but it was in limbo, withdrawn from the main life of the universities. The few people calling themselves historians of China were thought of by the Sinological colleagues to have taken up history because they were incompetent in the handling of difficult texts. . . (p513)

and Levenson goes on to further contrast the "Sinologist" (highly negative) with the "historian of China" (highly positive). What exactly is Levenson's position? Putting the most complimentary face on his view of "Sinology," we might suppose that he feels that Sinology, though it may have had a use in the past by working on "difficult texts," has now made the texts easy by its labors, so that the Sinologists themselves can be safely dispensed with, and the "historians of China," who alone know what the texts are good for, can proceed with their work. It's perhaps slightly ungrateful to dump the Sinologists so abruptly, but that's merely a point of style. The substance of the argument may well be true. Let's consider it.

A Test Case

If the world had been philologically made safe for the nonphilologists, then those who entered the easy terrain of Chinese antiquity would have produced, without straining after the solution of "difficult texts," irreproachable histories of early China: masterpieces of accuracy, edifices of correct interpretation. We happen to possess a textual record of what, in Levenson's opinion, such an edifice would look like. He wrote it himself (with Franz Schurmann). It is called China: An Interpretive History. It came out in 1969, shortly after the revolution, in the year of Levenson's death. Let's see how the Warring States areas of this book stack up, considered as a report of solved problems on every side, a sunlit and unproblematic landscape. What is the big picture like, the one that the philologists have been missing all this time? We take seven points.

Large Ideas

Our first point consists of one sentence.

The feudal ideal is unity. (p41)

It would be difficult, within the confines of one sentence, to be more wrong about feudalism (to avoid friction about the vexed term "feudalism," those who wish may substitute "indirect sovereignty"). The essence of indirect sovereignty is the ability to control large areas politically, without having the strength to dominate them militarily, or the political technique to incorporate them institutionally. Such a system is a stable but diverse pattern of local dominations, all the separate dominations adding up to effective general domination. It is a makeshift form of conquest, employed when actual conquest is beyond the power of the state in question. That political makeshift necessarily tolerates cultural disunity of all sorts, just as the mediaeval French kings tolerated it. Under those kings, the Burgundians and Normans and Angevins proceeded in their lively but culturally distinct ways. It would be more accurate to write "The essence of the feudal system is personally mediated and functionally limited central political control with a wide tolerance for local political and cultural diversity." To say that the feudal ideal is unity is to miss the entire point of feudalism.

Difficult Texts

Perhaps the second point will be handled better. Here it is:

The "Spring and Autumn" period takes its name from a chronicle, "Spring and Autumn Annals of the State of Lu." Confucius (551-479 BC), a native of Lu (in modern Shantung province), arranged this cryptic account of the reigns of twelve successive Lu rulers, all of them implicated in the internecine warfare of the times. (p41-42)

This is the usual myth, plus some additional errors introduced by the writer's carelessness. Not to get distracted by those new errors, let's consider the usual myth. The myth is that "Confucius arranged this cryptic account." He did not. Our received Spring and Autumn runs on for several years after Confucius's death. Confucius had nothing to do with the "Spring and Autumn" chronicle, either at the time of its composition or subsequently. That can be shown by sufficient exercise of skill in the "handling of difficult texts." Kennedy had refuted the basic proposition, in English and in 1942, decades before Levenson wrote his book. The Chun/Chyou is a primary source, an unedited and therefore fantastically valuable source, for these three centuries of the pre-Imperial period. Levenson does not appear to have examined it, or to have read the research reports of those who had examined it. Instead, he has made contact with the traditional myth about the text, and he has been content with that traditional myth. This is history?

Here is a third statement:

The literature that yields these terms and enshrines these ideas is primarily the Classics, treated as a canon first by Confucius, later (with additions and refinements) by Han and Sung Confucianists. (p43)

This is a wider version of the previous error. As we can learn by paying sufficient attention to the "difficult texts," Confucius, meaning the actual guy called Confucius, had nothing to do with what were later called the Classics. People in the middle and late Warring States sometimes claimed that he had done so, but this claim has no support in any text dating from within the century after Confucius's death. It arises only later, at a time when Confucius was beginning to be translated out of his original persona and into his later role as something of a culture hero. This device of a later claim without any earlier support is exactly the sort of thing that no "historian of China" worth his weight in weasels would accept without examination. To take only one of the more painfully obvious points, the list of Classics (apart from an earlier appearance in one of the Gwodyen documents) is first met with in the philosophical literature in the early 03rd century writer Sywndz. That list is still rudimentary. (It exists, in fact, in two different forms in SZ 1 and 2, of which SZ 1 is clearly later). The list of Classics continued to be fluid, both as to number and order, throughout Former Han. Everything we know from the late Warring States writings, and from the Han palace debates, which are voluminously recorded, shows that Confucius, the finite and historically tenable 05th century Confucius, could not possibly have treated these texts, or any subset of them, as a canon.

Levenson now moves on to mention one of this list of Confucius-established "Classics:"

(1) Book of Changes ("I-ching"). (p43)

Neither Sywndz (who of course we are not bothering to read) nor Mencius (who we are not bothering to read either; presumably these are "difficult texts," which seems to mean, texts written in Chinese) treated the Yi as authoritative. Neither Mencius nor Sywndz even mentioned the Yi. But both Mencius and Sywndz claim in some way to derive from Confucius. This perhaps puts a certain strain on the idea that Confucius himself recognized, or even read, the Yi. Dubs showed long ago, in a cogent article written in English and widely available in libraries, that Confucius had not done so. The way in which the Yi insinuated itself into the list of canonical works, and finally (in late Han) made it to the top of that list, as the most important of all the Classics, is an interesting story. It shows mumbo-jumbo winning out over an incipiently rational elite tradition. That development, the loss of reason and the victory of voodoo, is just the sort of development that a "history of China" might well set before us. But that story will never be told unless "historians of China" stop regurgitating the lies about Confucius, and start looking at the evidence for Confucius. Aren't "historians" supposed to be involved with evidence? We had always thought so. Levenson apparently has other ideas. History for him is something that doesn't take much time, or involve all these uncongenial "difficulties." It amount to repeating other people's rumors.


So much for texts. We have read four statements, and found four elementary errors. But fair is fair, and perhaps we should now skip ahead in the book, to pick up some nontextual matters that might show Levenson in a more flattering light. Here is a fifth statement, about a seemingly nontextual matter:

Then, much later, by the fifth or fourth century BC, pettier nobles on horses held the field [of battle]. (p67)

Oh, good. At last. Material culture, real horses, real battle, nothing oversubtle here, no "difficult texts" to puzzle over, no Sywndz (or Dubs) to read. Unfortunately, however, there is no evidence for riding astride, at any social level (great, petty, or peasant), in the 05c. There is scant and tenuous evidence for sub-elite riding astride in the middle 04c. There is no clear evidence that warriors on horseback played an important role, either in warfare or in the theory of warfare, until the 03c. No military treatise written before the 03c, or even within the early 03c, so much as mentions cavalry. The much-retold story of the Jau Kingdom adopting barbarian dress (to facilitate fighting on horseback) in the very late 04c also associates that adoption with an imitation of non-Sinitic (Hu) ways. That story in its present form is suspect (there are various versions of it, some of them purporting to describe another situation entirely), but the historical sequence as such is quite possible. From the sounder evidence, archaeological as well as textual, it is likely that elite horsemanship appeared in China around the end of the 04c, and though the Jau story may be merely emblematic, the development which it emblematizes seems to be not too far off chronologically. But the other thing which the story emblematizes, and which other evidence also supports, is the fact that cavalry did not arise as an evolution within Chinese military tradition, as Levenson (uncritically following the Sinitic unity myth) implies. Not at all. It arose by adoption from the non-Sinitic world. To verify this, evidence needs to be consulted, and this time it is archaeological evidence: horses and horse gear and cavalry swords and their furnishings found in tombs or in battlefield graves, that sort of stuff. For this task, it would seem, Levenson the "historian of China" is somehow too busy. The historians of Greece had acknowledged decades before this that archaeology was one of their important sources. The sort of "historians of China" envisioned by Levenson seem to operate in a realm somehow purer, and undoubtedly easier. Whether they get better results is precisely the question here being investigated.

Power and Politics

To investigate that question in a different setting, we may now look at the Chin Dynasty. Its establishment is beyond question the great event of Chinese antiquity. In fact, it is the event that put an end to Chinese antiquity. Here is something that the average trumpet-and-drum historian can be expected to get right, even if the large ideas and the concrete archaeological objects and the difficult texts prove elusive for him. Levenson writes:

It was in Ch'in, precisely where the public power of the state first encroached on the private power of nobles, that peasants began to be soldiers, no longer merely escorts and servants in war. (p67)

The use of armed infantry, and not merely menial foot auxiliaries to chariot warriors, which along with the invention of the bureaucratic state must rank as the key technical invention of Chinese antiquity, probably goes back in its origins to the very end of the Spring and Autumn. It was in fact this innovation that marked the end of Spring and Autumn, and ushered in the much more dangerous Warring States period. This would be in about 0500, give or take a little. Some would date the Warring States period from this vital innovation and its political underpinnings. That innovation would quickly become characteristic of the states in general. There is no evidence of a Chin invention of infantry, at this time or at any later time. The evidence, such as it is, tends to show that Chin was late, not early, in the process of converting to the new style army. The first textual record of an attempt to embody the large population into the class the of the militarily eligible implies a location not in far western Chin, but in far eastern Chi (it is reflected in the early Gwandz and the parallel Analects chapters, both of them eastern texts). Those documents date from the middle 04th century, and they show the new statecraft, and the new warcraft, getting their act together in a serious way. Consistently with these indications, the first large-scale battle to be fought by armies of the new type seems to have been the battle of Ma-ling, between Chi (eastern) and Ngwei (central) in 0343. Chi won. There is no sign of any participation of Chin in military encounters of this new sort until somewhat later. The distance between Chi in the east and Chin in the west is something like that between Boston and Berkeley. An American historian who proclaimed that a certain Tea Party took place in San Francisco Harbor would be courting ridicule from other American historians. Apparently the "historians of China" have nothing to fear, however, by getting things a couple of thousand miles wrong. Other "historians of China" see nothing wrong with it. On the contrary, the other "historians of China" greeted Levenson's book with raves and reverences.

They are only Chinese miles anyway, right? Who really cares? What difference does it make? To "historians of China," we are to conclude, it makes no difference at all.

Properly perceived, Chin is not an example of the Advantage of the First Move, as Levenson seems inclined to claim, but rather of the Advantage of the Second Move (if one wants a European label for this recurring topos, there is Trotsky's Law of Combined Development). The basic way in which Chin figures in the context of the rival states has here been drastically misperceived. Since Chin, as the eventual victor, was in a sense the most important of all the states, and thus the one that a "historian of China" should try above all others to get right, the lapse here is something more than a trifling detail. It is central.

Here is another Levenson statement about Chin, the last of the seven statements which we will consider:

It was the Emperor himself who in 212 BC initiated the project of"burying the scholars" (some four hundred), a mad assault on Confucianists (among others), to complement the reasoned proscription of Confucian ideas. (p71)

The Emperor may or may not have been mad, but he initiated nothing in this area. What he did, it seems, was to approve a recommendation by Li Sz. And what was Li Sz's recommendation? We have the text of it. In fact, and - here is where hateful philology comes back into the discussion - we have the text of it twice. One version is the Li Sz memorial as transcribed by Szma Tan, in one corner of the Shr Ji. It prescribes that circulation of the Confucian texts should be restricted to the officially approved scholars of the Chin Academy, and that private possession of them should be punished by a term at hard labor. The other version is the same memorial as tarted up and sensationalized by Szma Chyen, in another corner of the Shr Ji; it provides that those caught possessing the forbidden texts should be first executed, and then set at hard labor, a very neat trick if you can work it.

The contrast between the two versions of Li Sz's memorial had already been laid out (though admittedly not solved) by Bodde, in a work of - get this - 1938. Written in fairly clear English, and published by Brill, a prominent scholarly publisher. Something presumably hard to ignore. Levenson, never troubled by things hard to ignore, completely ignores it. The little complications of this Shr Ji doublet might be thought to be a problem easily within the powers of the most amateur of "historians" to unriddle (how, except in a Mozart opera, and even there as a joke at which the audience is supposed to laugh, do you first burn someone alive and then exile them for life?). Levenson does not unriddle them; he does not attempt to unriddle them. Confronted with (1) a genuine document and (2) a tarted up, cinematized, falsified document, his instinct as a "historian of China" is to choose the latter. He shows himself, as a historian, to be a sucker for the atrocity propaganda literature of earlier times.

And so it goes; pick your page.

The Bottom Line

Overall, Levenson's guiding insight about traditional China, explicit in both the 1964 article and the 1969 book, is that China is at bottom a yin/yang dualism between Confucianism (when people are in office) and Dauism (when they are out of it). This is essentially the Lin Yutang view of Chinese culture. On the way to reaching that view, Lin Yutang handles the sources much better than Levenson does. Lin doesn't find those sources particularly difficult, on the contrary, he moves among them with notable grace and poise. So if one wants to rest in that rather simpleminded conclusion about Chinese culture, Lin, not Levenson, would presumably be the citation of choice. It can of course be doubted that a Confucian/Dauist duality will adequately describe the Chinese mandarins of the late Empire. It can be further doubted that even an adequate description of the Chinese mandarins of the late Empire would be, in any modern sense, a "history of China" at the beginning of the Empire. Nobody purveying such a view should seek to be evaluated as a "historian of China." Levenson's invoking that rubric may well seem perilous in retrospect.

So, this was the 1964 revolution, and this is what came out of it. What lost out in it was the study of Chinese history. The study of Chinese history, in any serious sense, was tragically crippled by the palace revolution. It lost its eyes and ears, and was condemned to be led around by the nose by any sufficiently vivid cultural myth or propaganda pamphlet.


A polite person might have said in 1964 that classical Sinology deserved to survive just because it would be nice if it did so. Live and let live. An impolite but methodologically discerning person might have judged in 1964 that classical Sinology should get its priorities straight, and bestir itself about the many still unsolved aspects of the "difficult texts" and the no less difficult facts and interpretations which properly rest on the evidence contained in those texts, and stop fooling around, like the erudite but trifling Pelliot, with little things that happen to amuse them. Either way, we would still have had a classical Sinology in being, right through the Sixties. And if some difficult text had then come up that needed to be worked on, or if some hard question of historical interpretation had presented itself for solution, the chance of reading that text or solving that problem would at least have existed. The necessary people would have been around, and the "historians" could easily have inquired of them.

Instead of which, in our time, the problems drift on unsolved, and the "historians of China" bumble on unenlightened. It was not, after all, a very admirable revolution that produced this state of affairs. But effective it was. So effective that it will take a lot of work to get the enterprise of history, properly so called, back on its feet and functioning once more.

E Bruce Brooks
Warring States Project
University of Massachusetts at Amherst

Back to Reviews Index

21 Oct 2012 / Contact The Project