Warring States Project | A Summary of Methods and Results

A Summary of Methods and Results

What problem is the Project addressing? How does it approach that problem? What solutions it has reached, and what are their implications for our sense of the classical period? For the European scholars attending the WSWG 17 Conference (hosted by Leiden University in September 2003) Bruce wrote a short summary of our answers to those questions, by way of introducing ourselves. Here is a revised version of that summary.

The Large Problem

The problem is that traditional ideas about the classical Warring States period are often wrong, and never more so than with the texts on which our knowledge of that period is based. It follows that we don't really know very much about the formative period preceding the creation of the Chinese Empire in 0221.

If we read the Warring States texts as unitary compositions, we immediately encounter difficulties. Take, for example, the Analects. If all of that text reflects Confucius, who died in the early 05th century, why does it contain (in LY 12-13) whole lines also found in the 04th century chapters of the Gwandz, or (also in LY 12-13) a political theory identical with that argued by Mencius beginning in 0320? How can it respond (in LY 17:19) to the Mician school's critique of the three-year mourning custom, when Mwodz, the supposed founder of that school, is supposed to be, at earliest, a disciple of one of Confucius's disciples? Why does it duplicate almost exactly (in LY 18:5) a typically Dauist paragraph from the Dauist text Jwangdz, which dates from the 03c? This continual anticipation of ideas, and even stories, known to be later than Confucius is most disturbing to the conventional view of the Analects. In fact, it refutes that view.

And to take a point first noticed by Lyou Dzung-ywæn in the Tang dynasty, Why does the Analects refer to Dzvng Shvn as Dzvngdz ("Master Dzvng"), when Confucius is supposed to have been Dzvng Shvn's master? This cannot be the language of Dzvngdz's teacher. It can only be the language of Dzvngdz's own disciples.

Again, some WS texts seem to be aware of each other, in both directions. Thus, the Micians (in MZ 48:8) had criticized the Confucian custom of the three year mourning. Not only does the Analects respond to the Micians in LY 17:19, but the Micians answer back, in MZ 48:12, by holding the Analects argument up to ridicule, and deriding the Confucians as no more intelligent than a baby. The Mwodz must be at once earlier than, and later than, the Analects. This is logically impossible if the Mwodz is a unitary text.

In these and in many other ways, we can easily notice relations within one text, and between two or more texts, that the integral theory of those texts is powerless to explain.

An Interpolation

A first step in understanding such texts is to separate them into constituents, and purge them of later additions. Some of the later additions are in the form of later chapters, and some are in the form of interpolations: single passages inserted into previously written chapters. Here is an example of how one detects an interpolation in the Analects.

The First Page of the Analects, in a Tokugawa Japanese Edition

Among the various additions to texts, the interpolation is the easiest to understand and illustrate. The classic tests of an interpolation are three: (1) its content is inconsistent or anomalous in context; (2) if it is removed, the surrounding text will often close up to make a coherent and consecutive whole (unless the join has been smoothed by an unusually clever interpolator); and (3) it is often possible to suggest a reasonable motive for the interpolation. We here give the example of Analects 13:3.

(1). In this passage, the headstrong disciple Dz-lu is arguing with Confucius:

Dz-lu said, If the Ruler of Wei were waiting for the Master to run his government, what would the Master do first? The Master said, It would certainly be to rectify names, would it not? Dz-lu said, Is there such a thing? The Master is off the track. What is this about rectifying? The Master said, Boorish indeed is You [Dz-lu]. The gentleman, with respect to what he does not understand, will maintain an abashed silence. If names are not rectified, speech will not be representative. If speech is not representative, things will not get done. If things do not get done, rites and music will not flourish. If rites and music do not flourish, punishments and penalties will not be just. And if punishments and penalties are not just, the people will have nowhere to put hand or foot. Therefore, as to the gentleman: if he names something, it must be sayable, and if he says something, it must be doable. The gentleman's relation to words is to leave nothing whatever to chance.

Dz-lu's puzzled response suggests that "the rectification of names" is a novel idea. In Analects terms, it certainly is. Nothing like it occurs elsewhere in the text. The 13:3 picture of a society structured so that orders from above are carried out is in opposition with the picture, in LY 13:15, which deplores the state in which no one dares disobey the orders of the government. On the other hand, a top-down society is advocated by the contemporary statecraft texts, and Sywndz in particular has a whole chapter (SZ 22) on the rectification of names. Then on the basis of the content of the surrounding text, LY 13:3 is anomalous, and suspect as an interpolation.

(2) We now apply the second test. The surrounding passages may be summarized this way:

13:1. Dz-lu asks about government
13:2. Ran Yung asks about government

13:3. Dz-lu is baffled by the concept of rectification of names

13:4. The peasant skill of husbandry is not needed by ministers
13:5. The elegant accomplishment of memorizing poetry is worthless if not practical

It is obvious that apart from 13:3, we have successive pairs of sayings. Such a texture can be shown to be present in the rest of Analects 13, and in nearly all the other Analects chapters. This texture of paired sayings fully emerges in LY 13 only when such passages as the interruptive 13:3, which violate that pattern, are removed. Then on the basis of the form of the surrounding text, LY 13:3 is anomalous, and suspect as an interpolation.

(3) We now apply the third text. What could have been the motive for adding such a passage? One has already been suggested in stating why the content of 13:3 is suspicious: LY 13:3 affirms a doctrine important to the 03c philosopher Sywndz, who devoted a whole chapter to it, but is otherwise unknown to the Analects. But why should the Analects make a gesture to Sywndz? Here is one possibility. In 0254, Sywndz was hired away from Chi to be the governor of territory newly conquered by Chu; territory immediately to the south of Lu, and at the time an area over which Lu had had influence. Sywndz at once announced his intention to get rid of that dangerous thing, wrong thinking, in his territory. When in 0249 Chu completed its conquest of this territory, and Lu and its small neighbors ceased to exist as states, this policy was carried out, and many schools of thought (several posthumous Mencian schools, the Dau/Dv Jing group, and the Micians) ceased to produce new text. Between those dates, as an attempt to ingratiate themselves with the new power situation, some texts produced in Lu and vicinity wrote new material which might be more acceptable to Sywndz. LY 13:3 is the most obvious of these passages. Then a motive for its addition becomes obvious: the Analects school was bargaining with Sywndz for its survival. Under such conditions, much is possible that would be out of the question in more normal times.

Our three tests are thus met. Results so produced, and so checked, may be called literarily consistent.

Which is all very well, but is there outside confirmation, of this or any similar result?

Confirmation

Sometimes there is, and here is one case, not directly involving the Analects, but rather a text with which the Analects is itself involved: the Dau/Dv Jing.

Bamboo Slips from the Chu Tomb at Gwodyen

In June 1990, E Bruce Brooks revealed in a lecture, and in July 1994 he published in SPP 46, a view that the Dau/Dv Jing text, a work in 81 short chapters, grew by accretion over the period from c0340 to 0249. This specific theory is linked with, and consistent with, theories of the other major texts, including the Analects, so that a test of the DDJ theory is to some extent a test of the whole system of proposed dates.

The DDJ accretion theory contains a prediction. It predicts that if a copy of the DDJ from between the years c0340 and 0249 were discovered, it would lack some of its higher-numbered chapters, and the number of chapters missing would depend on the exact date of the copy. Such predictions are open to testing.

A test presently appeared. In 1993, a text of the DDJ was recovered from a tomb near the capital of the old state of Chu (some of the bamboo strips are shown above). The tomb could be dated circumstantially to the period between 0298 and 0278. The theory had predicted that a DDJ text from that period would probably contain material later than DDJ 53, but not anything beyond DDJ 70. The validity of our entire construct was thus resting on one question: is the Chu DDJ complete or not? And if not complete, is it incomplete in the way the theory predicts?

When the results of the 1993 discovery were released in 1998, the Chu DDJ turned out to be incomplete, and in the way we had predicted. That is, it contained nothing beyond DDJ 66. None of the last 15 chapters of the DDJ text were represented in the Chu selection. This is exactly what was predicted by the DDJ theory proposed in 1990 (before the discovery of the Chu text had been made), and published in 1994 (before the contents of that discovery were known). The prediction was made without knowledge of the evidence. The confirmation was exact.

Some have objected that perhaps the Chu DDJ text (which contains only 33 segments of DDJ, and thus omits much material from the range DDJ 1-66 also) were omitted by chance. But the chance that a selection of 33 items from a larger group of 81 items would include nothing at all from items #67 to #81 inclusive is 1 in 7,384 (for that computation, see the separate Arithmetic page). Statistically, this is so unlikely as to be impossible. The DDJ theory, and the system of text chronology of which it is a part, are thus stunningly confirmed by the Chu DDJ.

Such confirmed predictions, from evidence which could not have been known at the time the prediction was made, are considered decisive in the hard sciences. What kind of science history may be, we do not know, but pending a determination, we consider this confirmation to be decisive in our area also.

As a group, our philological findings also turn out to be consistent in another way, and this adds to the evidence that they are correct. For that other way, see the next page: