Education About Asia: Online Archives

How Do We Know about Early China?

Back to search results
Download PDF

Study of the distant past always begins with critical examination of sources. In the present case we might ask, “How do we know about early China?” We are speaking, after all, of a time removed from our own by two or more millenia. The sources upon which we base such knowledge are simply traces of the past, and these traces are largely of two types: written texts and physical objects.1 While physical objects should never be overlooked and give material reality to the written word, texts are ultimately of greater value. Jean Bottéro’s comment about the ancient Mesopotamian world can in general be applied to China as well:

The only data that come from the past and that can respond directly to all our questions are documents—texts. They are exact, detailed, precise, and most often unquestionable . . . . Documents therefore comprise the surest, the most complete, the most indispensable sources for our rediscovery of the past.2

But before placing this much confidence in the written texts of early China, we need to understand the formidable problems associated with their transmission and interpretation. Early Chinese texts are sometimes divided into two categories, so-called “hard texts” and “soft texts.” In the case of the former, what we see before us today are the very marks that were written or inscribed long ago, in some cases perhaps by the original author or by scribes working under the author’s direction. The most famous of these hard texts are the oracle inscriptions, carved into turtle shells or cat­tle scapulae,3 and the bronze inscriptions, sometimes cast and some­times scratched on the surface of early ritual bronze vessels.4 The oracle inscriptions are our best source for studying the last centuries of the Shang dynasty (circa 1250–circa 1045 BCE), while the longest and most informative bronze inscriptions are a valuable source for investigating the first centuries of the early Zhou dynasty (circa 1045 to circa 750 BCE). In addition, numerous other hard texts have appeared in recent decades, mostly excavated from tombs: legal documents, admistrative records, inventories of funeral items, calendrical notes, covenant texts, etc., and many more such writings are likely to be dug from the earth in the decades ahead.5 Reading hard texts requires not only a mastery of a difficult lan­guage that, at least in the case of the bone and bronze inscriptions, is significantly different from the language of the somewhat later soft texts called “classical Chinese,” but also mastery of epigraphy, for these texts are written in early script forms and sometimes present challenges on the most basic level of decipherment.

The following three oracle inscriptions are typical of such texts and reflect both the value and the limitations of these texts:

1. “Jiawu, divining: ‘The Many Yin people ought not under­take the harvesting of the millet.’”

2. “Crackmaking on jiachen day: ‘On the next yisi day, make an offering to Father Yi of penned sheep.’”

3. “Divining: ‘It is Father Yi who is cursing Fu Hao.’”6

The first of these examples is an act of divination about the mundane but critical activity of when to begin the harvest, the sec­ond concerns the timing of an offering to a deceased ancestor, and the third indicates that a deceased ancestor can pronounce a curse upon the living, a curse that proper sacrifice might avert. However revealing texts such as these might be of the practices and beliefs of the people in Shang times, they are limited in two important ways. First, both the oracle bone, such as those presented above, and the bronze inscriptions tend to be maddeningly brief, highly formulaic, and restricted in subject matter. Second, related to the previous limi­tation, these records should always be studied with their original rit­ual or religious function firmly in mind. For example, one scholar has recently cautioned against the practice of reading bronze inscriptions as records of historical events produced primarily for future generations.7 Whatever these records might tell us today, he argues, they were most probably addressed originally to deceased ancestors and contained information that was meant to impress, maybe on occasion even trick, the departed.

Soft texts, in contrast to hard texts, come to us through the hands of numerous intermediary editors and copyists. The earliest extant editions of most of these texts date from more than a millenium after they were originally written. As is the case with other ancient tradi­tions, techniques of textual criticism, developed and elaborated both in China and in the West, must be applied to guarantee the most reli­able possible versions of these sources. Recent discoveries of early versions of portions of these soft texts, usually in the form of inscribed bamboo strips, complicate this picture even more. These excavated texts are not exactly hard texts, for they themselves are most probably copies of previous versions, but they date to an earlier time than the soft texts we have customarily read. Many of these finds consist of bundles of bamboo strips that roughly parallel a sec­tion, but only a section, of some extant soft text. This and other evi­dence adduced by earlier scholars leads to the conclusion that the soft texts we possess today were actually edited together from smaller bundles at an early stage of textual transmis­sion. The homogeneity and “good order” of the text as it has been transmitted to us might mask a com­plex heterogeneity. As indicated above, scholars have long argued that numerous early Chinese texts are made up of layers—that is, some sections are much earlier or much later than other sections. To give one famous example, Chinese scholars demonstrated several centuries ago that the Confucian Analects, surely one of early China’s most important books, was composed of several lay­ers, and recently two scholars in the West showed just how complex that early history may have been.8 What this means is that such texts, which came about through gradual accre­tion, do not reflect a particular moment of time but a development that might have gone on over cen­turies.

Inscribed cattle scapula excavated at Anyang. (circa 1250 BCE).
Source: Cambridge Illustrated History of China By Patricia Buckley Ebrey.
Cambridge University Press 1996.

The texts used to reconstruct and talk about early China are filled with problems, as I tried to indicate above, and those problems should never be overlooked as we contemplate the Chinese past. For example, the Spring and Autumn period of Chinese history (722–481 BCE) have been enormously important in our reconstruction of early Chinese civilization. This is the period of Confucius (551–479 BCE) and, at least by traditional accounts, Laozi as well. It also is a period when the older Zhou political order had largely disintegrated and a number of indepen­dent states sometimes cooperated and sometimes vied across the early Chinese landscape. The Spring and Autumn is the first well-attested era of political regionalism in China, a regionalism that only intensified during the subsequent Warring States period and reappeared from time to time in later Chinese political history. Numerous histories have been written of this critical time, some at very impressive length.9 Several sources have been used for its his­tory, including bronze inscriptions and a wealth of excavated mate­rial objects. But by far the single most influential source is a text known as Zuo Commentary (Zuo zhuan), a work of just under 200,000 characters, the longest from pre-Han China. Precisely what is this text? As the title indicates, it was transmitted as a com­mentary to another text, Spring and Autumn Annals, a chronologi­cal list of notices of particular historical events compiled in the ancient eastern state of Lu and supposedly produced (or more like­ly re-edited) by Confucius himself. Most scholars today quite rea­sonably accept the reliability of Spring and Autumn Annals, for the short entries found in this text appear to derive from official state reports that are roughly contempo­rary with the events they describe. The problem is that these some­times cryptic entries have almost always been understood through much more extensive and “infor­mative” commentaries, particular­ly Zuo Commentary.10

So what can we say of the provenance of Zuo Commentary? First of all, there is some reason to believe that the text was not originally produced as a commen­tary at all, but it is almost certain that the author or, more likely, authors, had Spring and Autumn Annals close at hand. Although these authors lived perhaps 150 to 200 years after Spring and Autumn Annals concludes, events recorded in a single sentence in Spring and Autumn Annals are often presented with considerable detail in Zuo Commentary. The latter text, for example, includes many long and rhetorically com­plex speeches, purported to come from Spring and Autumn period speakers, even though there is not a single speech found in Spring and Autumn Annals. In this long and detailed text we would seem to have our key to the presenta­tion of Spring and Autumn history, and that is how Zuo Commen­tary has often been used. But we must ask, as always, what is the source of all of the material found in this influential text and is it reliable? There is no easy answer. One prominent contemporary scholar believes that the speeches found in Zuo Commentary are part of an oral tradition and that the transmission of those speeches was constantly adapted to later realities, making them much more a source for the study of Warring States (475–221 BCE) ideology than Spring and Autumn attitudes and events.11 Another believes the speeches may well derive from actual records kept during the Spring and Autumn period and that these, minus the insertion of a few later forgeries, are a reliable guide to what he calls “pre-Con­fucian” thought.12 One thing seems quite certain: the Zuo Com­mentary is neither wholly fictitious nor wholly reliable. Hints are found here and there throughout the text that the authors may have had more in the way of genuine Spring and Autumn records than just the Lu annals, but it is maddeningly difficult to sort this out, and they, moreover, give us virtually no help in this endeavor— which brings us to the issue of authorship.

One reads Zuo Commetnary longing for the guidance of an overt authorial voice. Note, for example, the following story which appears in Lord Huan year 15 (697 BCE):

Zhaizhong was monopolizing the government, and the Earl of the State of Zheng worried about this. He sent Zhaizhong’s son-in-law Yong Jiu to kill him. Yong was going to offer Zhaizhong a feast in the outlying dis­trict. Yong’s wife, named Ji, learned of the plot and said to her mother, “Whom should one hold dearer, a father or a husband?” Her mother said, “All men can be hus­bands, but one has only a single father. How can they be compared?” Consequently, she reported to Zhaizhong, “Master Yong has left his home and is going to offer you a feast in the outskirts. I am suspi­cious of him and so report.”

Zhaizhong killed Yong Jiu and exposed his corpse near the pond of the Zhou clan. The Lord of Zheng loaded the corpse into his wagon and left the state, say­ing, “He involved his own wife in the plot. He ought to have died!” In the summer, Lord Li fled to the State of Cai.13

This narrative is extremely terse and moves with great speed. In fact, it can only be followed with careful attention. No interpre­tive voice comes to our rescue, helping us to contextualize or understand the full significance of the story or to tell us where this particular story came from. “According to learned Persians, it was the Phonecians who caused the conflict . . . . However, where the Io incident is concerned, the Phonecians do not agree with the Per­sians,” so Herodotus reports in his Histories, written in Greece per­haps a century earlier than Zuo Commentary appeared in China.14 Herodotus remains at our side throughout his history chattily guid­ing our understanding and repeatedly telling us where he gets his stories (this does not mean, of course, that they are always true, but at least we sometimes know who the liars were!). The Zuo Com­mentary authors are almost entirely absent. Events take place and speeches are delivered, as in the passage cited above, with virtually no authorial interference. They never tell us whether they got a par­ticular account from “people in the state of Jin” or “people in the state of Qin,” although we the readers might sometimes surmise one or the other.

Still, we can learn something about these hidden authors, their concerns and their prejudices, from their writings. First, they were preoccupied with violations of traditional ritual. Passages such as the following abound: “That the Prince of Qi escorted the Lady Jiang to Huan is not the proper ritual. Whenever the daughter of a lord is married to an equivalent state and is a younger sister of the reigning lord, then a high-ranking marshal would escort her to show ritual courtesy to the former ruler.”15 The fact that the Prince of Qi escorted his younger sister is only noted in the text at all because it was a violation of ritual protocol. In fact, it is probably not wrong to say, as has one specialist, that in the ideal world of the Zuo Commentary authors there would be no history, only a rep­etitious record of ritual performance. Time in such a conception would be swallowed up by ritual rhythms, and history, at least as we understand it, would end.16

Second, the authors of Zuo Commentary do create a heroic type. Their hero is the good minister who can infallibly read the future course of events in the behavior, particularly the ritual behavior, of his contemporaries:

In the thirteenth year, in spring, Qu Xia of the State Chu set out to assail the Luo. Dou Bobi saw him off. As he was returning, Dou Bobi told his charioteer, “The maréchal is certain to be defeated. He lifts his feet high. His intentions are not firm.” So Dou Bobi met with the Viscount of Chu and said, “We must reinforce the army.” But the Viscount of Chu refused to do this.17

Dou Bobi, the typical Zuo Commentary hero, reads the future in the inappropriate gait of the maréchal and boldly warns his ruler, but to no avail. The experienced ruler knows that defeat will fol­low, which of course it does. Indeed, there are few surprises in this text, as almost all outcomes are signalled by amazingly insightful “predictions” of wise ministers. The authors, we might surmise, identify with this group of ministers and have created this text as a means of enhancing their own prestige and political influence, as a means of warning rulers that survival might well depend on heed­ing their precocious advice. “Objective history” is, of course, a chimera, for all historians have their prejudices and perspectives. That being said, the didactic intent of Zuo Commentary is exceed­ingly strong and colors the presentation of the past at virtually every turn.

Consequently, if we are to use Zuo Commentary at all as a source of Spring and Autumn history, we must do so with extraor­dinary caution. And the same is true for all other early Chinese texts. Their reliability must be subjected to the most rigorous examination, and they should constantly be checked against hard texts and physical objects that come to us from the world they pur­port to describe. A final example might illustrate how our image of the past can be revised through just such a process. For centuries, historians of the short-lived but incredibly important Qin dynasty (221–206 BCE) have followed the lead of writers who lived in the subsequent Han dynasty (202 BCE –221 CE) in decrying the harsh “legalistic” practices of the Qin government. The notorious First Qin Emperor, now enshrined in film as well as in popular fiction, supposedly built the Great Wall with the blood and bones of count­less corvee laborers, burned books, buried scholars, and imposed harsh standards and cruel laws upon his people.18 Such actions are usually cited as the reason the Qin dynasty fell so quickly! Many scholars (although by no means all) forgot that Han writers felt the need to justify the Han conquest of another dynasty that had been so recently founded and thereby might have exaggerated the faults of the dynasty that preceded them. Several years ago, sections of a Qin legal code were found inscribed on bamboo strips. Examina­tion of those texts, “hard texts,” has softened or at least complicat­ed our view of Qin legal institutions.19 Qin law assuredly had its harsh aspect, as did the legal codes of later dynasties, but to attribute the fall of the Qin primarily to excessive “legalism” almost certainly oversimplifies a complex historical event.

The study of ancient history, whether that of China or any other ancient civilization, will always be a difficult and problemat­ic endeavor. As with all such history, the final word can never be spoken. Scholars will continue to re-examine old sources, check them against the most recent excavations, and rethink conclusions that had been accepted earlier almost as final truths. While we should never speak of the distant past too confidently, careful scholarship can yield genuine advances—we can draw nearer to an accurate picture of a past that will in some ways always remain hidden from full view.


1. It is tempting to say “strictly limited” rather than “largely limited,” but we should keep Collingwood’s famous statement in mind: “. . . everything in the world is potential evidence for any subject whatever” (The Idea of History, [Oxford: Oxford University Press, 1946], 280). Along this same line, Lucien Febvre has said, “History is made with documents, to be sure—when there are some. But it can be made, it should be made, from all that the historian can per­mit himself to utilize. Therefore, from words. From signs. From landscapes and from tiles. From the forms of fields and from weeds. From eclipses of the moon and from the collars of draft animals. From the expertise of geologists on rocks and from the analysis of swords in metal by chemists” (My own translation from Combats pour l’histoire [Paris: Agora, 1953], 428).

2. Jean Bottéro, Clarisse Herrenschmidt, and Jean-Pierre Vernant, Ancestor of the West: Writing, Reasoning, and Religion in Mesopotamia, Elam, and Greece, translated by Teresa Lavendar Fagan (Chicago and London: The University of Chicago Press, 2000).

3. The standard work on this subject is David Keightley, Sources of Shang History: The Oracle Bone Inscriptions of Bronze Age China (Berkeley: University of California, 1978).

4. On this topic, see Edward L. Shaughnessy, Sources of Western Zhou History: Inscribed Bronze Vessels (Berkeley: University of California, 1991).

5. A good survey of much of this newly excavated material is Edward L. Shaugh­nessy, ed. New Sources of Early Chinese History: An Introduction to the Reading of Inscriptions and Manuscripts, Early China Special Monograph Series 3 (Berkeley: University of California, 1997). A complete list of excavated material can be found in Enno Giele, “Database of Early Chinese Materials,” on the Web at

6. Adapted slightly from the translations of Michael J. Puett in To Become a God: Cosmology, Sacrifice, and Self-Divinization in Early China (Cambridge and London: Harvard East Asian Center, 2002), 42–47.

7. Lothar von Falkenhausen, “Issues in Western Zhou Studies: A Review Article,” Early China 18 (1993).

8. E. Bruce Brooks and A. Taeko Brooks, The Original Analects: Sayings of Con­fucius and his Successors (New York: Columbia University Press, 1998).

9. While I have in mind here several Chinese-language histories, an English-lan­guage example, albeit much briefer than many Chinese treatments of the Spring and Autumn period, is Hsu Cho-yun, Ancient China in Transition: An Analysis of Social Mobility, 722–222 BC (Stanford: Stanford University Press, 1965).

10. A good English-language sample from sections of Zuo Commentary can be found in Burton Watson, trans., The Tso Chuan (New York: Columbia Univer­sity Press, 1989). A full translation, albeit difficult to use, can be found in James Legge, trans., The Chinese Classics, Vol. 5, The Ch’un Ts’ew with The Tso Chuen (1879; rpt., Taipei: Southern Materials Center, 1985).

11. David Schaberg, A Patterned Past: Form and Thought in Early Chinese Histori­ography (Cambridge and London: Harvard University Asia Center, 2001).

12. Yuri Pines, Foundations of Confucian Thought: Intellectual Life in the Chunqiu Period, 722–453 BCE (Honolulu: University of Hawai’i Press, 2002).

13. Lord Huan 15.2 (697 BCE). All translations from Zuo Commentary are my own and are part of an ongoing, new translation of the text undertaken in collabora­tion with David Schaberg and Li Waiyee. Readers will note that this story con­tains a common folklore motif, appearing for example in India as well, where a woman is forced to choose between the life of her husband or her father. The “correct” choice is of course one’s father. A woman can always find another husband, but she can only have one father!

14. Herodotus, The Histories, trans. By Robin Waterfield (Oxford: Oxford Universi­ty Press, 1998), 3.

15. Lord Huan 3.6 (709 BCE).

16. Schaberg, A Patterned Past, 270–76.

17. Lord Huan 13.1 (699 BCE).

18. The best fictional treatment of the Qin dynasty is Jean Levi, The Chinese Emper­or, translated from French by Barbara Bray (New York: Harcourt, 1987). Two of China’s best-known directors have recently made films about this period of time: “The Emperor and the Assassin,” directed by Chen Kaige and “Hero,” directed by Zhang Yimou. I should note here that Mao Zedong often identified himself with the First Qin Emperor in something of a reversal of the way the latter had been traditionally viewed. One of the most active movements of the years of the Cultural Revolution was entitled “Pi Kong, yang Qin,” which means “Criticize Confucius and praise the Qin.”

19. See A. F. P. Hulsewé, Remnants of Ch’in Law (Leiden: E. J. Brill, 1985), and Robin Yates, “Some Notes on Ch’in Law,” Early China 11–12 (1985–1987): 243–75.


Bottéro, Jean, Clarisse Herrenschmidt, and Jean-Pierre Vernant. Ancestor of the West: Writing, Reasoning, and Religion in Mesopotamia, Elam, and Greece. Trans. by Teresa Lavendar Fagan. Chicago and London. The University of Chicago Press. 2000.

Brooks, Bruce E. and A. Taeko Brooks. The Original Analects: Sayings of Confu­cius and his Successors. New York: Columbia University Press. 1998.

Collingwood, R. G. The Idea of History. Oxford: Oxford University Press. 1946.

Falkenhausen, Lothar von. “Issues in Western Zhou Studies: A Review Article.” Early China 18. 1993:145–71.

Febvre, Lucien. Combats pour l’histoire. Paris: Agora. 1953.

Herodotus. The Histories. Trans. by Robin Waterfield. Oxford: Oxford University Press. 1998.

Hsu, Cho-yun. Ancient China in Transition: An Analysis of Social Mobiity, 722–222 BC. Stanford: Stanford University Press. 1965.

Hulsewé, A. F. P. Remnants of Ch’in Law. Leiden: E. J. Brill. 1985.

Keightley, David. Sources of Shang History: The Oracle Bone Inscriptions of Bronze Age China. Berkeley: University of California. 1978.

Legge, James, trans. The Chinese Classics, Vol. 5, The Ch’un Ts’ew with The Tso Chuen. 1879. Reprint. Taipei: Southern Materials Center. 1985.

Lewis, Mark Edward. Writing and Authority in Early China. Albany: State Universi­ty of New York Press. 1999.

Marou, Henri-Irénée. De la connaissance historique. Paris: Editions du Seuil. 1954.

Pines, Yuri. Foundations of Confucian Thought: Intellectual Life in the Chunqiu Period, 722–453 BCE. Honolulu: University of Hawai’i Press. 2002.

Puett, Michael J. Cosmology, Sacrifice, and Self-Divinization in Early China. Cam­bridge and London: Harvard University Asia Center, 2002.

Schaberg, David. A Patterned Past: Form and Thought in Early Chinese Historiog­raphy. Cambridge and London: Harvard University Asia Center. 2001.

Shaughnessy, Edward L., ed. New Sources of Early Chinese History: An Introduc­tion to the Reading of Inscriptions and Manuscripts. Early China Special Monograph Series 3. Berkeley: University of California, 1997.

_____. Sources of Western Zhou History: Inscribed Bronze Vessels. Berkeley: Uni­versity of California. 1991.

Yates, Robin. “Some Notes on Ch’in Law.” Early China 11–12. 1985–1987:243–75.