(I dedicate this book to the memory of Wilbur Knorr, 1945-1997.)
The date of the Zhou Conquest of Shang is the most famous problem in ancient Chinese chronology. I will argue that the correct date was probably still known in the time of Confucius; but it was lost soon thereafter, and the matter has been debated ever since. In the past 25 centuries more than 40 dates have been proposed, and most of these still have their defenders. The problem cannot be solved by itself. Any proposed solution must be consistent with everything else that can be known about ancient Chinese history before and after. For this reason I have found that I could approach the problem only as part of a complete study of chronology from the earliest times to the end of Western Zhou. This has taken me almost two decades. The result is my book The Riddle of the Bamboo Annals.
In earlier chapters, I use the Annals, together with astronomical data explored by Pankenier, and late Shang inscriptions of all kinds, to work out a complete chronology down to the beginning of Zhou. (I derive putative dates in an original Annals-like chronology even for Huang Di, Zhuan Xu and Di Ku, though I consider these rulers legendary.) Pankenier has shown that the conjunction of 1953 corresponds to the 14th year of Shun in the Annals, when real power passed to Yu. Kevin Pang and I have shown that if one uses this date with Annals reign lengths, positing two-year gaps (from succession to accession, for completion of mourning) between reigns, then the Annals' date for the solar eclipse on the 1st of the 9th month of the 5th year of Zhong Kang must be 16 October 1876, when there was in fact a ring-form eclipse visible near the Xia capital.
I then extend the analysis backward and forward. I find 2287 to have been the originally intended date for Huang Di 1, confirmable in several ways. And I find the last year of the next to last Xia ruler Fa to be 1555, which has been shown by Pankenier to have been the actual last year of Xia; so I conclude that Di Gui (Jie) is an imaginary addition to Chinese history (and this too I confirm in several ways).
These derived dates for Xia are confirmed, next, by looking at Kong Jia, the 14th king, notable because he has a "gan" name. The obvious guess to try is that the gan is determined by the first day of the reign, which I had found to be 1577-69. The first day of 1577 in the Xia calendar was jiazi, first day in the 60-day cycle. The hypothesis seems to work well for Shang kings, who all have gan names. The rules appear to have been these: (1) no two successive reigns have the same gan; (2) if the succession year would cause this to happen, the gan is determined by the accession year (either two or three years later, in Shang; apparently a Shang king's death year was counted in his calendar only if he lived through most of it); and (3) gui is taboo; if the rules dictate gui, then the next day, jia, is used instead. These rules, together with Annals reign lengths (with a few corrections: Tai Wu's "75 years" is not possible) yield a complete tentative chronology for Shang.
It follows that royal mourning was as important in Shang as it was in Xia and in Zhou. The burdens of mourning on the successor king almost led to a usurpation at the beginning of Shang. To ward off the danger, it seems that a Shang king regularly coopted a royal brother to succeed himself while his own son and heir performed mourning. But eventually this led to the danger of usurpation of the succession by the brother, forcing a king to establish his chosen son before his own death. Late in the dynasty the son's formal reign-calendar sometimes began even before the father's death. Thus Wu Yi, who died in 1109, gave his son Wenwu Ding a calendar beginning in 1118; and Di Xin (1086-1041) began a second calendar, probably for his son Wu Geng, in 1068. These facts I discovered by analyzing bone inscriptions containing data about the annual ritual cycle (si) of the "5 sacrifices," using the work of Shima Kunio and Chang Yuzhi. This has allowed me to work out, tentatively, a complete schedule of exactly dated first days of the annual ritual cycle from 1120 to 1041 (Appendix III). This material, together with my revised dating of 56 Zhou bronze inscriptions (Appendix II) provides a firm basis for chronology for the last 350 years of the classic "Three Dynasties."
In chapters at the end of the book I use the Annals together with Western Zhou bronze inscriptions to establish a complete chronology for Western Zhou (Chapter Nine); and in the end chapters I try to show the probable steps by which, from the late 5th century to about 300, an originally correct chronology in chronicles ancestral to the Annals got distorted into the chronology given in the so-called "modern text."
---------------------
Here follows Chapter Eight, together with (1)
my complete reconstruction of the Bamboo Annals' chronology from Huang
Di to the end of Western Zhou; (2) my tentative dating of 56 Western Zhou
bronze inscriptions, included in my book The Riddle of the Bamboo Annals
(as Appendix II). This supersedes and corrects my dating of 68 Western
Zhou bronze inscriptions included in my 1981 paper "Xi Zhou zhi Nianli."
That paper was presented at the fourth conference of the Guwenzi Yanjiuhui
in Taiyuan, and is included in the present book Wu Wang ke Shang zhi nian
yanjiu. My article in HJAS 43 (1983) was a revision and expansion
of the Taiyuan conference paper.
Chapter Eight: The Date of the Zhou Conquest of Shang
Part I
8.0 In Chapter Four, I attempted to fix the chronology of
the Shang kings from Wu Ding's death (in 1189, I believe) to the end of
the dynasty 150 years later. Fraternal succession disappeared, as
kings became increasingly anxious about fraternal usurpation. As
kings tried to set up their heirs before their own deaths, chronology becomes
tangled. Although the Wenwu Ding calendar begins in 1118, Wu Yi actually
died in 1109, on a royal hunt "in the He-Wei area." This hunt, conventionally
capping a successful campaign (against the "Lord of the Yu Fang") must
in part have been intended as a warning to the rising Zhou power.
The warlike Zhou chief Ji Li (1127-1102) had been winning victory after
victory. Finally, when he came to court (probably in 1102) the Shang
king (who was probably Wenwu Ding renamed Di Yi) threw him in prison, where
he soon died. After that peace between the two powers was never more
than pretense. The "Duo Yi" chapter of Yi Zhou shu has Wu Wang after
his victory (in 1040) saying to Zhou Gong that "Heaven's withholding its
favor from Yin has been going on for sixty years, since before I was born";
i.e., from Ji Li's death on, Zhou had not regarded Shang's rule as sanctioned
by Heaven.
In 1086 Di Xin became king of Shang.
We read in the Annals that eleven years before the conjunction (of 1059),
hence in 1070, Wen Wang held a conclave of lords in his own court, an obvious
threatening move. The Shang king (says the Annals) staged a hunt
in the Wei valley the following year, clearly a response to the threat,
for this was the Zhou homeland. The next year was 1068, when Di Xin's
second calendar begins, and there must have been at that time a great assembly
of lords in the Shang caapital, from which Wen Wang would not dare to stay
away. In any case it is stated in the Annals (which is quite silent
about a calendar or an assembly) that in this year (that is, two years
after the assembly in Zhou) Wen Wang was imprisoned in You Li, where he
remained for the next seven years. (This is where he must have been,
then, when he addressed his sons (by letter?) about "planning for the succession"
on the occasion of the lunar eclipse of March, 1065, recorded in Yi Zhou
shu #23 "Xiao Kai.") On Wen Wang's release in 1062 he found increasing
support among the other lords. In 1058 Di Xin is said to have recognized
him as having the authority "to conduct punitive expeditions of his own,"
seen by the Zhou as the "mandate," an authority Wen Wang was certain to
exercise anyway. In 1056 he declared himself king, with a new calendar.
In 1055 he held another convocation of regional lords. It would seem
(see 6.10) that Di Xin ignored the threat, choosing soon after to march
east again against the Eastern Yi, perennial enemies (there had been a
great campaign against them earlier, in 1077-76). Wu Wang succeeded
in 1049, and we continue to argue how long it took him to mount the final
campaign.
8.0.1 Summary of the argument: I have published three
articles concerning the problem of the date of the Conquest, in 1983, 1984,
and 1985. Each is partly right and partly wrong. I explain
this matter first, in 8.1. In 8.2, I set out the basic assumptions
I am using. In 8.3, I present the analysis of the Annals that leads
me to the date 1040. In 8.4, I offer twelve arguments of confirmation.
In 8.5, I tabulate the main events, with dates, during the century concluding
with Zhou Gong's surrendering full royal power to Cheng Wang, effective
1030.
8.1. Brief critique of relevant published articles by DSN:
1983: Harvard Journal of Asiatic Studies 43, 481-580: "The Dates of Western Chou"
Right: (1) Each Western Zhou king had two calendars,
a calendar counting from his succession year; and a calendar counting from
the year of completion of mourning, i.e., two years later, the accession
year. The Bamboo Annals normally uses the latter. (But as Shaughnessy
(1991 p. 155 n. 60) has noticed, the succession calendar was used early
in a reign, and the accession calendar only later. For example, in
Xuan Wang's reign, the succession date 827 governs the calendar through
810, and dates are counted from 825 only after completion of mourning for
Gong He in 809. See Appendix II.)
(2) Wen Wang named 1058 his "mandate" year, and proclaimed
himself universal king with a royal "Zhou" calendar beginning 1056.
The reason for this two-year delay (explained in Nivison 1983, pp. 530-531)
was probably to allow all mourning obligations among the people to expire
before they were required to recognise the new government by using the
new calendar.
(3) Wen Wang died in 1050, at the end of a reign of 2
+ 50 years. The Annals dates for Wen Wang are 1113-1062, displaced
backward twelve years (together with other pre-Conquest Zhou dates) by
redating of the conjunction of 1059 to 1071.
(4) Lunar phase terms in Western Zhou dates should be
interpreted approximately as did Wang Guowei: They name lunar quarters,
or the first days of quarters. (For the full proof, see Chapter Seven.
In brief: the prefatory statements in the "Kang Gao" chapter of the Shang
shu, long recognized as out of place, is actually an alternative preface
to the "Shao Gao" chapter. A comparison of the two texts shows that
the term zaishengbo must name a day that is either the 6th or 7th of the
month; therefore jishengbo, which should be the next day, must be the first
day of the second quarter. Examination of Xuan Wang inscriptions
then shows that Wang's schema is right.
Wrong: The date of the Conquest was not 1045, it was 1040. I had misread the Shiji: The shijia chapters say that the Conquest was in Wu Wang's 11th year; the "Zhou Benji" says that the Zhou army had crossed the Yellow River by the 12th month of the 11th year, and that the victory was won on jiazi in the 2nd month. The historian was actually following standard Western Han practice, taking "12th month" to be the name of the post-winter-solstice month, as it is in the Xia calendar. Thus he understood "12th month" as the name of the first month of the Shang calendar, so for him "2nd month" was two months later in the same year. In doing this he was misreading his own source (which as I explain below was in error anyway); so the Shiji is worthless as evidence for a "12th year" date. (I also gave incorrect dates for Gong Wang, Yih Wang and Xiao Wang; and I gave an incorrect argument for the (correct) date 1105 for Di Yi.)
1984: Early China 8 (1982-1983), 76-78: "1040 as the Date of the Chou Conquest"
Right: 1040 is the correct date for the Conquest.
Wrong: I argued for this by supposing that Wen Wang's reign was just 50 years, 1101-1052, so that 1040 was Wu Wang's 12th year. Actually, Wen Wang did die in 1050, and (as I realized later) the Conquest was not in a "12th year." Seeing that this argument was unsatisfactory, I returned to 1045 as Conquest date in my publications from 1985 through 1989. I had resolved the problem by 1990 (see below), by discovering how to explain "12th year" as an ancient error, and by discovering an unnoticed feature of the Zhou calendar, namely, that it took as winter solstice day a day that was two (or three) days late.
1985: Guwenzi Yanjiu 12, October, 445-461: "Guoyu 'Wu Wang fa Yin' tian xiang bian wei"
Right: The astrological description of the sky at the time of the launching of the Conquest campaign, as given in the Guoyu, "Zhou Yu" 3, is not a true report but an incorrect calculation made centuries later.
Wrong: I argued that this calculation was inserted into the Guoyu in the 1st century BCE, and was presumably made at this time. (Hence Sima Qian does not use it, and Liu Xin does.) Furthermore, I was again assuming that the Conquest was in 1045. Therefore my reconstruction of the incorrect calculation was completely wrong. I now know that the calculation was based on observations in the early or middle 5th century BCE, and must have been made by a historian who knew that the date was 1040, 2nd month, jiazi; but in analyzing this date he used bad science: (1) he accepted the "Yin Li" 76-year intercalation cycle (the zhang-bu system) as completely accurate; and (2) he believed that Jupiter circled the zodiac in exactly 12 years. Actually, the zhang-bu system gives a ganzhi day date a day early for each three centuries back, and the Jupiter cycle is about 11.86 years. Therefore (1) he reasoned that a Xia-calendar jiazi victory day would have been the 1st of the 3rd month. (See 8.4.11.1 (b) below.) He was thus led to believe that the intended calendar must be the Zhou calendar, not the Xia calendar, so that the campaign must have started late in 1041, i.e., not in the winter solstice month (as it in fact did), but 30 + 29 days earlier; and he calculated (2) that in that year Jupiter must have been in Chun Huo (Quail Fire), as it would seem to be if one counted by twelves back from any observed Quail Fire year in the first half of the 5th century. (More below, 8.3.1; see also 11.7 in Chapter Eleven.)
8.2. Basic premises:
(a) Reconstruction of chronology must use, or explain, the dates in the "modern text" Bamboo Annals. Shaughnessy's discovery (1986) that a slip was moved from the Cheng Wang chronicle to the end of the Wu Wang chronicle shows that the text is genuine. (The slip was moved (about 425 BC) long before the text was buried in ca. 300 BCE. The Zuo zhuan assumes that events mentioned in the slip occurred in Wu Wang's reign; and the "Jin Teng" chapter of the Shang shu assumes that Wu Wang lived longer. See Chapter Eleven.))
(b) In the Bamboo Annals, the 5-planet conjunction put in Di Xin 32 = 1071, and said to be in Fang, must be read as referring to the conjunction in Jing in 1059 BCE, as suggested by Needham (1959 p. 408 n. c) and confirmed by Pankenier (EC 7 p. 4). (And the erroneous date and location must be explained and due allowance made for the error; i.e., one must not simply dismiss the text because it is in error.) The Conquest cannot be earlier than 9 + 5 years later (actually it was 9 + 10 years later), because Wen Wang died nine years later.
(c) The month and day dates in the "Wu Cheng" chapter of the Shang shu, as quoted by Liu Xin, are valid (but misinterpreted by Liu): The only years in the possible time range that satisfy these data are 1045 and 1040 (as I explain in detail in Chapter Seven).
8.3. Basic argument for 1040: The calendars for Cheng Wang through Mu Wang in the Bamboo Annals, when Mu Wang's first year is corrected (by using bronze inscriptions) from 962 to 956, and when allowance is made for 2-year mourning intervals, show that Cheng Wang's succession year must be 1037 BCE (and that the first year of Wen Wang's royal Zhou calendar must have been 1056 BCE, 100 years before Mu Wang 1). An editor who knew Cheng Wang's first year, but did not admit mourning intervals, would have dated Kang Wang two years early, to 1007, which is the Annals' date, instead of 1005 (succession) and 1003 (accession). The Xiao Yu ding inscription, 25th year (see Appendix II), shows that 1003 is correct as accession date. Therefore the editor was correct in believing 1037 to be Cheng Wang's "first year." This editor, however, dates the Regency 1044-1038, making Wu Wang die in 1045; and this is impossible. So putting the Regency before the beginning of Cheng Wang's reign was a mistake. The Regency must have begun in 1037, which must be Cheng Wang's succession year. If Wu Wang died two years after the victory, then the Conquest date must be 1040.
8.3.1 Objection: The Lu shi Chunqiu and the Bamboo Annals say that the victory was in Wu Wang's 12th year (and the Shiji seems to be misusing a source that said the same). This implies that the date must be 1045.
Reply: This is what I supposed in HJAS 43; But "12th year" is an error. A calculation (as above) made ca. 475-450 BC would have concluded that the campaign started in late 1041, with Jupiter in Chun Huo. If the calculator knew that Jupiter was in Chun Huo also in the Mandate year (the year after the conjunction), and that Wen Wang died in the 9th year of the Mandate, he would have to suppose that the Mandate year was 2 x 12 years before 1041, i.e., 1065, so he would suppose that the 1056 calendar was Wu Wang's. His received information must have been that Wu Wang's death was in the 12th year (sc. of his own calendar) and the Conquest was in the 17th year (sc. of the royal calendar). Not knowing that two different calendars were involved, he would suppose that the two dates had gotten confused, and would try to correct them. This could be done in two steps: (1) One might suppose that Zhou Gong's 7-year Regency was the seven years prior to Cheng Wang's accession year (1042-1035), rather than the seven years beginning with Cheng Wang's succession year (1037-1031, the correct dating). This would put the Conquest in the 12th year, i.e., 1045, if Wu died two years later. (This mistake probably had already been made by others.) (2) Wu Wang's death could then be moved from his 14th year to his 17th year, by the slip move that Shaughnessy discovered. This was probably the motive for the move, which other evidence shows must have been made in Warring States (rather than in the Jin Dynasty); e.g., the Zuo chuan and the "Jin Teng" chapter of the Shang shu both agree with this altered chronology. (A later Warring States "correction" in the Annals moved these dates back another five years. See Chapter Twelve.)