In the Yin Oracle-Bone inscriptions, there are records of five lunar eclipses, all of which are Bin 賓 diviner group inscriptions.Footnote 1 For the past seventy years, scholars in the fields of Oracle-Bone inscription studies and the history of Chinese astronomy have conducted research on those inscriptional records, hoping not only to find correct absolute dates for those eclipses,Footnote 2 but also to reconstruct chronologies of the Yin (i.e. Shang) and Western Zhou dynasties.
One of those five lunar eclipses is carved on the reverse of Yingcang 英藏 885 and 886.Footnote 3 Unlike the other four lunar eclipses, the ganzhi 干支 date of this eclipse has been controversial.Footnote 4 To calculate the absolute date of this eclipse, it is necessary to correctly ascertain its ganzhi date. Therefore, in this article I will first test different opinions on the ganzhi date of that eclipse against the paleographic evidence to examine whether those opinions fit the pattern of inscriptional records of lunar eclipses. Based on these analyses, I will draw a conclusion on the ganzhi date of this eclipse. Then I will use an innovative approach to find the absolute date of this eclipse.
1. Pattern of Inscriptional Records of Lunar Eclipses
Below are inscriptions directly related to those five lunar eclipses.
Eclipse I:
[1] 旬壬申夕月有食。 《合集》11482（反）
On the night of the tenth day renshen (ganzhi day 9), the moon was eclipsed. Heji 11482 (reverse)Footnote 5
Eclipse II:
[2] 癸未卜, 爭貞: 翌甲申易日。之夕月有食。甲霧不雨。 《合集》11483
Cracking on guiwei (day 20), Zheng tested: “On next day jiashen (day 21) it will be sunnyFootnote 6.” This night, the moon was eclipsed. On day jia(shen), it was foggy and it did not rain. Heji 11483
Eclipse III:
[3] 三日乙酉夕月有食, 昏。八月。 《合集》11485
On the night of the third day yiyou (day 22), the moon was eclipsed and it was dark.Footnote 7 (It was in the) eighth month. Heji 11485
Eclipse IV:
[4] 六日甲午夕月有食。 《合集》11484
On the night of the sixth day jiawu (day 31), the moon was eclipsed. Heji 11484
Eclipse V:
[5] … 日己未夕
庚申月有食。Footnote 8 《英藏》885反On the night of … day jiwei (day 56)
gengshen (day 57), the moon was eclipsed. (Or: On the night of … day jiwei, it . On gengshen, the moon was eclipsed.) Yingcang 885 (reverse)[6] 七日己未
庚申月有食。 《英藏》886反The seventh day jiwei (day 56)
gengshen (day 57) the moon was eclipsed. (Or: on the seventh day jiwei, it . On gengshen, the moon was eclipsed.) Yingcang 886 (reverse)
Chen Mengjia 陳夢家, Chang Yuzhi 常玉芝, and Feng Shi 馮時Footnote 9 have all observed the fact that records of Eclipse I–IV clearly state on which night those four eclipses happened. Since lunar eclipses normally occur at night, one may assume that one ganzhi date only should be enough to record the date of a lunar eclipse. In other words, it seems unnecessary for the Yin to inscribe the word xi 夕 (night) in those records. Nevertheless, this was done not just in one record, but in all four. Therefore, it is reasonable to conclude that scribes attached importance to the times of eclipses, and it is a pattern in inscriptional records of lunar eclipses that on which night a lunar eclipse occurred should be clear.
In inscriptions [5] and [6], records of Eclipse V, the word xi does not occur immediately before the expression yue youshi 月有食. Chen Mengjia, Chang Yuzhi, and Feng Shi have noticed such a difference between records of Eclipse V and those of the other four eclipses. Chen Mengjia and Feng Shi have also made efforts to explain why the word xi is missing after gengshen in these two inscriptions.
1.1 Chen Mengjia's explanation
Chen Mengjia regarded gengshen as the ganzhi date for Eclipse V.Footnote 10 In comparing these records, he noticed the difference between records of this lunar eclipse and those of the other lunar eclipses, as stated above. In order to explain this apparent discrepancy, he proposed that the word xi refers to the second half of a night in the Oracle-Bone inscriptions. He thought that Eclipse V started in the first half of the night of November 15, 1218 b.c.e. As a result, he did not think the Yin should use the word xi in records of Eclipse V.
Chen Mengjia's explanation appears to be based upon two kinds of unreliable evidence. The first was the absolute dates he selected for Eclipse I, IV, and V. According to his reconstructions, Eclipse I and IV began after midnight, whereas Eclipse V began before midnight. Hence, he thought that the word xi in the Yin Oracle-Bone inscriptions might refer to the second half of a night. It is unfortunate that there is no consensus on absolute dates for those Yin lunar eclipses. He himself even said that he had guessed (caice 猜測) those dates. This kind of evidence clearly does not lend any credibility to his explanation.
The second piece of evidence he cited was an assertion by Zheng Xuan 鄭玄 of the Western Han Dynasty. Zheng Xuan wrote in a note that the time before chen 晨, “dawn,” was called xi. However, neither Zheng Xuan nor Chen Mengjia provided any evidence to support that assertion, which alone cannot support Chen's interpretation of the word xi in the Oracle-Bone inscriptions.
Due to the problems with his evidence, it was not credible for Chen Mengjia to say that from midnight to the time before dawn was called xi in the Oracle-Bone inscriptions. As far as I know, no scholars accept his conclusion that the reason for the lack of the word xi in records of Eclipse V is because that eclipse started before midnight.
1.2 Feng Shi's explanation
Feng Shi also holds the view that the ganzhi date for Eclipse V is gengshen. Below is his explanation in 2001 as to why the word xi does not appear right before yue youshi in inscriptions [5] and [6]:Footnote 11
庚申月食緣何不作“庚申夕”，確實非常奇怪。認為該辭省略了 “夕” 字顯然十分牽強。然則如從卜辭于月食發生的前一日夜晚記有 “壹 (曀)”字
(Feng Shi's transcription of graph 
)
的情況考慮，認為前夜的陰沉天氣一直持續到第二天夜晚，致使殷人只能在陰晴間偶爾觀測到月食，而並不知其準確的交食時間，故未記 “夕”字。這種解釋至少不違背卜辭的本義。
Why was (the time of) the lunar eclipse on gengshen not recorded as “on the night of gengshen”? This indeed is extremely strange. It is obviously unusually farfetched to think that the character xi was omitted in this inscription. However, it is recorded in the inscription that the weather was cloudy on the night immediately before the occurrence of the lunar eclipse. If one takes that into consideration, one may think that the cloudy weather of the previous night continued all the way to the second night, which made the Yin able to observe that lunar eclipse occasionally between cloudy and clear weather, and they did not know the accurate time of the eclipse; so they did not record it with the character xi. Such an explanation is at least not contrary to the original meaning of the inscription.
Obviously, Feng Shi thinks that the omission of the word xi in his version of records of Eclipse V is so strange that he has to come up with an explanation for it. He also states that it would not do simply to say that the character xi was omitted.
So the question is whether his explanation works. It is his view that the Yin did not use the word xi because they could not know the accurate time of that eclipse, since they could only occasionally observe the eclipse due to cloudy weather that had started the previous night. Such an explanation is unacceptable for a simple reason: the Yin did not need to know the exact time of an eclipse to record that it happened at night. It is common sense that as long as they observed a lunar eclipse at night, even only occasionally and without knowing its exact time, they could record that it happened on that night. According to Feng Shi's scenario, the Yin indeed observed the eclipse occasionally without knowing its exact time. Theoretically, the Yin should have written: gengshen xi yue you shi 庚申夕月有食, rather than an inscription without the word xi. It appears that his explanation does not answer the question as to why the word xi does not occur after gengshen in records of Eclipse V.
1.3 Chang Yuzhi's explanation
Chang Yuzhi is aware of the apparent difference between records of Eclipse V and those of other eclipses. Her opinion is that jiwei xi gengshen in [5] and jiwei gengshen in [6] specify the time of Eclipse V: the inscription [5] clearly states that it happened on the night of jiwei and [6] implies the same.Footnote 12
It is apparent that there are different understandings about the ganzhi date for Eclipse V in inscriptions [5] and [6]. Chen Mengjia, Feng Shi, and other scholars think its ganzhi date is gengshen. If so, those records do not specifically state on which night Eclipse V occurred and its time is not clear; thus, they are not in agreement with the pattern of inscriptional records of lunar eclipses. Chang Yuzhi and others think both jiwei and gengshen are used to specify the time of Eclipse V and that Eclipse V started from the night of jiwei. If so, records of Eclipse V are in agreement with the pattern. At this point, it is necessary to address the issue of the ganzhi date of Eclipse V.
2. Ganzhi Date of Eclipse V
As shown above, for jiwei xi gengshen in [5] and jiwei gengshen in [6], there are two readings: one is to insert a syntactic break after graph , making gengshen the date of Eclipse V; the other is not to use such a break, the result of which is that both jiwei and gengshen are relevant to this eclipse. In the case of the first reading, records of Eclipse V will not be in agreement with the pattern of inscriptional records of lunar eclipses. This certainly sounds problematic. Moreover, it is questionable whether the graph has been deciphered correctly. Would the syntactic break have been required according to some decipherments? Furthermore, would the decipherments of those who do not favor the syntactic break be acceptable?
2.1 Arguments for a break after graph
2.1.1 Dong Zuobin's argument
In 1945, Dong Zuobin 董作賓 published his Yin Lipu 殷曆譜. He presented a systematic study of Yin lunar eclipses. However, he could not decipher graph , as shown by the following statement:
此或指天象言, 尚待考。Footnote 13
It might refer to a celestial phenomenon, and it remains to be deciphered.
Even though Dong Zuobin did not know what word the graph represented, he did not show any hesitation in picking the day gengshen as the only date of Eclipse V. Furthermore, he did not mention any possible relation between day jiwei and the eclipse, even though jiwei is part of the inscriptions of Eclipse V. His reason for doing so was that he thought the word represented by the graph he could not decipher might refer to a celestial phenomenon unrelated to the eclipse. Edward L. Shaughnessy states that “it is never acceptable methodology to prove one unknown with another unknown.”Footnote 14 It would seem that since Dong Zuobin was doing just that, his methodology, and therefore also his argument, are unacceptable.
2.1.2 Chen Mengjia's argument
Chen Mengjia made the following statements:Footnote 15
“夕”一定指晚上的氣候 ⋯⋯
The expression xi- must refer to some weather at night …
“夕”之義不外乎指夜間有星無雲或無星有雲。
The meaning of the expression xi- can only be that either there are stars but not clouds in the sky, or there are clouds but not stars in the sky.
Clearly, Chen Mengjia did not decipher the graph. In addition, the meaning of youxing wuyun 有星無雲, “there are stars but not clouds in the sky,” is exactly opposite to that of wuxing youyun 無星有雲, “there are clouds but not stars in the sky.” He did not even provide any evidence for relating graph to “clouds” or “stars.” It is obvious that he did not know which word was represented by the graph.
As in the case of Dong Zuobin, Chen Mengjia did not understand which word was represented by the graph; yet, he excluded the possibility of there being any relation between day jiwei and the eclipse, even though they appear so closely in inscriptions [5] and [6]. For the same reason as stated above, it is difficult to accept his interpretation.
2.1.3 Kwang-Chih Chang's argument
Kwang-Chih Chang 張光直 followed Dong Zuobin's later transcription of Eclipse V: graph was read as yin 垔 and there was a period after the word yin, making gengshen the date of Eclipse V.Footnote 16
Actually, it was Ye Yusen 葉玉森 who first proposed this yin reading of the graph.Footnote 17 To him, the bottom part of the graph depicts a vessel, while the top part depicts steam rising up. Therefore, he suspected that this graph represented the word yin, a kind of sacrifice. The way he deciphered the graph is normally called wang wen sheng yi 望文生義 (eliciting a meaning from the shape of the graph), which is often discouraged. Moreover, Li Xiaoding 李孝定 pointed out that the bronze graph for the word yin is 
, which is dramatically different from the bone graph, and that it is wrong to interpret graph as yin 垔.Footnote 18
Kwang-Chih Chang mentioned Long Yuchun's 龍宇純 paleographic study of this bone graph to support the reading of the graph as yin. However, Long Yuchun's opinion is that Ye Yusen's interpretation was incorrect.Footnote 19
Since graph cannot be transcribed as yin, Kwang-Chih Chang did not make a case for there being a syntactic break between this graph and gengshen.
2.1.4 Fan Yuzhou's argument
Fan Yuzhou 范毓周 cites two inscriptions in which graph is followed by animal names. Based upon those examples, the following is his conclusion:Footnote 20
從上述用例看，字顯然是與
、
大略相類的一種祭名，其在月食四刻辭中的用例中大概也應如是。
Judging from examples above, character obviously is the name of a sacrifice, which is more or less similar to ce and you . Perhaps, in the inscriptions of Lunar Eclipse IV, should also be used this way.
In the Yin Oracle-Bone inscriptions, that graph indeed may appear immediately before animal names. In such a context, it is acceptable to take it as a verb, meaning either “to use animals in a specific sacrifice” or “to deal with animals in a particular way.” It is incorrect for him to interpret it as a noun. What is more important is whether the inscriptions of Eclipse V provide the context just mentioned. The answer is that they do not because there are no animal names after the graph in inscriptions [5] and [6]. Moreover, among the 31 examples Zhang Bingquan 張秉權 collected of ganzhi date + + ganzhi date, no animal names appear in any of them.Footnote 21 Apparently, there are two contexts for graph . Within the context of there appearing no animal names after the graph, it cannot be considered as a sacrificial verb. Therefore, it is not credible to interpret the graph as the name of a sacrifice and to put a syntactic break to separate jiwei and Eclipse V.
2.1.5 Zhang Peiyu's argument
Zhang Peiyu 張培瑜 followed the view that there is a break between graph and gengshen for this reason: such a reading would make the absolute date for Eclipse V compatible with that of Eclipse III in terms of which month they occurred according to Yin calendar.Footnote 22 However, since 1999, he has followed Qiu Xigui's 裘錫圭 interpretation of the graph and has argued against a break between the graph and gengshen.Footnote 23 Moreover, in 1999, he selected absolutes for Eclipse V and III, regarding those dates as compatible.Footnote 24 Obviously, even Zhang Peiyu does not think his 1989 argument for a syntactic break after the graph and before gengshen is valid.
2.1.6 Feng Shi's argument
In 1990, Feng Shi transcribed graph as zhuo 斲 and put a comma after it without any further explanation.Footnote 25 Since then, Feng Shi has transcribed the graph differently and has discussed which word this graph represents. Below is his transcription of inscription [5]:Footnote 26
七日己未壹（曀），庚申月有食。
On the seventh day jiwei, it was cloudy; on gengshen, the moon was eclipsed.
我們認為，此字似應從饒宗頤先生釋壹，讀為曀，指天氣陰沉。
We think it seems that (we) should follow Mr Rao Zongyi to interpret this character as yi 壹 and to read it as yi 曀, meaning the weather is cloudy.
Feng Shi did not cite how Rao Zongyi 饒宗頤 deciphered graph as the word yi 壹. Actually, he only mentioned briefly in one sentence: 與金文壺字下體相同，當是 “壹” 之本字 (Graph is the same as the lower part of the bronze graph for character hu; it should be the original form of character yi.)Footnote 27 While the first half of the sentence is accurate, the second half is not at all certain. He did not show the evolution from the bottom part of the bronze graph for hu to character yi and did not provide examples to substantiate his claim. If that is already a weak interpretation of the graph, then it is an even more serious problem for him to read it as yi 曀, meaning “cloudy weather” without any evidence. This lends no credibility to putting a syntactic break after the graph.
2.1.7 Summary
Dong Zuobin, Chen Mengjia, Fan Yuzhou, Kwang-Chih Chang, Zhang Peiyu, and Feng Shi did not decipher graph so as to show a syntactic break between this graph and the ganzhi date gengshen in inscriptions [5] and [6] was warranted. Moreover, as discussed in section 1, there is a pattern of inscriptional records of lunar eclipses: the Yin attached importance to making clear in inscriptional records concerning which night a lunar eclipse happened. Putting such a syntactic break in records of Eclipse V would make them violate the pattern. For these two reasons, their arguments for the syntactic break are unacceptable.
Without any analyses of inscriptions [5] and [6], Zhao Quemin 趙卻民, Zhang Peiyu, Kwang-Chih Chang, Yan Yiping 嚴一萍, Wen Shaofeng 溫少峰 and Yuan Tingdong 袁庭棟, Chen Zungui 陳遵媯, Feng Shi, Lao Gan 勞幹, Xu Zhentao 徐振韜, Zheng Guang 鄭光, and Peng Diejun 彭瓞鈞, Qiu Jincheng 邱錦程, and Zhou Hongxiang 周鴻翔 all insert a syntactic break at that place in their transcriptions.Footnote 28 Nevertheless, they provide no evidence for establishing such a syntactic break.
2.2 Arguments for not putting a break after graph
Over the past sixty years, scholars have presented various arguments for interpreting “ganzhi date + + ganzhi date” as one structure in the language of the Yin Oracle-Bone inscriptions, thus attempting to establish a time period from a certain time in jiwei (day 56) to a certain time in gengshen (day 57) as the time of the occurrence of Eclipse V. By interpreting inscription [5] in this way, the eclipse would have started from the night of jiwei. Although inscription [6] does not specifically state the eclipse began from the night of jiwei, it makes it unmistakably clear. As a lunar eclipse cannot last longer than eight hours, if the eclipse happened at some time from jiwei to gengshen, it must have started at some point on the night of jiwei and ended sometime during the morning of gengshen. Thus, inscriptions [5] and [6] make clear on which night the eclipse happened; therefore, these two inscriptions do not break the pattern of the inscriptional lunar eclipse records. This is one advantage of this argument over those that favor a syntactic break between graph and day gengshen. However, the question remains whether that graph indeed represents a word that holds the “ganzhi date + + ganzhi date” together.
2.2.1 Homer H. Dubs's argument
It was Homer H. Dubs who first deemed that jiwei gengshen should be taken as an expression for the time of Eclipse V. To him, the eclipse on December 27–28, 1192 b.c.e. “must have been a striking event. It began at 8.48 p.m. local time at Anyang on cyclical day 56 and became total at 9.53 p.m. Totality lasted for an hour and three-quarters. The moon began to reappear at 11.37 p.m., and the eclipse ended at 42 minutes after midnight, on the morning of cyclical day 57. The debated unknown character in the record then probably means something like ‘midnight’ or ‘continuing into’, indicating that this eclipse endured into the morning of day 57.”Footnote 29
Apparently, Homer H. Dubs did not decipher the unknown graph. He simply thought that such an interpretation would fit the circumstances of the lunar eclipse on December 27–28, 1192 b.c.e., and he might well be wrong in matching this eclipse with Eclipse V. In addition, he chose this eclipse because he held the opinion that a Yin day started at midnight, which is not supported by any inscriptional evidence. If a Yin day did not start at midnight, but rather at su 夙, which is supported by evidence in both Oracle-Bone inscriptions and later transmitted texts,Footnote 30 then this interpretation of the graph would not be possible.
Indeed, Homer H. Dubs proposed a new way of interpreting graph . Nevertheless, he did not substantiate his claim.
2.2.2 Chou Fa-kao's argument
Concerning the question of when is the start of a Yin day, Chou Fa-kao 周法高 was in agreement with Homer H. Dubs. Moreover, he followed Ye Yusen's opinion on the unknown graph :
As for the decipherment of the character, I think that Yeh Yu-sen reads it correctly as 禋 …
In my opinion, when this character is used to mean a sacrifice, it is the same as 禋; when it is used between the notations for two cyclical days, it is a phonetic loan for the character 因, of which, Karlgren says (370 a.): … yin … continue (Lunyu).Footnote 31
Chou Fa-kao followed Ye Yusen's interpretation of that graph, despite the fact that Li Xiaoding had already demonstrated that Ye Yusen was mistaken, as cited in section 2.1.3 above. Hence, it served no purpose for Chou Fa-kao to propose a phonetic loan word yin for this graph.
Moreover, Chou Fa-kao misunderstood the meaning of the word yin 因 used in the Lunyu. Its original context is as follows:
“殷因于夏禮，所損益可知也；周因于殷禮，所損益可知也。” If the word yin is translated as “to continue,” then the translation of this sentence could be: “The Yin continued etiquettes of the Xia, and what parts of the Xia etiquettes were abandoned by the Yin and what was added to the Xia etiquettes by the Yin can be known; the Zhou continued etiquettes of the Yin, and what parts of the Yin etiquettes were abandoned by the Zhou and what was added to the Yin etiquettes by the Zhou can be known.”
In the context above, the word yin means that the subject of that sentence “continued” something that existed before; it does not mean that the subject itself continues. Obviously, Chou Fa-kao intended to use the latter and non-existent meaning of the word yin to explain the structure jiwei gengshen. Such a rendering of the word yin is unacceptable.
2.2.3 David N. Keightley's argument
David N. Keightley discussed the graph in jiwei gengshen in 1978, 1982, and 2000.Footnote 32 In his 1982 article, he enunciated his interpretation of the records of Eclipse V, a relevant part of which I cite below:
I believe that the word 
(Grammata Serica Recensa, no. 483b) that Chang, following Tung Tso-pin, reads as yin (it figures in his eclipse inscription no. 5) was not, in this context, a sacrificial term, but a word, *d'ŭg/t'ou(– > *tŭk/cho; 
, 斲 (Grammata Serica Recensa, no. 1235a,b), meaning “cleave” (as, in other contexts, “to cleave sacrificial meat”); it referred to the no-man's-land period in the dark hours of night when one kan-chih day cleaved to another, and dreams, toothache attacks, childbirths, and lunar eclipses were all recorded as occurring … Chang's translation of his eclipse no. 5—“Seven days later, Chi-Wei (we performed the) yin (ceremony). Keng-Shen, the moon eclipsed” (p. 324)—requires that the eclipse has occurred on day 57 (keng-shen) in the 60-day cycle, rather than day 56. My translation … —“On the seventh day (after the original divination and prognostication of harm), in the t'ou period when chi-wei (day 56) cleaved to keng-shen (day 57), the moon was eclipsed”—has the eclipse straddling two days.
I assume that David Keightley has followed Yu Xingwu 于省吾 to read graph as .Footnote 33 Yu Xingwu pointed out that the bottom part of bronze graphs for hu 壺 such as 
, 
, and 
clearly is the bone graph and the top part of those bronze graphs is the lid of a hu. Therefore, the bone graph is a vessel hu without its lid, which later became . In Xu Shen's 許慎 Shuowen Jiezi 說文解字, Yu discovered that is a variant form of 
, which refers to a kind of wine vessel, pronounced dou. Thus, it is acceptable to read the bone graph as dou.
In the structure jiwei gengshen, however, that bone graph could not be a wine vessel. Keightley read it as dou, meaning “to cleave.” The word “cleave” has different meanings. According to Keightley, the structure jiwei gengshen “referred to the no-man's-land period in the dark hours of night when one kan-chih day cleaved to another”. He has also made it clear that in records of Eclipse V this period of time straddled two days.
In sum, the interpretation of bone graph as dou is supported by paleographic evidence. Furthermore, Keightley's translation has demonstrated that his reading is appropriate in the context of Eclipse V. Therefore, according to Keightley's argument, there is no break between graph and gengshen.
2.2.4 Ken-ichi Takashima's argument
Ken-ichi Takashima first expressed his interpretation of bone graph in this context in 1979.Footnote 34 Recently, he has evaluated some other interpretations and presented his own:Footnote 35
I would, therefore, maintain the general outline of what I suggested (1979: 54): the graph should be transcribed as dou (though Shuowen [14a/4a] says “liquor jar”), which later acquired the semantic determinative 斤 “axe (=cut)”, yielding the word zhuo 斲 “to cleave, cut”. Its use in the sacrificial context is related to its temporal use, the common ground being the meaning of zhuo as “to cut, cleave” …
To sum up, we would like to suggest that the formula NP1 + 斲 + NP2 meant literally “(when) NP1 ‘cleaved/cut’ NP2” … In Takashima (1979: 54), I characterized the verb zhuo as “punctual,” rather than “durative” … In everyday usage, however, the time reference could have been looser, indicating, for instance, the approximate period when the king was asleep as in the above inscription. The phrase “之夕斲甲子” could, therefore, be rendered “(about the time) that evening passes over and on to the jiazi day.” This is tantamount to “during the transitional period from NP1 to NP2” or “while NP1 was ‘cleaving’ NP2.”
Ken-ichi Takashima offers linguistic criticism as to why other decipherments of the bone graph are weak, while Yu Xingwu's decipherment is sound from the point of view of paleography. Even though he reads the graph as the word zhuo 斲, his translation “during the transitional period from NP1 to NP2” works within the context of inscriptions [5] and [6].
2.2.5 Qiu Xigui's decipherment of graph
In 1993, Qiu Xigui presented a paper in which he read the bone graph as the character min 皿.Footnote 36 Since then, his interpretation of this graph has been followed by almost all scholars in the field of Oracle-Bone inscriptions in China. Chang Yuzhi, Zhang Peiyu, Xia-Shang-Zhou duandai gongcheng zhuanjiazu 夏商周斷代工程專家組, and Li Yong 李勇 are obvious proponents.Footnote 37 However, there are some problems with Qiu Xigui's decipherment.
In the Yin Oracle-Bone inscriptions, bone graphs for min 皿 are written as 
, which is clearly quite different from . However, why then does Qiu Xigui think they represent the same character? One of his two main reasons is that the ways they are used are the same (用法相合). He cites examples to make this point. For instance, the expression min wuniu 皿五牛 appears in Yingcang 1977. As in the case of , min is used before sacrificial animals. In Heji 31151, there is the following expression: hui geng min xin jiu 惠庚皿辛酒. This is an example in which min is used between two ganzhi dates, yet another way in which is used.
It is true that both graph and min appear to be used in the same contexts. But that does not prove that these two different graphs represent the same word. For example, in English there are expressions such as “Monday and Tuesday” and “Monday or Tuesday,” in which “and” and “or” are used within the same context. However, it would be incorrect to say “and” and “or” are the same word based upon those two expressions. In the same way, it is weak to say that and refer to the same word only because they occur in the same inscriptional contexts.
Qiu Xigui's other main reason is that the shape of and that of are related. He also cites two examples to demonstrate this point. The first example is an inscription on Heji 32181. In this inscription, there is the expression meng sanyang 盟(
）三羊. To him, the graph for meng consists of two parts: 
and 
; these two parts share 
. To Qiu Xigui, and are obviously the same character.
There is a problem with this example. The middle of is , whereas the middle of is rectangular. Obviously, and are not the same graph.
Qiu Xigui's second example is the expression 
羊 on Tie 67.2. He thinks that the graph before yang is a variant form of either or min. Therefore, and min are related in terms of their shape.
The logic behind Qiu Xigui's explanation of this example sounds strange. The shape of and that of min are indeed somewhat similar. This could be confusing for some scribes. If a graph is inscribed that bears a resemblance to both and min, that may mean it is a mistake. It does not mean that and min are related in terms of shape.
The above two examples cited by Qiu Xigui do not fully support his opinion that and min are related in terms of shape. As a result, his second reason for reading graph as min is invalid. His first reason, as analyzed above, is weak. Therefore, it appears that Qiu Xigui's case is unsubstantiated.
In addition, there is a problem with the meaning Qiu Xigui has assigned to min. His opinion is that min in this context should be read as xiang 向, meaning “head towards.” Jiazi xiang yichou refers to the time when jiazi is going to be over and yichou is going to start, which should not be long.Footnote 38 Chang Yuzhi has further stated that ganzhi xi min ganzhi does not refer to a time period between those two days and does not straddle those two days; it only refers to the time period when the night of the previous date is going to be over before dawn and does not include the dawn of the second day.Footnote 39
According to Qiu Xigui and Chang Yuzhi, ganzhi xi min ganzhi refers to a short time period of the night of the first date just before dawn of the second date. However, a lunar eclipse often lasts several hours. These two time periods do not match, which is a problem for reading graph as min.
2.2.6 Summary
Among those scholars who do not favor a syntactic break in the expression jiwei gengshen, there are four interpretations of the graph: yin 因, dou , zhuo 斲, and min 皿. For yin, there are two problems: one, paleographic evidence shows that it is incorrect to interpret this graph as yin. See section 2.1.3; and two, the meaning “to continue” of the word yin does not fit within the context here. For min, there also are two problems: it has not been totally proven that the graph represents the character min; and the time referred to by ganzhi xi min ganzhi appears to be too short for a lunar eclipse.
Within the context of Eclipse V, it is appropriate to read the graph as dou or zhuo meaning “to cleave.” There are two reasons that “to cleave” does not literally mean “to cut.” First, day jiwei is not “a knife,” and therefore cannot “cut” day gengshen. Second, day jiwei is inanimate, and therefore cannot “cut” day gengshen with a knife. However, if “to cleave” is understood as “to penetrate,” a meaning seen in http://dictionary.reference.com/browse/cleave, then jiwei dou/zhuo gengshen makes sense. It refers to a time period that straddles day jiwei and day gengshen.Footnote 40
The following information was already cited by Yu Xingwu: 
is the so-called suwen 俗文 (popular form) of zhuo 斲. There is a sentence in the Lushi chunqiu 呂氏春秋: “大匠不斲，大庖不豆 (a great craftsman does not do cleaving, and a great cook does not do cleaving.) Yu Yue 俞樾 thought that dou 豆 is a phonetic loan word for dou 剅 meaning to “cut/cleave.” According to the Jiyun 集韻, 剅 and are the same word.Footnote 41 Therefore, /剅 is pronounced as dou, and its meaning is “to cleave.” Therefore, /剅 is a better choice than zhuo 斲.
In the future, there might be an even better decipherment of graph . Nevertheless, for now, dou /剅 fits the context. Accordingly, the time of Eclipse V is not gengshen (night). Rather, it is jiwei (xi) dou gengshen, which refers to the time period when the night of jiwei cleaves to gengshen day. It straddles two days, as proposed by David N. Keightley. Interestingly, a total lunar eclipse in 2011 can be used to illustrate what time period jiwei (xi) dou gengshen actually refers to.
3. Red Moon on June 16, 2011
On the morning of June 16, 2011, a total lunar eclipse occurred in China. At its totality, the moon became red, which was splendid to many sky watchers. For me it was special, not because it was red, but because I realized it could help people understand the circumstance of Eclipse V inscribed in Yingcang 885 and 886.
3.1 Time specifications of the eclipse on June 16, 2011
The following time specifications of the eclipse in 2011 are available on the NASA website:Footnote 42
Moon enters umbra: 02:23
Start of totality: 03:22
Maximum eclipse: 04:13
End of totality: 05:03
Moon leaves umbra: 06:02
Umbral magnitude: 1.700
Duration of total phase: 1 h 41 m
Duration of umbral phase: 3 h 39 m
If observed in Beijing, the whole eclipse would not have been able to be seen because the moon had already set minutes after the total eclipse ended at 5:03 a.m.
SkyMap Pro9Footnote 43 provides the following time specifications for this lunar eclipse:
Moon enters umbra: 02:22:24
Start of totality: 03:21:57
Maximum eclipse: 04:12:24
End of totality: 05:02:50
Moon leaves umbra: 06:02:22
Umbral magnitude: 1.705
Duration of total phase: 1 h 40 m 52s
Duration of umbral phase: 3 h 39 m 58s
In Beijing, the whole eclipse was not observable. The sun rose at 05:04 and the moon set at 05:11 on June 16. After the moon set at 5:11 a.m., this eclipse was definitely not observable.
It is apparent that the above two sets of time specifications of this lunar eclipse are almost the same. Moreover, it is convenient to use SkyMap. Thus, I use SkyMap for time specifications for other lunar eclipses.
This eclipse started at night and ended when the moon set after sunrise. How then would the Yin people have recorded this eclipse?
3.2 Hypothetical inscriptional record of eclipse on June 16, 2011
In terms of the ganzhi date, June 15 and 16, 2011 are xinchou 辛丑 (day 38) and renyin 壬寅 (day 39), respectively. Had the Yin observed this eclipse, how would they have recorded it? The answer is not that simple. Scholars hold different opinions on the start of a Yin day. It is assumed that those different opinions would affect how it would be recorded in the language of the Oracle-Bone inscriptions. I will make clear that this is not the case by exploring how the 2011 eclipse would be recorded according to those opinions.
There are two schools of thought concerning the start of a Yin day: one is that it started at midnight, and the other is that it began sometime around daybreak. Scholars such as Homer H. Dubs, Chou Fa-kao, and David N. Keightley have all shared the former view, while many Chinese scholars hold the latter opinion.
3.2.1 If a Yin day started at midnight
For those who believe that a Yin day began at midnight, the eclipse on June 16, 2011 occurred totally during the time of day renyin and had nothing to do with day xinchou. The easiest way to record this eclipse would be to write an inscription such as renyin yue youshi 壬寅月有食, “on renyin, the moon was eclipsed.” However, this simple sentence is insufficient because it does not clearly state on which night the eclipse occurred: during the time period from midnight to six o'clock in the morning, as it actually did, or during the time between sunset and midnight? Clearly, more words are needed to show that it actually happened during the early morning on June 16, 2011.
It is unfortunate that there is not a single time word in the language of the Yin Oracle-Bone inscriptions that specifically refers to the time between midnight and sometime around sunrise. As a result, since a proper record of this eclipse cannot be produced simply by adding another word, other appropriate time words for recording this eclipse would need to be examined.
In the Yin Oracle-Bone inscriptions, time divisions after midnight (still nonexistent in the Oracle-Bone inscriptions) before noon of a Yin day include su 夙, dan 旦 (or ming 明, zhao 朝), dacai 大采 and dashi 大食.Footnote 44Dan refers to the time period from dawn to sunrise and su refers to a short time before dan. Apparently, both of these time divisions are relevant to the eclipse and could be used to record it. That inscription would read: renyin su dan yue you shi 壬寅夙旦月有食, “from su to dan on renyin, the moon was eclipsed.” Nevertheless, it would still not accurately represent the longest part of this nearly three-hour long eclipse. Therefore, it appears that a longer time division would be needed.
There are plenty of examples to show that a Yin day included two parts ri 日 and xi 夕, which may refer to the whole daytime and whole nighttime, respectively.Footnote 45 This is true even in English today. For example, the sentence “you are welcome to stay the night here,” means that “you may stay here till next morning”; it does not mean “you may stay here till midnight only.” Therefore, the duration of this eclipse from 2:22 a.m. to daybreak could be considered as part of the night of June 15, i.e., day xinchou. On the other hand, in the case of the short part of this eclipse that occurred around daybreak, it is clearly part of June 16 (i.e., day renyin). Therefore, in order to record this eclipse accurately, the two durations would have to be put together to produce such a record: xinchou xi dou renyin yue you shi 辛丑夕剅壬寅月有食, “during the time when the night of xinchou cleaved to renyin, the moon was eclipsed.”
3.2.2 If a Yin day started around daybreak
For those who maintain that a Yin day started around daybreak, it would be clear that the lunar eclipse on June 16, 2001 could be divided into two parts: the first duration, when it was dark and before daybreak, belonging to the night of June 15 (i.e., day xinchou); and the second duration occurring at daybreak and a short time after that on June 16 (i.e., day renyin). To record this eclipse, the two durations would have to be combined. The result will be the same as that of section 3.2.1: xinchou xi dou renyin yue you shi 辛丑夕剅壬寅月有食.
This eclipse could also be recorded as xinchou dou renyin yue you shi 辛丑剅壬寅月有食, “during the time when xinchou cleaved to renyin, the moon was eclipsed.” It is completely understandable that if the expression xinchou dou renyin were used to specify the time of this eclipse, the period when xinchou cleaved to renyin would necessarily include some portion of the night of xinchou. Thus, the word xi can be omitted.
3.2.3 Same hypothetical records
Whether or not one believes that the start of a Yin day was midnight or daybreak, it is a significant speculation that hypothetical inscriptional records of the lunar eclipse on June 16, 2011 would in either case be the same: xinchou xi dou renyin yue you shi 辛丑夕剅壬寅月有食 or xinchou dou renyin yue you shi 辛丑剅壬寅月有食. Moreover, based upon current lunar eclipse records in the Oracle-Bone inscriptions, there is no other way to record this eclipse accurately in the language of the Oracle-Bone inscriptions. Therefore, this eclipse could only be recorded as xinchou xi dou renyin yue you shi 辛丑夕剅壬寅月有食 or xinchou dou renyin yue you shi 辛丑剅壬寅月有食.
The above hypothetical inscriptional records of the 2011 eclipse are the same as those of Eclipse V inscribed in Yingcang 885 and 886. It follows that Eclipse V started some time on the night of jiwei and was over shortly after sunrise on gengshen.Footnote 46 This is the key to understanding the dating of Eclipse V.
4. Absolute Date of Eclipse V
4.1 Possible dates
With the advancement of modern astronomy, it is indeed fast and easy to calculate circumstances of lunar eclipses that have happened in ancient history. For the beginning of the Yin period, very few, if any, scholars still believe that it could be earlier than 1400 b.c.e. For the end of the Dynasty, the latest date among historians up to now is 1018 b.c.e.Footnote 47 Yin King Zugeng 祖庚 was the fourth king, and there were seven kings after him. David N. Keightley has estimated that these seven kings ruled for about 130 years.Footnote 48 The possible last year for the reign of King Zugeng, as well as the possible last year of Bin diviner group inscriptions or Period I inscriptions, should be (1018 + 130 =) 1148 b.c.e.
Between 1400 b.c.e. and 1148 b.c.e., there were only two lunar eclipses that straddled jiwei and gengshen, according to both above-mentioned views on the start of a Yin day. Their circumstances are as follows:Footnote 49
I. Eclipse on December 27–28, 1192 b.c.e.:
Moon enters umbra: 1192 b.c.e. December 27 20:20:43
Start of totality: 1192 b.c.e. December 27 21:25:38
Maximum eclipse: 1192 b.c.e. December 27 22:17:05
End of totality: 1192 b.c.e. December 27 23:08:33
Moon leaves umbra: 1192 b.c.e. December 28 00:13:28
Umbral magnitude: 1.658
Duration of total phase: 1 h 42 m 55s
Duration of umbral phase: 3 h 52 m 45s
II. Eclipse on August 14, 1166 b.c.e.
Moon enters umbra: 1166 b.c.e. August 14 03:14:02
Start of totality: 1166 b.c.e. August 14 04:16:59
Maximum eclipse: 1166 b.c.e. August 14 05:07:42
End of totality: 1166 b.c.e. August 14 05:58:24
Moon leaves umbra: 1166 b.c.e. August 14 07:01:21
Umbral magnitude: 1.628
Duration of total phase: 1 h 41 m 24s
Duration of umbral phase: 3 h 47 m 19s
On August 14, 1166 b.c.e., the sun rose at 05:24 and the moon set at 05:29 in Anyang. This lunar eclipse was partially observable to the Yin in the capital.
4.2 1166 b.c.e. as the absolute date for Eclipse V
Between the above two lunar eclipses, the one on December 27–28, 1192 b.c.e. is not a match for Eclipse V due to the three reasons stated in footnote 46. It is a fact that this eclipse in 1192 b.c.e. happened entirely during the night of jiwei and would have been recorded as jiwei xi yue youshi 己未夕月有食, “On the night of jiwei, the moon was eclipsed.” This inscriptional record appears to be perfectly acceptable, as in the case of those for Eclipse I–IV listed at the beginning of section 1 of this article.
As summarized in section 3.2.3, inscriptional records of a lunar eclipse recorded as ganzhi date (xi) + dou + ganzhi date correspond to lunar eclipses starting sometime during the night of the first ganzhi date and ending shortly after sunrise on the next day. The eclipse on August 14, 1166 b.c.e. started at 3:14 a.m., the dark night of jiwei, and ended at 5:29 a.m., shortly after sunrise on gengshen. Its time specifications match the circumstances of Eclipse V, whose time was jiwei (xi) + dou + gengshen, “the time when (the night of) jiwei cleaved to gengshen.” Therefore, the absolute date of Eclipse V is August 14, 1166 b.c.e.
Conclusion
With regard to records of Eclipse V carved on the back of Yingcang 885 and 886, the timing of this eclipse has been hotly debated for sixty years, mainly because scholars hold different opinions on the graph . After testing decipherments of the graph against paleographical evidence and the pattern of inscriptions of lunar eclipses, it has been found that reading this graph as dou 剅, meaning “to cleave,” fits the context of lunar eclipse inscriptions. In the inscriptions of Eclipse V, the expression jiwei xi dou gengshen 己未夕剅庚申 specifies the time of Eclipse V, meaning “when the evening of day jiwei cleaved to day gengshen.” Eclipse V straddled those two dates.
The spectacular red total lunar eclipse on June 16, 2011, starting at 2:22 a.m. and ending at 5:11 a.m., was only partially observable in China. Had the Yin watched and recorded it, as analyzed in section 3, according to current attested time divisions in the Oracle-Bone inscriptions and a pattern of extant inscriptions of lunar eclipses, the record could only be xinchou (xi) dou renyin yue youshi 辛丑(夕)剅壬寅月有食, “during the time when (the night of) xinchou cleaved to renyin, the moon was eclipsed,” regardless of whether one holds the view that a Yin day began at midnight or around daybreak. The structure of hypothetical inscriptional records of the June 2011 eclipse is identical to that of Eclipse V as recorded in the language of the Oracle-Bone inscriptions. Furthermore, as explained in footnote 46, there are three reasons for rejecting the view that an eclipse that ends not long after midnight could be recorded in the same way as Eclipse V. Therefore, according to this innovative approach to the dating of Eclipse V, it must have been similar to the eclipse that occurred on June 16, 2011, starting at night and ending shortly after sunrise.
The timeframe of Eclipse V is from 1400 b.c.e. to 1048 b.c.e. During that time period, the lunar eclipse on August 14, 1166 b.c.e. is the only one that corresponds to the circumstances of Eclipse V. Thus, the absolute date of Eclipse V is August 14, 1166 b.c.e. In 1998, Chang Yuzhi and I independently put forward this view. In this article, I reach the same conclusion by means of a novel approach.
deng zi 釋殷虛卜辭中的
