Wèi Emperor Míng, Cáo Ruì, died on the 22nd of January, 239, reportedly aged thirty-six years1 by traditional reckoning. In the old system, the count of the years of age starts at one rather than at zero at birth and the count is increased upon the start of a new calendar year, not the individual birth date. However, Péi Sōngzhī noticed a problem with this count:
Your Servant Sōngzhī notes: Wèi Wǔ [Cáo Cāo] in Jiàn'ān Ninth Year [204] Eighth Moon settled Yè, Emperor Wén first accepted empress Zhēn, Emperor Míng should have been born in the Tenth Year [205], calculating to reach to this year [239] Standard Moon, the total is only thirty-four years. At the time the Standard Moon had changed, with the former year’s twelfth moon became the present year’s standard moon, which could increase the count to thirty-five years, not to thirty-six.2
Most likely Péi Sōngzhī was seeking to point out a simple calculation error. A calculation error surely would not be too surprising giving the potential confusions caused by the Jǐngchū calendar reforms, as Péi Sōngzhī himself pointed out. The calendar reforms had moved the start of the new year to be earlier by one month, and Cáo Ruì had died on exactly the new year’s day of the new calendar. The start of a new year would increase the years count under traditional reckoning, but apparently someone had made another error somewhere, leading to a doubled increase to 34 to 35 to 36 instead of 34 to 35.
Qīng dynasty era scholar Pān Méi argued for a different age estimation entirely:
Yánkāng Inaugural Year used the Wǔdé enfeoffment on Cáo Ruì to be his state, the Basic Annals records his years were fifteen when enfeoffed as Marquis of Wǔdé. Yánkāng Inaugural Year was Jiàn'ān Twenty-fifth Year [220], the Emperor’s years were fifteen, then he was born in Jiàn'ān Eleventh Year [206].3
By this count, Cáo Ruì would have been 33 in traditional reckoning, increased to 34 if using the earlier New Year’s date from the Jǐngchū calendar reform.
The late-Qīng-early-R-early-PR era scholar Lú Bì (1876 - 1967) however argued that 36 was not an error at all, but an intentional message from Chén Shòu to secretly hint that Cáo Ruì was the son of Yuán Xī. To argue against the dating of the Wǔdé fief to 220 advocated by Pān Méi and others, Lú Bì wrote:
Sir Hóu said that Zhōu Fāngshū mistakenly divided Yánkāng Inaugural Year and Huángchū Inaugural Year into two years; this statement is correct. However the various scholars all rigidly cling to the writing “Yánkāng Inaugural Year, aged fifteen, enfeoffed as Marquis of Wǔdé,” and therefore suspect the Records writing beginning to end is uneven. Note that in the Wén jì [SGZ 2] before Huángchū Inaugural Year, it recollects many past events, not only events in Yánkāng Inaugural Year. Wèi Emperor Míng’s enfeofment as Marquis of Wǔdé should be before Yánkāng. Note the Cháng Lín zhuàn [SGZ 23] annotation citing Wèi lüè says: “Jí Mào transferred to Resident to the Marquis of Wǔdé. Twenty-Third Year [218] in connection to the case of his clansman Jí Běn and others rebelling, was arrested. It happened Chancellor of State Zhōng [Yáo] testified Mào’s and Běn’s degree of relation was already separated, and therefore he was able to not be connected,” so Jí Mào in becoming Resident to the Marquis of Wǔdé was an event in Jiàn’ān Twenty-Third Year, Wèi Emperor Míng’s enfeoffment as Marquis of Wǔdé also should have been at this time. If this event with the records written age of fiteen at enfeoffment of a marquisiate are in agreement, then in Jǐngchū Third Year the age of thirty-six also is no mistake. I humbly say [Chén Shòu] Chéngzuò in this writing, was in fact a “bent brush,” those reading the histories working back from the years and moons, verified by the lady Zhēn being bestowed death, by Wèi [Emperor] Míng for a long time not being able to be established as successor, then Yuánzhòng being whose son can be known without speaking of it.4
As a side note, that first line about Yánkāng 1 and Huángchū 1 being accidentally counted as separate years (Jiàn’ān 25 = Yánkāng 1 = Huángchū 1 = 220 AD) is a very plausible explanation for a mistaken extra year increase in a years of age count, which Lú Bì apparently acknowledged as a valid argument, but did not accept, instead favoring his own argument for a Yuán Xī paternity.
Being the most recent of the famous historical commentators on the text, Lú Bì has tended to be treated as the “final say” on the subject, and probably why so many English language commentators now love to claim that Cáo Ruì was Yuán Xī’s son.
There are serious problems with Lú Bì’s argument, however.
First, Lú Bì’s objection that the SGZ 2 could be recounting events that occurred prior to 220 apply just as much as to the Wèi lüè annotation to SGZ 13, and arguably even more so. The SGZ 2 writing makes the point of explicitly giving a date just prior to the record of the Wǔdé enfeoffment,5 which would seem to be taken as an intentional structuring to date the enfeoffment to that year. The annotation to SGZ 13 meanwhile has the recorded transfer before the dating of Jiàn’ān Twenty-Third Year (218), which admittedly could imply the event is supposed to be considered as taking place before 218, but could also be interpreted as a more general statement on Jí Mào’s overall career in a summary form, before giving a separate more detailed anecdote dated to a specific year. Furthermore, the use of titles anachronistically in the historiography is well known and demonstrated (consider how many times people are referred to by a posthumous name, such as say, I don’t know, “Wèi Emperor Míng”), meaning that the identification of Jí Mào’s appointment as Resident to “Marquis of Wǔdé” could very well refer to an appointment to attend to Cáo Ruì’s person and not to the specific Marquisate of Wǔdé.
Additionally, if comparing accuracy of details given in the core SGZ and in the Wèi lüè, I would tend to favor the SGZ as less likely to contain errors, since the Wèi lüè, while certainly still a valuable source, does suffer from several rather infamous known inaccuracies,6 which makes it more plausible that Wèi lüè may have simply gotten the dating of an appointment wrong.
Second, Lady Zhēn being bestowed death and Wèi Emperor Míng not being established as a successor cannot possibly be arguments for paternity fraud. Yè was captured in the Eighth Moon. If Cáo Ruì had been born in the same calendar year, his birth would have been within four months of the lady Zhēn and Cáo Pī first meeting, so Cáo Pī would immediately have known the boy was not his (contrary to popular belief, the ancients did know how pregnancy worked).7 No paternity fraud could have taken place if Cáo Ruì was born in the Ninth Year (204).
One could perhaps try to argue back that perhaps Cáo Pī decided to adopt the boy as his own son despite knowing it was not his biological son, but then he would have known what he was getting into from the beginning, meaning that the thought of Cáo Ruì not being his biological son should not have factored in to Cáo Pī’s later decision to order the lady Zhēn’s death in 221, some seventeen years later, and he certainly would not have established Cáo Ruì as his successor at all.
And this leads to one more serious fatal flaw to the argument that 36 is correct and evidence of a conspiracy covering up a Yuán Xī paternity: the siege of Yè.
Ninth Year [204] Spring Standard Moon, [Cáo Cāo] crossed the Hé, dammed the Qí river to enter the Bái canal to connect a provisions route. Second Moon, [Yuán] Shàng again attacked [Yuán] Tán, leaving Sū Yóu and Shěn Pèi to defend Yè. His Excellency [Cáo Cāo] advanced the army to the Huán river, [Sū] Yóu surrendered. At once arrived, attacked Yè, building earthen mounds and tunnels.8
Fifth Moon, demolished the earthen mounds and tunnels, building an encircling moat, diverting the Zhāng river to flood the city; inside the city those that died of hunger were over half.9
Eighth Moon, Shěn Pèi's elder brother's son Róng in the night opened the city's east gate he was defending to let in the troops. Pèi opposed in battle, defeated, captured Pèi alive, beheaded him, Yè was settled.10
Unless you want to argue Yuán Xī was capable of sneaking past the siege lines, which included a moat that cut off all communication for the last three months of siege, undetected, despite his fame and stature (perhaps he was capable of invisibility or teleportation?) he must have been outside of Yè for at least three months due to the moat built in the Fifth Moon, but more likely at least six months due to the siege first beginning in the Second Moon, and even more likely for years due to being appointed to serve in an entirely different province.11
We lack an exact date on when Yuán Shào installed his son Yuán Xī as Inspector of Yōu province, but we can estimate that it may have been sometime soon after 199, as that was the year that the warlord Gōngsūn Zàn, the previous main power in Yōu province, was destroyed and the territory of Yōu conquered by Yuán Shào.12 One might object that just because he was appointed to the territory, he need not necessarily have been sent immediately, or even necessarily gone there, given that we have examples of other individuals holding appointments they either did not actually physically go to or else ran from mobile offices outside the territory itself, but we are also explicitly told that after being defeated by Cáo Cāo, Yuán Shàng fled north to join Yuán Xī, so Xī most likely was physically in Yōu province for some time in order to have established himself well enough to become a source of refuge for his brother.
So, even supposing if Cáo Ruì had been born in Jiàn’ān Ninth Year (204) and too old to be Cáo Pī’s biological son, Yuán Xī could not have been the father either on account of not being in the same city, and very likely in a completely different province. Supposing if Cáo Ruì had been born in Jiàn’ān Tenth Year (205) instead, late enough for his paternity to be in doubt, then that makes Yuán Xī paternity even more implausible due to all those months that Yè city was under siege.
In summary, even if one were to wish to doubt the paternity of the Emperor Míng, Yuán Xī would actually be one person we can be fairly certain was not the father.
Aside from the previously mentioned sources of confusion such as mistakenly counting a single year as separate calendar years (Jiàn’ān 25 = Yánkāng 1 = Huángchū 1 = 220 AD) or the confusion as to whether or not to use Jǐngchū Calendar’s alternate New Year’s Day in the count, there is one other rather boring but still somewhat plausible explanation possibly worth mentioning: a simple transcription error.
Consider “six” 六 and consider “three” 三: if the top and bottom lines of “three” were damaged and obscured, the top line could be misinterpreted as a dot and the bottom line could be misinterpreted into two dots, causing “three” to be mistaken for “six.” Note also that replacing “thirty-six” with “thirty-three” sets a birth year to Jiàn’ān Eleventh Year or 206 (if not including the extra year added by using the Jǐngchū calendar’s alternative New Year date), which matches perfectly with Pān Méi’s SGZ 2-SGZ 3 correspondence argument of Cáo Ruì being 15 (traditional reckoning) in 220.
《三國志·魏書三·明帝紀》三年春正月丁亥,太尉宣王還至河內,帝驛馬召到,引入卧內,執其手謂曰:「吾疾甚,以後事屬君,君其與爽輔少子。吾得見君,無所恨!」宣王頓首流涕。即日,帝崩于嘉福殿,時年三十六。
《三國志注·魏書三·明帝紀》臣松之桉:魏武以建安九年八月定鄴,文帝始納甄后,明帝應以十年生,計至此年正月,整三十四年耳。時改正朔,以故年十二月為今年正月,可彊名三十五年,不得三十六也。
潘眉《三國志考證》:「延康元年以武德封曹叡為國,本紀載年十五封武德侯,延康元年即建安二十五年,帝是年十五,則生於建安十一年也。」
盧弼《三國志集解》:侯氏謂周方叔誤分延康元年、黃初元年為二年,其説誠是。惟諸家皆拘泥“延康元年,年十五,封武德侯”之文,遂疑志文前後參差。按《文紀》黃初元年以前,多追述往事,不盡為延康元年之事,魏明封武德侯,當在延康以前。按《常林傳》注引《魏略》雲“吉茂轉為武德侯庶子。二十三年,坐其宗人吉本等起事,被收。會鐘相國證茂、本服第已絕,故得不坐”,是吉茂之為武德侯庶子為建安二十三年事,魏明之封武德侯亦當在此時。若此事與志文年十五封侯相合,則景初三年年三十六,亦不誤矣。竊謂承祚此文,實為曲筆,讀史者逆推年月,證以甄夫人之賜死,魏明之久不得立為嗣,則元仲究為誰氏之子,可不言而喻矣。
《三國志·魏書二·文帝紀》五月戊寅,天子命王追尊皇祖太尉曰太王,夫人丁氏曰太王后,封王子叡為武德侯。
Yes, I’m talking about Wèi lüè’s version of Liú Shàn’s childhood annotated to SGZ 33, which Péi Sōngzhī absolutely demolished as completely ridiculous in his comments.
See for examples these quotations from various encyclopedias and treatises compiled during Hàn, which all pre-date Cáo Pī:
《論衡 》母懷子氣,十月而生。The mother harbors the child’s aura, ten moons and then birth.
《白虎通》人所以十月而生者何?Why are people born after ten moons?
《孔子家語》人十月而生。People are born after ten moons.
Note that a moon is approximately 29.5 days and thus shorter than the modern month which is usually 30 or 31 days, hence “ten moons” and not “nine months.”
《三國志·魏書一·武帝紀》九年春正月,濟河,遏淇水入白溝以通糧道。二月,尚復攻譚,留蘇由、審配守鄴。公進軍到洹水,由降。旣至,攻鄴,為土山、地道。
《三國志·魏書一·武帝紀》五月,毀土山、地道,作圍壍,決漳水灌城;城中餓死者過半。
《三國志·魏書一·武帝紀》八月,審配兄子榮夜開所守城東門內兵。配逆戰,敗,生禽配,斬之,鄴定。
《三國志·魏書六·袁紹傳》出長子譚為青州,沮授諫紹:「必為禍始。」紹不聽,曰:「孤欲令諸兒各據一州也。」又以中子熈為幽州,甥高幹為并州。
《後漢書·卷七十四上·袁紹傳》於是以中子熙為幽州刺史,外甥高幹為并州刺史。
《後漢書·卷七十四上·袁紹傳》四年春,擊公孫瓚,遂定幽土,事在瓚傳。