Succession Struggle and the Ethnic Identity of the Tang Imperial ...

Succession Struggle and the Ethnic Identity of the Tang Imperial House
Author(s): Sanping Chen
Source: Journal of the Royal Asiatic Society, Third Series, Vol. 6, No. 3 (Nov., 1996), pp. 379-
405
Published by: Cambridge University Press on behalf of the Royal Asiatic Society of Great Britain and
Ireland
Stable URL: http://www.jstor.org/stable/25183243 .
Accessed: 11/08/2011 18:21
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
http://www.jstor.org
Cambridge University Press and Royal Asiatic Society of Great Britain and Ireland are collaborating with
JSTOR to digitize, preserve and extend access to Journal of the Royal Asiatic Society.

Succession Struggle and the Ethnic Identity of the
Tang Imperial House*
SANPING CHEN
Introduction
One of the questions that has prompted this study is this: despite the fact that various
Xiongnu tifttyL and Xianbei j^$. groups dominated the political arena in Northern China
for almost 300 years between the collapse of the Western Jin ? and the founding of the Sui
RS, and the usually unacknowledged
several hundreds more
years,1 why
fact that their descendants continued to do so for
is it that one could learn so little about their cultural
heritage in traditional historiography?2 Even the linguistic affinity of the Tuoba ifi&$t
Xianbei remains to this day
a matter of
controversy.3
In his narration of the Turco-Persian Ghaznavid sultanate in the eastern Iranian world,
David Morgan
Although
had this
interesting
the Ghaznawids were of Turkish
observation:
origin, there seems to have been little that was
Turkish about the way in which their empire was run, or indeed about the culture
We should, however, remember that our sources were written by Persian
might have been unlikely
present.4
Not
by coincidence, Herbert Franke, while
conspicuous "conquest" dynasty
A Chinese official
history
to
identifiably
they patronized.
contemporaries,
lay much stress on the non-Persian . . . elements that may have been
in Chinese
like the Yuan-shi is not
history,
discussing
the
commented:
very explicit
who
legitimation of the most
about the Buddhist and Lamaist
elements inherent in Yuan statehood, and one has to turn to the Tibetan and Mongol
*
The author would like to thank several
colleagues
and friends for their
help and comments. In particular, he
thanks Professor Samuel Adshead of the University of
Canterbury for
furnishing him with a
copy of the late
Joseph Fletcher's unpublished article on blood tanistry, and Dr Chungmo Kwok for extensive editorial and
other help.
1
Sima Guang WJ^3fe et al., Zizhi tongjian JSfcHIMr (Beijing, 1956; hereafter
ZZhTf), 108.3429, comments
by Hu Sanxing ^JH^
.
2
Even on those external "barbarians" who might be considered China's
primary "security concern", the
records were still astonishingly meagre, as observed by Denis Sinor, "Central Eurasia", in Orientalism and
History,
ed. D. Sinor
(Bloomington, 1970), pp. 93-119.
3
S.
-
Chen, "A-gan revisited the Tuoba's cultural and political heritage", fournal of Asian
History, forth
coming.
4 David Morgan, Medieval Persia, 1040-1797 (London, 1988), p. 22. See also C. E. Bosworth, The Ghaznavids
(Edinburgh, 1963), p. 56.
JRAS, Series j, 6, 3 (1996), pp. 379~405

sources,
ful.5
even
though
We shall demonstrate that both Morgan's
380 Sanping Chen
the latter ones are mostly relatively late and sometimes unreliable and fanci
and Franke's observations would prove pertinent
to the Tang ^ records too. However, unlike the case of the Ghaznavids, which was
contested
left
today
by Fuad Koprulu,
to question
the "all-Chinese"
were few alternative sources on the Tang,
politically
but also
culturally,
a modern Turkish historian,6 there is not a single Xianbei soul
Tang history;
the vast land from
which had
Samarqand
A related issue is the so-called process of sinicization.
historiography,
there has been this convenient but somewhat
apparently quite well entrenched, by
either a "native"or a "conquest" regime.
undoubted strong "Northern influences", were
the process of "sinicization" of the Xiongnu
considered
completed by
albeit consistent with the traditional
scrutiny.
In her
study
of
zation" (or "sinification")
appear
a more
which various Chinese
and unlike the Yuan 7E world, there
totally dominated, not
Largely
to the Sea of Japan.
only
based on the standard
arbitrary categorization,
dynasties
now
were classified as
In this scheme the Sui and Tang, though
invariably regarded
and Xianbei groups
as native
then.8 We shall demonstrate that this mainstream
historiography,
early modern China, Pamela
appropriate,
is an obsolete
Crossley
may
concept.9 "Ethnicity,"
not
at least morphologically (and politically)
quite
regimes,7
with
in northern China was
stand an
viewpoint,
in-depth
contends that the term "sinici
or ethnic
study,
would
and a more correct
substitute. However, even she seems to admit that an abstract notion of "ethnicity"
more self-evident, nor more
The
question
of
appropriate
strate in this essay, via concrete
Tang
and other "native"
attributes of a dynasty,
house and the
on
characterizing
contemporary
the Tang
imperial
as well
clearly defined, than the "obsolete" concept of sinicization.
nomenclature
examples
houses
as a conspicuous
Confucian
notwithstanding,
we shall try
to demon
is no
and facts, the marked contrast between the
regarding
gentry
as a "native" or
contend that the first half of the Tang might
Xianbeo-Chinese) regime.10
succession and other
politico-cultural
cultural gap between the Tang imperial
class. In the end,
"basically
be more
aptly
sinified"
we raise serious doubts
imperial
house. We
called a Sarbo-Chinese (or
5
Herbert Franke, Prom Tribal Chieftain
to Universal Emperor and God: The Legitimation of the Yuan
Dynasty
(Munich, 1978), p. 609.
6
See Bosworth, op. cit., p. 3.
7
See for examples John King Fairbank, "Synarchy under the Treaties", in Chinese Thought and Institutions, ed.
J. K. Fairbank (Chicago, 1957), pp. 204-31, in particular p. 204, and Karl A. Wittfogel, Oriental Despotism;
a
Comparative Study of Total Power (New Haven, 1957), p. 182.
8
David Honey, "Stripping off felt and fur: an essay on nomadic salification", Papers
on Inner Asia. Ancient
Inner Asia, XXI (1992), seems.the only exception
in
trying
to include the Tang in the "conquest dynasties". Yet in
addition to his very curious exclusion of the preceding Sui, he still considers the Tang house "basically sinified"
(p. 2). Therefore in his otherwise colourful essay on nomadic sinicization covering the entire Chinese history from
the late Shang
to the Qing,
not a
single
item or case
pertaining
to the Tang
was found.
9
Pamela K.
Crossley, "Thinking about ethnicity
in
early modern China", Late
Imperial China, XI (1990),
pp. 1-34.
10
For the etymology Xianbei < *Sarbi, see for example Peter B. Golden, An Introduction to the History of the
Turkic Peoples (Wiesbaden, 1992), p. 69.
as

Succession Struggle and the Ethnic Identity of the Tang Imperial House 381
The Ethnic Identity of the Tang Imperial House
The conventional view that the Tang represented
rests on the premise
by
the time of the
that the Li 2j5 were either of Han
founding
a "native" Chinese
origin
or had
dynasty very much
"basically
of the dynasty. We shall contend that neither was true.
sinified"11
The well-documented fact that the Li had intermarried with various Xianbei and other
non-Han families for generations notwithstanding,
the clan's own ethnic
origin
has been a
controversy oft-debated as a result of
conflicting
evidence. For
example,
the noted
Tang
authority Chen Yinke Hdft& had written no fewer than four articles trying
to prove the
Li clan's Han
origin
on the paternal
descended from the Tuoba Xianbei, had
side.12 Yet the opposite proposition, namely
equally
solid if not
that the Li
stronger evidence, as acknowl
edged by Hu Rulei iMiQS's biography of Li Shimin ^ffiS.13 Moreover, the studies by
Chen Yinke and others have shown that the official histories compiled during the Tang had
been
subjected
house's "barbarian"
to much
background.
political doctoring
The newest
in order to mask and conceal the
proof
is the recent
imperial
archaeological discovery
showing
that another
contemporary prominent
Li clan, namely
that of Li Xian ^S" the
Northern Zhou M Grand General with the same
Longxi fiSffi ancestry claim, was in fact of
unmistakable Tuoba Xianbei descent.14
As many have maintained, ethnic
Northern
persistent
aristocracy
as a whole had
origin
is not of real
supposedly "basically
or occasional "atavistic" appearance of Northern influences
significance here, since the
sinified"
by
notwithstanding,
then. Some
could argue that little was found in the standard histories to suggest that the Li regime was
culturally anything
We
beg
but a "native" Chinese
to differ. The earlier quotes
dynasty.
on the Ghaznavids and on the Yuan
demonstrate how one-sided sources created biased or even false politico-cultural images
an ethnic
regime.
Careful examination of the historical sources of the era reveals many
of the Li clan's non-Han cultural traits and
marked distance between the
these issues, as well as the
imperial
contemporary
brief summary of some most remarkable cases.
(i) Language.
had been carried on by
term Guoyu i9fj? "national
identity. What may be more
house and the traditional Chinese
awareness of this difference. The
important
In a recent article this author has shown that the Tuoba Xianbei
the Li clan as their first or
language"
was
kept
one
clearly
gentry regarding
following
of
cases
is the
tongue
is a
family language.15 Moreover, even the
for a while
during
the Tang.16
11
Honey, "Stripping off felt and fur", p. 2.
12
They
are
"Li-Tang shizu zhi tuice wenti ^M$3k2Lfflcffl\ \ "Li-Tang shizu zhi tuice houji
^/SR$?;yE?,H?p5 ", "San lun
Li-Tang shizu wenti EM^S^Mf^M. "? and
"Li-Tang Wu-Zhou xianshi
shiji zakao "
All can be f?und m his collected works Chen Yinke
^^jSy^fctttSfrPS^
xianshenq lunwenji
IWWB6?a2*Craipei.i974).
13
Hu Rulei, Li Shimin zhuan ^tSSfll (Beijing, 1984), p. 2.
14
Ningxia Zizhiqu Bowuguan ^J[gtp(SW#|l& > "Mu faJue jianbao 3?f?JSffif#", WenwuJCfyj
(1985. n), pp. 1-20. This
implication of Li Xian's newly unearthed tomb
inscription on the ethnic origin of the
Tang imperial family was pointed out to me by Dr Chungmo Kwok via private communications.
15
S. Chen, "A-gan revisited".
16
Lui Pansui
$JKH?' "Li-Tang wei fanxing kao $^^Sftt^"", Nu shida
xueshujikan ^cSBX^^&^f J 1-4
(no date), pp. 1-5, first made this important discovery based on an entry in Ouyang Xiu
SJ$?iJ?
et al, Xin Tang
shu
Isff^tfip (Beijing, 1975; hereafter XTSh), 44.1160.
The

contemporary Chinese
gentry's
Yanshijiaxun Mfc^MW.17
382 Sanping Chen
attitude toward this was best reflected in a noted passage in
(ii) Affinity. The Tang was the last Chinese dynasty before the Manchus to marry off
royal princesses
gongzhu
^i "princess"
to the steppe khans and chieftains,18 so much so that the word
was
for "wife"!19 In the meantime, the
establish matrimonial relations with the
the honour of an
the late Tang era.20
simply regarded by Ziya Gokalp
leading
imperial marriage persisted
imperial
Chinese gentry families
house. More
konchuyl
as an ancient Turkic word
strikingly,
steadfastly
their
for more than two centuries, lasting
refused to
rejection
of
well into
(iii) Clan relationship. The Tang represented a unique case in Chinese history in that the
imperial
house bestowed its own clan name Li not
only
on a few Han Chinese persons but
more
frequently
on ethnic leaders and chieftains, be
they Turk, Tangut, Uighur,
Kitan or
Iranian.21 The Zhuxie %M Shatuo Wb Turkic tribe founded the Later Tang &M
dynasty
based
largely
on having
received this
imperial
period (841-6), the main business for a Kirghiz embassy,
Emperor Wuzong S^, was to register
other hand, the Tang imperial
clans' social
prestige
and privilege.24
house made
themselves with the
repeated
honour.22 As late as in the Huichang
as per
imperial
an
efforts to suppress the
imperial
edict
by
clan office.23 On the
leading
(iv) Clothing. It is well-known that the Tang fashions were under heavy foreign
influence.25 The love of exotic
style,
of Emperor Xuanzong "2C;|c .26 Much of this
and in sharp
contrast to the Tuoba
in which the emperor took
custom and make-up
Emperor
personal responsibility
the Xianbei women,27 most of the so-called
of the Tuoba
gentry,
military
we note that the felt hat
Taizong's brother-in-law, was later labelled
represented
became feverish
the steppe heritage.
during
For
gentry
the reign
example
Xiaowen 2?jCs wholesale sinicization drive,
"regular
for
abolishing
dresses 'SfBR." during
tradition.28 However, as to how this was viewed
even the leisure dress of
by
the Tang
were
the traditional
personally popularized by Zhangsun Wuji -UlS^fe? ,
by
the Confucian historians as "devilish".29
17
Yan Zhitui j|J|j2lJj?, Yanshijiaxun, edition used: Wang Liqi 3Etu^,
Yanshi jiaxun jijie ffitt&Wv&nr,
rev.
ed. (Beijing, 1993), 1.50-1. For an
English translation of the passage, see for example Arthur F. Wright, The Sui
Dynasty (New York, 1978), p. 36.
18
See the list in
Wang Pu 3:7lt, Tang hui yao JSW"Jc (Taipei, 1963; hereafter THY), 6.75. More extensive lists
were
given in Wang Tongling BE-fPHH^
, Yang-Sui Li-Tang xianshi xitong kao wm^J^5utS^iW%", Nw
shida xueshu jikan,
19
Ziya Gokalp
II.2 (no date), pp. 1-23.
(translated by R. Devereux), The Principles of Turkism (Leiden, 1968), p. no.
20 XTSh 119.4306, 172.5205-06, ZZhTJ 248.8036 and Li Fang $9j, Taiping guangji ^^f-jfUeB (Beijing,
1961), 184.1379. Read also Chen Yinke, Tangdai zhengzhishi shulun gao 0ft^?p3?3?l$iS5 (Chongqing/
Shanghai, 1944/1947),
21
Wang Tongling,
pp. 57~9
op. cit., pp. 20-3, has a rather extensive table on this issue. Still the important Kirghiz
case is
missing in that table.
22
See for example ZZhTJ 172.8879, Hu Sanxing's
notes.
23 XTSh 2i7b.6i50.
24 ZZ/iTJ 195.6135-36, 200.6138.
25 Read for instance Xiang Da
[njj||, Tangdai Chang'an yu Xiyu wenming Sfvil5?J^S^fcfc9t (Beijing,
1957)- _
26 Yuan Zhen7c|j|'s poem Faqu ??jfj (Quan Tangshi ^^S$ [Shanghai, 1986], 419.1025) may serve as a
succinct review of the fad.
27 Wei ShouSf|ltJc, Wei shu $$tf?t(Beijing, 1974) 19.469 and 21.536 (also ZZhTJ 140.4357 and 142.4434).
28 Liu Xu MSSJ/m Tang shu ||/if# (Beijing, 1975; hereafter JTSh), 45.1938, 1951, XTSh, 24.527-28, and
THY 31.577-78.
29 XTSh 34-878.

Succession Struggle and the Ethnic Identity of the Tang Imperial House 383
Another telltale case, as recorded in
Tang yulin 0tq#, was
Emperor Xuanzong's feeling
of
alienation ^-S merely
at his chief minister Zhang Yue 36h$?'s "Confucian dress".30 These
incidents again point at a gap between the Tang ruling class and the Confucian gentry in
this regard.
(v)
sucking
Social mores. Several noticeable
and
famous (or notorious)
foot-kissing betrayed
matrimonial relations. The
etiquettes practised by
the clan's non-Han cultural
the Li clansmen like breast
heritage.31
But the most
custom was the Li clan's record on levirate and other scandalous
practice,
as we have
pointed
out elsewhere, reflected a key
northern legacy in the Tang house, namely the lack of clearly defined and recognized
generational
boundaries on the
steppe.32
In addition to the many well-known cases, we
note the recently unearthed tomb inscription of the wife of the Turk general Ashina Zhong
ffl$M$&, which reveals yet another marriage of Emperor Taizong's with his former
in-laws.33 The case was not
such incidents that were
have been looked upon by
house's
ethnicity finally
Confucian standard
raised.34
governing
(vi) Another interesting
reported anywhere
similarly suppressed
in existing records, suggesting
even more
in the official histories. As to how this would
the Confucian gentry, suffice it to
became an open issue in the Southern
a women's
proper
say that when the Tang
Song,
behaviour" was the first
"violations of the
question
cultural trait was the "barbarian" childhood names fashionable
among the Northern aristocracy, the Sui and Tang houses included. Both Sui Wendi JCW
and Yangdi M$? had such names.35 So did Yang Yong ^fcpj and Li Jiancheng ^IfeiSc,36
the two princes who shared more than
examine later. There is good
but the real point
just being
an unsuccessful heir apparent,
evidence that some of these names were of Buddhist
is their non-sinicized forms. For
was also the name of a Central Asian Turk chief.37 An
example,
as we shall
origin,
Sui Wendi's name Naluoyan
intriguing story
is that a passage in
fTSh indicated unmistakably that Li Shimin also had such a childhood name.38 But
nowhere could this name be found in any records. One could
only
conclude that the
emperor
made sure his "barbarian" name became an absolute
imperial
taboo. Another case
is the childhood name Zhinu $?& of Li Zhi ^?p the future Emperor Gaozong j!(f^.39 His
father's use of a proverb: "Having
borne a wolf. . ." to describe Li Zhi's character40 leads us
30
Wang DangiEJH, Tang yulin (Changsha, 1939), 4.101. This was corroborated by XTSh 125.4407.
31
Liu Pansui, op. cit., was the first to point
out that both were non-Chinese etiquettes, though Liu's citations
were far from complete.
32
"A-gan revisited".
33
"Tang Ashina Zhong
mu
fajue jianbao Mffl!^$&^^Wi1Si%i'\ Kaogu %"? (1977.2), reprinted
in Tujue
yu Huihe lishi lunwen xuanji $SJ8rISIH1 tyjte$*jSfcfrijfipl,
ed. Lin Gan (Beijing, 1987), ii, pp. 408-19.
34
Zhu XiMM, Zhuzi yulei $tfrWM (Beijing, 1986), 136.3245.
35
Wendi's childhood name was
Nanluoyan SPiSS- See Daoxuan ilflr. Xu gaoseng zhuan ?||jtrf{M$ (Taipei,
1988), 26.900. Yangdi's
was Ame
for
fpffirSee example ZZhTJ 179.5577.
36
Yang Yong's childhood name was Gandifa
?^jtfef?j {ZZhTJ 179.5575). As we have pointed out in "A-gan
revisited", this name can be identified with a similar childhood name
Qizhifa cLMJfc of the "Xianbei-ized"
warlord Feng Ba $ij$t. See Wei shu, 97.2126. Li
Jiangcheng's childhood name was Pishamen flfet^P? (XTSh
79.3540).
37
ZZhTJ 212.6735.
38
JTSh 64.2415.
39
For this childhood name, see for example ZZhTJ 197.6206.
40
ZZhTJ 197-6208.

to submit that this
term for "wolf".41
seemingly
384 Sanping Chen
Han name was but a corrupted
or masked
proto-Mongolian
(vii) Yet another issue showing the marked contrast between the Tang imperial house
and the Confucian gentry was the monarchs' extravagant patronage
of the performing
? music, dance, drama, and other entertainment, much to the horror of the Confucian
moralists. Worse still, the Turco-Xianbei emperors often showed little reservation in
bestowing
on these artists, considered of the same social class as house-slaves and prostitutes
by
the traditional Chinese
gentry, prominent
and
prestigious
titles. One such artist was
even enfeoffed with a
princedom by the Northern Qi Htff, a
precedent the Sui Emperor
Yangdi
once wanted to follow to benefit his favourite and talented Kuchaean musician Bai
Mingda 6Kill (who would continue to serve the Tang with distinction).42
Tang emperors
were both criticized
to these artists.43 The third, Emperor Gaozong,
the artists undue
privileges.44
by
Confucian ministers for similar
arts
The first two
appointments given
also received similar criticism for according
We further remark that Emperor Xuanzong
was the last
Tang emperor
to show this
passion
for the
performing
arts.45 The same fervour was not to
be observed until the coming of the Shatuo Turkic regimes.46
The above items illustrate the Li clan's cultural identity in the eyes of the contemporary
Chinese gentry class. In addition, we find the views of two other
parties, namely
the Turks
and the Li clan members themselves, on this
subject suggestive
too.
First the Turks in the Orkhon inscriptions, probably the only independent historical
source of the era, called the Tang power Tabgach,
or Tuoba, fully
two centuries after the
collapse
of the last Tuoba
regime.47
Because of the
paucity
of data, it is difficult to ascertain
the Turks' exact
geographic perception of contemporary East Asia. But first Sui shu PffiF
clearly
showed that the Turks were well aware of the existence of the Southern state of
Chen ?fc.48 Then even after several hundred years al-Kasyari
Tawyac
was
only part of Sin. Moreover, and against
Russian word KHTafi, al-Kasyari
"[t]he
Secondly,
appeared
also gave
an etymology
name of a tribe of the Turks who settled in those
the attitude of the early Tang regime
to be a most useful indicator of its own
the inevitable
regions"!49
stated
analogy
of the name
unmistakably
of the modern
Tawyac:
that
It was
toward the traditional Chinese gentry
self-identity.
In addition to the
regime's
persistent
efforts in
depressing
the
prestige
of the traditional
gentry
class as mentioned
earlier, Li Yuan $#$, the founding emperor, had this explanation for his son Li Shimin's
growing political independence and aspiration:
41
This term was well attested as the clan name Chinu BfcJK., which became Lang

This boy has
longer my
Succession Struggle and the Ethnic Identity of the Tang Imperial House 385
long been commanding troops in the provinces. Instigated by educated men, he is no
son of the old days.50
Though cited by
a great many authors, few have noted the critical importance of the fact
-
that the
phrase
"educated men" was
literally
dushu Han *MRH "educated Han" in
fTSh.51 Sima Guang, most likely based on late (Song ^ dynasty) understanding, changed
it to a more
elegant
word
wholeheartedly corrupted by
private imperial quarters
shusheng itF^i , editing
out the crucial
the "educated Han" as the father had
and in the presence of his consort
implication. Perhaps,
charged,
Empress Zhangsun
the son, in the
of Tuoba
descent, would also call the famous courtier Wei Zheng W&L a tianshe Han B9#3|
"house-owning
Han
peasant".52
These two
quotes
are the best reflection of the Li clan's
ethnic self-identity, for in the period immediately proceeding the Sui and Tang, the term
Han {? as used in such phrases was always a derogatory appellation by
other such Northerners for the Han or otherwise sinicized
people.53 Although
the Xianbei and
the slur was
gradually picked up by other courtiers during Wu Zetian jRJIlK's reign, an episode
recorded in Chaoye qianzai MSfHslM clearly showed that the traditional meaning and
target were not lost among the gentry
of the term could be found than the father/son's
"house-owning peasant", typical
to nomadic
historians
tribespeople.
actually corrupted
class members.54 Besides, no better "classical
of a sedentary, agricultural
The cases are also
the original
stories.
If one finds the above arguments and evidence
represented just
Northern exposure,
identity
Tang,
application
good examples
fragmentary,
the occasional and transient remnants of the
can be established by
a far clearer and better focused
a systematic analysis
as we shall endeavour to show in the
following
picture
Central
of how
usage"
to "educated men" and
or
of the
of the succession
sections.
imperial
The Case of Crown Prince Chengqian jjfcjfc
Let us first examine the case of Emperor Taizong's original
Kingdom
but
foreign
"elegant' editing by
argues that the cases
imperial
heir apparent
struggles
house's
previous
clan's ethnic
in the early
in some detail,
because it
epitomizes
not
only
some crucial
political
and cultural issues of the era, but also
the problems with both standard historiography and modern scholarship in studying these
issues. In addition, this
episode
also sheds
light
on the so-called process of "sinicization".
The case is succinctly described (or one may say glossed over) in The Cambridge History of
China which has this to say about Taizong's
The prince apparently was
began
intelligent
and capable,
first heir
apparent:
... As he grew older, however, the heir apparent
to behave in ways which seemed both abnormal and scandalous to the Chinese courtiers, and
50
ZZhTJ 190.5959.
51
JTSh 64.2415-16.
52 Liu Su, , Sui Tangjiahua Ffttgig (Beijing, 1979), 1.7; Liu Su #Jjfc, Da Tang xinyu ^Sff^
$?)$?
(Beijing, 1984), 1.13. Again Sima Guang edited it to a mere
-
tianshe weng 0^^"
"old house-owning peasant"
in
ZZhTJ 194.6096.
53
Read for
example
Chen
Yuan^g, Tongjian Huzhu biaowei 3?96S5ct^SS (Beijing, 1958), pp. 323-4,
though Chen quoted only
a small number of cases.
54 Zhang Zhuo iffii** ? Chaoye qianzai (Beijing, 1979), 4.89.
-

he may well have been mentally
used the Turkish
language,
This is indeed an accurate narration
take the official lines, especially
Arthur Wright
"strange neurosis", but he does
Edwin
calls Prince
style".57
Pulleyblank gives
386 Sanping Chen
unbalanced. He began
to reject his Chinese identity and heritage,
and dressed himself and his entourage in Turkish costume . . .55
according
the alleged
"mental
to the standard
problems",
also calls the whole affair "a scandal" caused
Chengqian's "strange
point
Tang
records. But can we
at their face value?
by Prince
out its connections with the Li clan's
Chengqian's
family history.56
what we think the most relevant observation on this case: he
neurosis" "atavistic
We shall demonstrate that the whole affair was
hardly
predilections
for the Turkish life
a "scandal"; neither did Prince
Chengqian
have a
"strange neurosis"; nor were the
predilections
"atavistic". We contend
that Prince Chengqian
was not much different from a
typical Li boy of the time, just
as the
young
Li Shimin once was, and the so-called scandal was no more than the norm of the Li's
family life.
Let us look at Li Shimin the father himself. By many stories including
his own
admission, Li Shimin was as "wild" while young
as Prince
Chengqian
was
reported
to have
behaved: a
passion for archery and horsemanship (which continued all through his life and
were exercised with his largely ethnic, Turkic in particular, imperial guards) but little
knowledge
seasoned with
As for
of the classics.58 Yet the same acts of his son Chengqian
rejecting
exaggerated
the "Chinese
or made-up
identity"
were
depicted,
"deviate" details, as scandalous offences.
and
taking
on "Turkish"
things,
no doubt
we have
demonstrated elsewhere that Li Shimin himself had a well-established "Turkic" or "tribal"
identity
and was conversant in Turkic.59 In particular,
Li Shimin, while still a prince,
became a "sworn brother" of several prominent
Turk
persons.60
This
identity
was even
reflected on his death and his mausoleum.61 If anything, Prince Chengqian could only be
accused of
being
too filial a son of the many aspects of the clan's life his father found best
absent from the official history. The suppression of Li Shimin's childhood name in all
55 Denis Crispin Twitchett (ed.), The Cambridge History of China (London, 1979), iii, p. 236.
56 Arthur Wright, "T'ang T'ai-tsung and Buddhism", in Perspectives on the T'ang, ed. A. F. Wright and D.
Twitchett (New Haven, 1973), pp. 239?63. The particular
57
Edwin G. Pulleyblank, "The An Lu-shan rebellion
comment is on
p. 253.
and the origins of chronic militarism in late T'ang
China" in Essays
on
T'ang Society, ed. J. C. Perry and B. L. Smith (Leiden, 1976), pp. 32-60. See p. 37.
58
See, e.g., ZZhTJ 192.6034, Quan Tang
wen ^Jftfc (Beijing, 1983) 9.106, THY 26.506, Da Tang xinyu
9.133, etc. For continuing his old habit while an emperor, see ZZhTJ 192.6021-22, 192.6042.
59 "A-gan revisited".
60
They include Tuli 3*3l| (Tolis) Qaghan
of the Eastern Turk (ZZhTJ 191.5992), the loyal Turk general
Ashina Simo
RfJrlfiPSS (Cefu yuangui JU^JtuUL [Beijing* J96o], 980.11516) who was
given
the imperial clan
name Li (JTSh 194.5156, XTSh 215.6037), and the Western Turk prince, later Duolu Bj|j|?* Qaghan (JTSh
I94-5I83)
61
Cen Zhongmian QWtffa, Sui Tang
shi Rf^lfi (Beijing, 1957), pp. 140 and 142, has claimed that certain
features of Tang Taizong's mausoleum were imitative of Turkish burial custom, a notion followed by Howard J.
Wechsler, Offerings of Jade and Silk (New Haven, 1985), p. 81. We contend that, rather than mere imitation, it in
fact reflected the Tang imperial house's steppe background. The Tang's respect for non-Han burial customs both
within and without the Chinese heartland was reflected in Tang Taizong's condemnation of the Eastern Turks'
adoption of the (Han) tomb burials in violation of their ancestral customs
(Cefu yuangui, 125.1501; see also Ilkido
Ecsedy, "Ancient Turk [T'u-chiiehl burial customs", AOH, XXXVIII [1984], pp. 263-87; the particular passages
are found on pages 276-7) and Tang Xuanzong's edict allowing
a surrendered Turk official to "be buried
according
to the native [Turkish] customs" (Cefu yuangui, 974.11446).

Succession Struggle and the Ethnic Identity of the Tang Imperial House 387
records is a good example. Chengqian's
"scandalous
relationship"
entertainer,62 a Turco-Xianbei trait as examined earlier, is another.
Here is a typical
that would have seemed
Emperor Taizong's
case of how modern authors tend to go
quite
natural after
recognizing
to great lengths
the Li's Turco-Xianbei
six famous horses whose reliefs were carved at the
imperial
with an actor
to explain things
identity.
mausoleum
(another Turco-Mongol trait), according to Edward Schafer in his famous study
of the
Tang exotica, "came to T'ai
Tsung
from the Turks"
partly
because one of them had the
name
Teqin Biao $??$&& "prince's roan".63 This far-fetched interpretation for the Turkic
word
tegin
is hardly necessary after
Tuoba Xianbei title,64 and
Emperor Taizong
secondly,
realizing,
was none other than a prince.65
first that the term was also a well-recognized
when that very horse was
being ridden, the future
Regarding
the convenient
allegation
that Prince
Chengqian
had "mental
problems",
it is
interesting
to notice that the original heir apparent by the dizhang jjjj||| principle of the
Qing flf Emperor Shengzu Hffl. (Kangxi Jfl#S) supposedly had the same disorder.66 One
cannot
help noticing the use or abuse of psychiatry
to
punish cultural or
political dissidents
in a more modern context.
Let us further remark that the incoming Emperor Gaozong jS?^ turned out no less
"scandalous" than his hapless elder brother Prince Chengqian in "imitating" the steppe
life-style by marrying
revealing
panions
was Gaozong's
in their inner
unbalanced" Prince
his father's concubine. Not as well-known but
probably
more
conscious decision to order that his own sons had Turk com
palace.67
Chengqian
speak the Turkic language.
reign
Finally,
was
it is no accident that Prince
of Xuanzong,68
At the very least,
certainly
Chengqian
not the only
was
we now know that the
the last Turco-Xianbei emperor of the Tang
"mentally
Li clan member who liked to
eventually "rehabilitated" in the
in our view, who,
incidentally and despite his ministers' opposition, also liked to have Turk companions
during
entourage
his outings. However, the presence of a Turkic
was a sure
sign
that the Turco-Xianbei
period
To summarize, there is little doubt that Prince Chengqian
who fell victim not
efforts
by
sinocentric views and
with the historical
ing
only
to the vicious succession
the Confucian historians and the
image
tendency
and
issue of political legitimacy
to extract several
general
legacy
language interpreter
was
struggle
gradually coming
was
just
but also the
imperial court, prompted by
and the latter's overt concern, sometimes even
in the Central
in Xuanzong's
to an end.69
another normal Li lad
joint propaganda
the former's
obsession,
it would leave behind, not to mention the command
Kingdom
observations about the succession
"sinicization" from this case and many of its precedents,
sections.
at the time. Moreover, we are able
struggles
to be discussed in the
and the process of
following
62
ZZhTJ 196.6191.
63
Edward H. Schafer, The Golden Peaches of Samarkand: a
Study of T'ang Exotics
(Berkeley, 1963), p. 65.
64
Peter A. Boodberg, "The language of the T'o-pa Wei", HJAS,
I (1936), pp. 167-85.
65
See Emperor Taizong's essay "Liumatu zan m
/\fiHU1t"> Quan Tang wen, 10.124.
66
Silas H. L. Wu, Passage to Power:
K'ang-hsi and His Heir
Apparent, 1661?1772 (Cambridge, Mass., 1979),
PP- 133-6
67 THY 55.943-44, XTSh 1995661, ZZhTJ 201.6363.
68 XTSh 80.3565.
69 THY 27.521.

388 Sanping Chen
But before that, let us
point
to a case
bearing
a
striking
resemblance to Prince
Chengqian's alleged plots of revolt: the case of the Northern Wei's Crown Prince Tuoba
Xun .70 However
35HSLjfy
there was a
key difference: at that time the Northern Wei was not
the
only
and
all-dominating regime
in China. It was thus unable to
gloss
the incident over
as
perhaps
another
"family
scandal" caused
by
a
"mentally
unbalanced" crown
prince.
In
short, it did not enjoy,
as its Sui and Tang
successors did, a
monopoly
on
historiography:
the incident was
clearly recorded in Nan Qi shu *$0Wr"? the history of the rivalry of the
Southern Qi Jf,
as a backlash to the Tuoba
Emperor
Xiao wen's wholesale sinicization
drive,71 which
according
to
many people
was motivated
by
the latter's ambition to end the
North-South partition,72 a deed that was eventually accomplished by the Sui and Tang. In
addition to
preparing
us for
addressing
the common threads in these cases, it may also
help
explain why the two Tuoba political
as well as
biological heirs were so anxious to claim
their
alleged
Chinese ancestry.
The Succession to the Throne in the Early Tang
The case of Prince Chengqian was hardly an aberration in the imperial family, as the official
records would have us believe. As shall be demonstrated, one of many characteristics that
set the Tang apart from other "native" Chinese
struggle
around the succession to the throne. During
century and a half of the dynasty,
dynasties
a period
was the persistent political
that covered
nearly
not once did the succession follow the time-honoured
the first
Han
dizhang primogeniture principle.
Nor was there a
single
time when the process
ran
automatically without strife (and indeed bloodshed). Even after that, as observed by Chen
Yinke, the position of the official heir apparent (by the dizhang rule or not) was anything
but firmly established.73
was
It should be pointed
strikingly
categorization.
out that exactly
akin to the Qing regime,
The latter, not burdened
on this point
the
supposedly
a typical "conquest" dynasty
by
"native" Tang dynasty
in the conventional
the need to maintain a Han Chinese facade,
simply (after a period of failed experimentation) abolished the institution of an heir
apparent altogether.
This
imperial
similarity points
to the root of the incessant succession
house's Turco-Xianbei steppe origin
and heritage.
struggles
The magnitude
of the Tang:
the
and duration of
this
quite unique (among
the so-called "native"
dynasties)
characteristic were far too
great
to be simply ascribed to some
lingering Northern influences. It is the reflection of the true
ethnic/cultural identity of an imperial house that had spared no efforts in presenting itself as
a bona
fide Han regime in all historical records.
One cannot
Southern
help noticing
the numerous succession
struggles
in the Western
Jin
and the
dynasties. However, in addition to the northern influence74 and often direct
involvement, exemplified by the case of Liu Yuan $)$i|, the founding emperor of the
70
Wei shu, 22.558.
71
Xiao Zixian
jjf^Jjf, Nan Qi shu (Beijing, 1972), 57.996.
72
See for example Chen Qiaoyi
Li
g?0tf?' Daoyuan yu Shujing zhu M^LtcMt^J^H (Shanghai, 1987),
pp. 31, 34-5.
73 Chen Yinke, Tangdai zhengzhishi, pp. 52-3.
74 Chen Yuan, op. cit., p. 252.

Succession Struggle and the Ethnic Identity of the Tang Imperial House 389
Xiongnu Former Zhao Bftlfi dynasty,75 these incidents all fall into the general pattern of
every
new "native"
dynasty that, largely
as the
ripple
effect of the
preceding military
campaigns and difficulties in establishing
a new
imperial order, the first and sometimes the
second succession were
mentioned here survived much
native house
during
trouble. Furthermore, the
invariably problematic.
beyond
the period, namely
external involvement, while the Tang
Turco-Xianbei courts),
within the
imperial
clan.
struggles during
cases
In essence, none of these "native"
dynasties
this "bottleneck"; whereas the only long-lived
the Eastern
Jin, was
the Southern
remarkably
dynasties
(and many of their precedents
as observed by Chen Yinke, were all confined to the
free of similar
all had extensive
in the Northern
The
disruptive
and often
bloody
succession process of the
Turco-Mongol regimes
on the
steppe characterized by fratricide and other forms of bloodshed within the ruling house has
long been noted. However, the most methodical treatment of the
subject appears
to be that
of the late Joseph Fletcher. He used the term
"tanistry" for the general principle
on the
steppe that the leadership of a tribe or other polity should be passed on to the best qualified
member of the
usually
chiefly
or
ruling
capital
house.76 And the eventual choice of a new leader was
the result of some form of contention. In a fine article Fletcher discussed in detail
how this Turco-Mongol
Peter Boodberg
was
tradition manifested itself in the Ottoman
Empire.77
among the first to refer to the Turco-Mongol traditions, including
the not-unusual father-son
enmity
in
particular,
in
discussing
the Sui
imperial succession.78
In an
unpublished presentation
on
tanistry in Ottoman, Indian and later Chinese (Jin &,
Yuan t?, Ming B| and Qing dynasties) regimes, Fletcher also mentioned briefly the Sui
case and the case of Li Shimin.79 Nonetheless, as we shall demonstrate, this tradition in fact
left much more than what Fletcher had called "traces" in the history
would also reveal some
process of "sinicization".
Let us briefly
the Tang.
The
uninterrupted
review the
interesting phenomena
long history
of succession
in what had been
struggles
stretch started with Li Shimin. The future
of the Tang,
traditionally
which
area
called the
in the first two centuries of
Emperor Taizong's
meticulously planned
and
impeccably
executed manoeuvre to
replace
his elder brother
Jiancheng the heir apparent culminated in the famous Xuanwu Gate "SfJ^P? coup d'etat in
626. We shall later discuss other aspects reflected in this blatant
honoured
dizhang principle.
historian Zhao Yi $$? that this murderous fratricide
male
offspring
challenge
But not as well-known was the observation
of the two slain brothers almost
developed
extending
by
to the time
the Qing
to the execution of all the
into a case of patricide
as well,80
as is
generally
believed
regarding
the Sui
Emperor Yangdi's
accession to the throne.
A very similar drama was to be staged in the reign of Taizong by his sons, though with a
75
fin shu
(Beijing,
76
Joseph Fletcher,
77
Joseph Fletcher,
1974), 101.2648 and passim.
"The
Mongols: ecological and social
perspectives", HfAS, XLVI (1986), pp. 11-50.
"Turco-Mongolian monarchic tradition in the Ottoman Empire", Harvard Ukrainian
Studies, III/IV (1979-80),
78
Peter A.
Boodberg,
pp. 236-51.
"Marginalia
to the histories of the Northern dynasties", HfAS, IV (1939), pp. 230-83,
in
particular p. 266.
79
Joseph Fletcher, "Blood tanistry: authority and succession in the Ottoman, Indian Muslim, and later
Chinese empires", The Conference on the Theory of Democracy and Popular Participation, Bellagio, Italy, 1978.
80
Zhao Yi, Nian'ershi Zhaji iJ'.ZljfcrtJB
annotated by Du Weiyun $$??$. (Taipei, 1975), 19.410.

390 Sanping Chen
somewhat different
ending
due to
changed
circumstances and
personalities,
whose
impli
cations will be discussed later. It is worth noting that in the aftermath of the alleged
patricide/fratricide scheme of 643 plotted by Taizong's original heir apparent Prince
Chengqian,
the choice of the new crown
prince
was
partially prompted by
the need to
avoid future fratricide (which of course did not prevent it from happening
at all after the
death of Taizong).81 Another episode was the promise by one of the players Li Tai ^^
(Prince Wei ft) to kill off his son if he was appointed heir apparent so that after his death
the throne would go laterally
to another son of
Taizong's.
This
promise
did not sound so
unnatural to the ears of Taizong (and indeed as we shall see would be redeemed en masse
by
a great-grandson
Into the
reign
of his).82
of the next
Zetian, the focus of the succession
emperor and with the emergence of a strong woman Wu
struggle
shifted somewhat from
brotherly
acute parent-child enmity. Empress Wu showed little hesitation in having
apparent, both her own sons, killed in 675 and 684 repectively.83
who contrasted her apparent lack of qualms with a mother's natural
the case that one of the princes
was
perhaps
contest to
two heirs
A few modern authors84
feelings
and argued
not born of Empress Wu did not seem to
recognize
the entrenched tradition of filicide in the Sui and
Tang houses, and
Empress
Wu's
merciless killing of her own
baby daughter and several grandchildren later in her life.85
What is more
interesting is this: after having deposed two puppet emperors, both her
own sons, and enthroned herself, Empress Wu had both put
court
supervision
that
effectively
for
under house arrest with strict
cut off all communications with the outside.86 In the
meantime, the empress realized that she would still have to pass the throne on to no one else
but her blood offspring in the end.87 This awkward situation bears a striking resemblance
to the Ottoman Court's
form but nonetheless
Kafes (cage) system.88
"violating
flesh and blood", would be institutionalized
beyond
the year 833.89
After the restoration
This measure, perhaps
the fundamental human relations
(705) of the Tang dynasty
by Emperor Xuanzong
in a less draconian
by incarcerating
one's
and enforced at least
under Wu's son Emperor Zhongzong
4*^,
we soon see another case of filicide in a succession
struggle:
the
killing
of
Zhong
zong's heir apparent Prince Chongjun Itlfi after the latter's failed coup attempt against his
81
ZZhTf 197-6197, 199.6280-81.
82
ZZhTf 197.6195.
83
ZZhTf 202.6377,203.6419.
84
For example,
see
Huji jjjjj^j,
Wu Zetian benzhuan
SSll^^'ff (Xi'an, 1986), pp. 60-1.
85
ZZhTf 199.6286-87, 204.6467, 207.6557.
86
fTSh 7.135, ZZhTf 204.6473, 205.6490.
87
ZZhTf 204.6474-75, 206.6526-27.
88
Anthony D. Alderson, The Structure of the Ottoman Dynasty (Oxford, 1956), pp. 32-6; Jack Goody (ed.),
Succession to
High Office (Cambridge, 1966), pp.
20-1. For a detailed description of this part of the Ottoman palace,
see Norman Mosley Penzer,
89
ZZhTf 244.7886. We
The Harem (London, 1936), pp.
may also note that the Ottoman
197-201.
Kafes
was
hardly unique in the Turkic political
sphere: the later Ghaznavids in Afghanistan and Northern India also
developed
a similar policy
after incessant
succession struggles among the Sultan's family members. See C. E. and Decay
Bosworth, The Later Ghaznavids: Splendour
(Edinburgh, 1977), p. 38. Corresponding
to the "birth control" measures in the Kafes,
a remarkable
episode was Xuanzong's repeated attempts in aborting
his consort's pregnancy for fear of further antagonizing
Princess Taiping (JTSh 52.2184). One notes that the incident was recorded due largely
to the fact that the
pregnancy Xuanzong (then the crown prince) tried to terminate produced in the end none other than the future
emperor Suzong.

Succession Struggle and the Ethnic Identity of the Tang Imperial House 391
father.90 But not long afterwards the emperor in 710 fell victim to
viricide/patricide
as his
empress and daughter Princess Anle $$l conspired to seize the imperial power for
themselves.91
In a
typical Turco-Mongol
manoeuvre after the death of the khan, and
against
the
age-old tradition of lineal succession in the Central Kingdom, Princess Taiping J?SF, sister
of the deceased emperor, and her nephew Prince Longji ISfeS (and as many have claimed,
the latter's
seemingly
indifferent father too) together plotted
a successful coup
to
put
Longji's father back on the throne. At this time, the principle of tanistry, namely to let the
ablest heir of all succeed, was so entrenched in the Tang house, prevailing
over the
dizhang
principle stipulating instead the eldest heir born of the chief consort, that Longji's elder
brother Chengqi J^S steadfastly refused to be appointed the heir apparent, to avoid an
almost certain
repeat of the Xuanwu Gate incident.92
Thus was created a most
interesting triangular power
structure between the
Emperor,
his full sister Princess Taiping, and the young but capable Crown Prince Longji. The
triumvirate showed deep rifts from the very beginning.93 Although Prince Longji, later
Emperor Xuanzong, apparently
made sure that not much was left in the
history
records
that would cast doubt upon his filial relationship with his emperor father, there was
ample
evidence showing that Emperor Ruizong ^f^r? did feel daunted by this intellectually and
militarily gifted son, and the menace of patricide, which the Li boys (and girls) often
showed little reluctance in plotting
Ruizong's heavy dependence
his son's
power.94
and
executing,
on sister Princess
would seem a major
Taiping against
factor in Emperor
and as a way of balancing
No sooner had the son forced the aunt to commit suicide
preemptive military move than the father decided to go into genuine retirement for good.
In many aspects Emperor Xuanzong
could be called the last
Turco-Mongol
or Turco
Xianbei monarch of the Tang ruling house. Military and political expansion continued and
reached its zenith
during
his
reign.
The emperor also showed remarkable openness and
reception
to
foreign cultures, music and dance in
particular,
which was
eagerly
and
wholeheartedly imitated by
an
equally receptive populace, with various Hu tf\ fashions and
styles
soon
becoming
the
dominating vogue of the time.95
Xuanzong's dependence
on and
trust in his non-Han ethnic
generals
were also
unprecedented (which
would
eventually
cause his fall from grace). It is indeed very tempting
to compare the first half of his reign
with that of the brilliant Manchu Emperor Shengzu (Kangxi), the best emperor the Central
Kingdom
article.
ever had
according
to many. But this would be a little beyond
by
the scope of this
At issue here is this: just like Emperor Kangxi,96 Xuanzong was dogged by the problem
of succession and the choice of heir apparent. What is most remarkable and revealing of his
Turco-Xianbei heritage is that in the spring of 737 the emperor had three of his sons
including his first heir apparent killed in a single day.97 Such a resolute act of filicide, later
90
ZZhTJ 208.6611-12.
91
ZZh TJ 209.6641-42.
92
ZZhTJ 209.6650.
93
ZZ/i TJ 210.6656-57.
94
See for example the story in ZZhTJ 210.6673-74, fTSh 97.3051 and XTSh 125.4406.
95
JTSh 45.1957-58.
96
See Silas Wu, Passage to Power.
97
ZZhTJ 214.6829.
a

392 Sanping Chen
indignantly denounced as a "breach of heavenly principles" by the "native" Song dynasty
historian Fan Zuyu lEfiS ,98 would put Xuanzong
on an
equal footing with, say, the
Ottoman sultan Selim I (Selim the Grim), who was believed to have killed three of his four
sons,99 leaving
but one
The radical measure
(Suleyman
the Magnificent)
to succeed him.
(plus further acts of filicide100) did not relieve Xuanzong
in the end
from the now
by
almost trademark
Tang
succession trouble. In the wake of the An Lushan
5?3Slij Rebellion, another heir apparent son would come to the parting of the ways with
Xuanzong and declare himself the new emperor (Emperor Suzong jjf^), handing his
father the coup de grdce.101 Furthermore, after both had returned to the recovered Tang
capital, Xuanzong would find himself almost assassinated in a kidnap plot devised and
executed by his son's most trusted courtier.102 With all his loyal ministers, attendants and
eunuchs dead or exiled, this "retired emperor" would eventually die a bitter, lonely and
helpless
old man under virtual house arrest.
Such an unceremonious exit of in our view the last Turco-Xianbei monarch of the Tang
by
no means
spelt
the end of the
imperial
house's
Turco-Mongol-style
succession
prob
lems. Indeed even
prior to that Emperor Suzong had to overcome and kill another royal
brother, Prince Yong ^C (Li Lin ^3sfc) who, apparently with the encouragement of
Emperor Xuanzong,103 began
to harbour imperial ambitions too. A footnote is that this
episode of fratricide had profound effects on the life and career of the two greatest Tang
poets Li Bai ^S& (Li Po) and Du Fu tfcfS ,104 Despite
a further act of filicide which saw one
of his able sons Prince Jianning ??$? killed and the eldest one, the future Emperor Daizong
j&fc gravely endangered,105 Suzong
chaos of yet another fratricidal coup d'etat in 762.106
The Tang's
would die, of natural causes or otherwise, amid the
succession troubles and the "unstable institution" of heir apparent
from over after this. What is more, there would emerge much more
successions than the almost unbroken lineal tradition of the
were many gradual
but
important changes
struggles (some of which will be covered in later
this article that the early Tang
was a Sarbo-Chinese
characteristics, we feel our documentation of
frequent
were far
lateral
early Tang. However, there
in the aspects and circumstances of succession
tanistry
sections).
regime
As for the major
contention of
with strong Turco-Mongol
cases could stop
at this point.
following sections will be devoted to various aspects and implications of the Tang
succession
struggles.
98
Fan Zuyu, Tangjian 05 (Shanghai, 1937), 9.78.
99
On November 20, 1514. See Alderson, op. cit., p. 30.
100
XTSh 82.3608, ZZhTf 216.6916.17.
101
ZZ/iTJ 218.6975-76, 218.6982.
102
ZZhTf 221.7094-9S
103
ZZ/iT/ 218.6983, 219.7007.
104
Li Bai signed
on with the losing side of this fratricidal struggle for the throne, for which he almost received
a death sentence. He was
eventually exiled to Yelang |$$p,
which was
by far the most serious crisis in Li's life. See
for example XTSh 202.5763. This in turn led Du Fu to write several immortal poems on his deep concern over his
friend's fate.
105 ZZh Tf 219.7013.
106 ZZh Tf 222.7123-24.
The

Succession Struggle and the Ethnic Identity of the Tang Imperial House 393
The "Oedipus Complex"
Arthur Wright first proposed that the Sui Emperor Yangdi represented an instance of the
Oedipus complex.107 Recently
Victor
our view, the "Oedipus complex" depiction
analysis.
Xiong repeated
and in fact reinforced the case.108 In
here has overstretched the
It is obvious from our narration that acute father-son
characteristic of the first two centuries of the Tang.
trace the phenomenon
back to most earlier Northern
enmity
sphere
and
of psycho
rivalry
In addition to the Sui, one can
dynasties
established
groups, as well as in the An Lushan-Shi Shiming jfeSW rebel regime. Boodberg
to the point:
it was a much more
general
and wider
phenomenon
by
on the steppe.
were
easily
various tribal
was more
Wright's Oedipus complex diagnosis
would make this father-son
adversary
a rather
personal trait of Emperor Yangdi. Even if the label would stick, then it has to be extended
to so many other political figures of steppe origin
or
background, both in China and
elsewhere, that any utility
the term had would appear lost. This father-son
enmity
was so
typical of Turco-Mongol culture that the vendetta against
a dead father or the rush to
reverse his policies
could be
clearly
shown in a wide
variety
of cases, ranging
from the
Ghaznavid sultan Masud in the Turco-Iranian world,109 to late Manchu monarchs like
Emperor Gaozong Jft^ (Qianlong $?|Sfe) of the Qing.110 But if one invokes the Oedipal
drive on the pretext
of its being
a universal human trait, then in addition to
specificity,
hence the very
reason for
introducing
the term, one would face the
daunting
task of explaining the relative scarcity of similar cases in the two Han SI dynasties,
as well
as in other stable and prolonged
It is our view that the so-called
bypassing
"native"
the real socio-economic and
dynasties.
Oedipus complex
politico-cultural
losing
the
is in effect a convenient way of
issues. We note that even the
Shakespearean play Hamlet which has often been cited as a case of such a complex may in
fact have actual historical succession
struggles
As many modern researchers have concluded, contrary
pastoral
in the background.111
nomadism was a very complex way of life which
and great effort. In contrast to agriculture,
in decision-making
spring
storm could
way of life, the relatively
to conventional
misconceptions,
required sophisticated planning
it in fact allowed a much smaller margin
of error
and a very low tolerance of natural and man-made disasters, as a severe
easily wipe
out the entire stock in a matter of days. The
low economic return on a per-acreage-of-land
constant threat of tribal and other warfare meant that tribe and other
be very personal
and militaristic, without the
luxury
of a
highly
mobile
basis as well as the
polity leadership must
large, permanent civil bureau
cratic establishment. All these factors demanded that the
leadership
remained with young,
industrious and energetic figures who could respond quickly and forcefully to a
change of
107
Arthur Wright, "Sui
Yang-ti: personality and stereotype", in
Confucianism
and Chinese Civilization, ed. A.
Wright (New York,
108
victor Cunrui
1964), pp. 158-87.
Xiong, "Sui Yangdi and the building of Sui-Tang Luoyang", The Journal of Asian Studies,
LII (1993), pp. 66-89.
109
Bosworth, The Ghaznavids, pp. 230-4.
110
FengErkang^flfJjfc, Yongzheng zhuan
MlE&(Beijing, 1985), pp. 555-9.
111
Lilian Winstanley, Hamlet and the Scottish Succession
(Cambridge, 1921) and, much more
recently, Stuart M.
Kurland, "Hamlet and the Scottish succession?" Studies in English Literature, 1500?1900, XXXIV (1994),
pp. 279-300.

394 Sanping Chen
or emergency in the environment, and who would be able to lead the tribes into successful
military campaigns.
This leads to the notion that
socio-political
life on the
steppe required
a faster pace
or
metabolism in
replenishing its leadership than in a
sedentary agricultural society which,
with a
large
civil administration machine
feeding
on immense and reliable revenues from
intensive farming, could afford to allow irresponsive and politically uninterested elderly
monarchs to continue to
occupy the throne
(incidentally
this was
exactly
what had become
of Emperor Xuanzong, the last Turco-Xianbei monarch of the Tang). This in our view
was the real root of the
semi-legendary steppe
tradition found as far
away
as in the Khazar
empire, that a khan had a
preordained time limit on his reign,112 after which he would be
murdered.113
We see also how Turco-Mongol traditions adapted
or shall we say mellowed in their
transition to the Central Kingdom. Unlike the harsh political realities on the steppe where a
khan
normally
would have to hang
a sedentary society opened up
albeit as our cases displayed
steppe
alternative.
Following
decessors, the Tang, especially
on until his natural or not so
a less draconian
option:
the life as such would seem not
the position
greatly
infrequently
violent death,
of a retired emperor,
more enviable than the
the Turco-Xianbei traditions of its Northern-dynasty pre
in its first century and half, had numerous such retired
emperors, whereas none was found in more than four centuries of the two Han
This is one more reason why
regime.
the early Tang
We may further observe that in subsequent major
should be more
"native"
aptly
called
dynasties, namely
dynasties.
a Sarbo-Chinese
the Song
and the
Ming, only
a handful more such cases were found, almost
invariably consequences
of catastrophic military invasion from the north.114 Only in the Qing dynasty did we find
another case of purely domestic nature, which curiously enough was allegedly prompted
by
namely
the "filial wish" of Emperor Gaozong
the
length
of the reign
that his
reign
of his grandfather Emperor Shengzu.115
did not exceed a pre-set limit,
Having
examined the father-son feature, the maternal
aspect
in the
alleged Oedipus
complex should also be reviewed to do it full justice. First we remark that once the mother
changed
herself into a "surrogate"
father in the sense that she became a contender for the
throne, as in the case of Empress Wu Zetian, the mother-son relation would turn out to be
much less romantic than the
complex prescribes,
one more
proof
of the
phenomenon's
politico-cultural rather than the alleged psychosexual root. Also the supposed affection
was
evidently not
always reciprocal
as
political filicide by the mother was noted from the
late Tuoba Wei period onward.
Nevertheless, in comparison with the fate of the fathers, royal matriarchs did seem to
fare much better, even in extreme cases like Empress Wu Zetian
(and
the grandmother
112 Linghu Defen, Zhou shu (Beijing, 1971) 50.909; also Li Yanshou ^Ji?|?, Bei shi :jtj& (Beijing, 1974),
99.3287, and Du You ttl&, Tong dian jft^ (Shanghai, 1935), 197.1068.
113 Sir James G. Frazer, "The killing of the Khazar kings", Folklore, XXVIII (1917), PP- 382-407;
D. M.
Dunlop, The History of the fewish Khazars (Princeton, 1954), p. 97. An extensive exposition of ritual regicide can
be found in Sir James G. Frazer, The Golden Bough, 3rd ed. (New York, 1935), iv, pp. 9?119.
114
A comprehensive summary of most such "retired emperors"
can be found in Zhao Yi, op. cit., 13.281-6.
115
See for instance Arthur William Hummel (ed.), Eminent Chinese of the Ch'ing period (1644?IQ12)
(Washington, D.C., 1943-4), p. 372.
of

Succession Struggle and the Ethnic Identity of the Tang Imperial House 395
Emperor Xiao wen of the Tuoba Wei116). However,
too universal trait of Oedipal drive,
we think a more
instead of the either too individual or
plausible explanation
lies elsewhere. In
this
regard
classical historians would seem more down-to-earth than
psychoanalysis
learning modern authors when Hou Han shu $c9I4F provided the following hint in
describing the Wuhuan J?^|| 's unmistakable steppe traits of patricide
[They]
and fratricide:
kill their father and elder brothers when angry, but would never hurt their mother, for
mothers have a clan whereas fathers and elder brothers do not
themselves.117
Here
we
clearly
factor on the
steppe
see remnants of matriarchy
than one's subconscious
[have
someone to take] revenge for
at work, which in our view is a more realistic
Oedipal
desire for one's mother, for this can
also
explain
the Northern women's
traditionally strong role, attested
by
their
prominent
social status in the Northern dynasties,118 and culminating in Empress Wu
becoming the
one and only woman emperor in Chinese history, the latter of which could only happen in
a Sarbo-Chinese
dynasty
with
deeply-entrenched Turco-Mongol
traditions. A
mother's near-sacred status and
complete
dominance of her son in our view also
provide
a
most natural explanation for the unprecedented Tuoba Wei custom of killing
an heir
apparent's
birth mother before
crowning
the prince
son.119 The case of Empress-dowager
Hu120 demonstrates that this draconian rule was not an unwarranted and overcautious
precaution.
Other Aspects of Tang Tanistry
The established Han tradition of
having,
at least
formally,
an heir
apparent
means that
succession struggles in the Central Kingdom usually happened prior to the death of the
current ruler and around the institution of heir
apparent. This may be the most
important
departure from the steppe where the hell of succession wars normally broke loose with the
death of the khan. But this feature or "adaption"
China, as it was also attested in other
to the
"conquest" regimes,
sedentary society was not unique
the Ottoman
Empire
particular.121
In his pioneering treatise on the political history of the Tang, providing the first
systematic examination of the phenomenon
of the unstable
Tang
steppe
institution of heir
apparent (to which this author owes much of the
inspiration
for the current
work), the late
Tang studies authority Chen Yinke has analyzed the importance of controlling the
Xuanwu Gate in the
capital
in numerous
coups d'etat during
the
early Tang.122 However,
we consider the discussion of the key factions of people instead of geography in these
incidents a more
consequential topic. Not
only would one understand the reason for the
notoriety of the Xuanwu Gate in the period, but it would also shed light on the process of
116 ZZhTf 134.4187, 137.4302.
117 Fan Ye $JR|L Hou Han shu (Beijing, 1965), 90.2979.
118 Yanshi jiaxun jijie, 1.48. Arthur Wright (The Sui Dynasty, p. 35) has also noted Yan's observation.
119
The case of Lady Gouyi &3~*&LK
of the Han, from which the Tuoba custom had allegedly taken its cue,
was
evidently
an
exceptional, ad hoc measure.
120
ZZhTf 152.4739.
121
See for
example Joseph Fletcher, "Turco-Mongol monarchic tradition in the Ottoman
122
Empire".
Chen Yinke, Tangdai zhengzhi shi, pp. 41-4.
in
in

how this Sarbo-Chinese
"native"
dynasty.
regime slowly
For the era under examination, namely
elements in the succession
struggle
the cause for the Xuanwu Gate's
396 Sanping Chen
was various
prominence
evolved into
something
the first 150 years of the Tang,
imperial guard
in these coups during
more in line with a
one of the key
units. This in our view was
the period.
For
tanistry
struggles
in an
agrarian society,
Fletcher has introduced the term
"surrogate
nomads" for
the equivalent of tribes and tribal military elites who would fight out the succession wars
on the steppe.
For the
early Tang,
the
"surrogate" qualifier
would seem almost
superfluous,
for the ranks of imperial guards were filled with people not just of nomadic origin but
literally fresh from the steppe. A good example was the three hundred Turk troops Prince
Jiancheng planned
were said to be so numerous that
doubt
mostly military. Many
to use to attack Li Shimin's residence.123 In fact at times these
they
were
filled half of the positions
actually
figures
at court,124 which were no
mentioned or even named in succession
struggles,125 hardly the best indication of a "native" dynasty, though the standard records
have
"barbarian"
undoubtedly suppressed
figures
in setting
or played
the course of Chinese
down as much as possible
history.
the role of these
Another
important
dimension of the unmistakable
Turco-Mongol
character of the
early
Tang was its continued military and political expansion in almost all directions. Again
a
great number of non-Han ethnic
generals
were used in this endeavour,126 a rather
Mongol
forces in the Qing's
advances one millennium earlier. As the
and naturally
unique
an even
case in Chinese
march into Central Asia, an exact
greater number of troops
history
imperial guards actively participated
if we exclude the
repeat of the Tang's
in these
campaigns,
their
prestige
at the court and their role in the succession
struggle
continued.
The ancient Orkhon inscriptions mentioned explicitly that the Turk troops had fought the
wars for the Tabgach/Tuoba emperors.127 But this complaint masked the other side of the
coin: these Turk
generals'
and soldiers'
important
role in choosing
would serve under, as well as the grossly out-of-proportion positions they
Tang
The
court.
expansion
of the Tang was
gradually
Arabs and the Tibetans. A new Uighur power
checked
by
was also
meantime, the unstoppable process of "sinicization" was
One of the
signs was the role of personal
the advent of two new
emerging
slowly
but
the very emperors they
filled at the
powers:
on the steppe.
steadily taking
slaves and eunuchs in the succession
the
In the
its toll.
struggle.128
Emperor Xuanzong
who started as such a
strong Turco-Xianbei monarch would see
himself gradually turn into a uninterested and disengaged emperor, of the kind the later
"native" Ming dynasty
would see many of. The appearance of the second Eastern Turk
empire has invariably been hailed as a conquered people casting off the Tang yoke. Few
123
XTSh 79. See also Howard J. Wechsler, Mirror to the Son of Heaven: Wei Cheng
at the Court of T'ang
Vai-tsung (New Haven, 1974), p- 71
124
THY 73.1311; ZZhTJ 1936098.
125
Li Duozuo 255^r$? (JTSh 109.3296-97) was a typical case.
126
See for
example Chen Yinke, "Lun Tangdai
zhi Fanjiang yu fubing |^/?f^^Sr7r$IW?:^
> Chen Yinke
xiansheng lunwenji, i, pp. 665^77, and Zhang Qun "&%&, "Tangdai Fanjiang chutan
^fr^B^FfiJ^", Journal of
Oriental Studies, XIX (1981), pp. 1-38.
127
See for example Talat Tekin, op. cit., p. 264.
128
For example, in
Emperor Xuanzong's military
move to eliminate Princess Taiping's supporters, which
culminated in the princess's suicide and Emperor Ruizong's
final "retirement", two
important players were Wang
Maozhong -F^fr
a
family slave of Korean descent, and Gao Lishi j(|>fjdr, a eunuch. See 106.3252.
ZZhTJ 210.6683, JTSh

Succession Struggle and the Ethnic Identity of the Tang Imperial House
397
realize that the
development might
be more
appropriately
viewed as the consequence of the
growing alienation felt by
a
(junior) partner in a Sarbo-Turco-Chinese joint
venture that
was
tilting
more and more toward
agrarian traditions.129 In short, it was not unlike the case
of Outer Mongolia gaining independence
Chinese
dynasty.
after the end of the Qing,
With the onset of the An Lushan Rebellion, the outward
expansion
a Manchu-Mongol
came to a sudden
stop. But the stubborn steppe heritage of succession struggles did not go right
political
it underwent
The
guards
key
arena with our last Turco-Xianbei monarch. However, the fundamental aspects
in deciding
the end of the glories
a sea change.
transformation was that of the court eunuchs
the outcome of the succession
of the expansion
wars. The new
struggle,130
replacing
power-brokers
a natural
the
off the
regular imperial
development
with
of
would continue this
role until almost the very end of the
Tang.
Fletcher's term
"surrogate
nomads" would now
do full justice
to them. The coming of the eunuchal power at the Tang
court
closely
resembled, for example, the development in the declining years of the Safavid dynasty of
Iran after the latter's
practice
of
appointing
the
royal princes
to
provincial governorships
(also an early Tang policy) was
replaced by their confinement in the harem (similar
measures were
adopted by
the
Tang too, as discussed
earlier)
in order to avoid succession
contentions.131 With the control of imperial guards gradually falling into the hands of the
court eunuchs and the guard ranks filled mostly by rich playboys from the capital,132 the
political
importance
drama was now
played
of the Xuanwu Gate, together
was soon lost. However, these new
article and will be pursued
in separate
out within the walls of the inner
changes
studies.
with the prestige
are mostly beyond
The Process of Sinicization
palace.
The case of Prince Chengqian had many historical precedents and parallels.
Tuoba Xun, the Northern Wei heir apparent,
Sui heir
Chengqian's
succession
classical Chinese
apparent Yang Yong
uncle. Careful
struggles
study
and Taizong's
of the era: the "bandwagon"
scholarship
two other
prominent
The
strategic
of the many guard units,133
elder brother Crown Prince
of these cases reveals an
effective tool in such contentions when the regime was hard
to rule the entire Central
Kingdom.
In the case of Emperor Yangdi
the scope of the current
In addition to
cases were that of the
interesting pattern
Jiancheng,
to the
of sinicization and the patronage of
or other "authentic" Chinese literature and arts as a most
pressed
of the Sui, we see his marriage
for political legitimacy
at a young age
129
Pulleyblank ("The An-Lu-shan rebellion", p. 40) seems the only author to have noted this Turco-Chinese
partnership, albeit failing
to
recognize the crucial Xianbei factor which, like the Manchus, was the key element
binding the steppe and the
agrarian communities into a true
empire.
130
See Chen Yinke, Tangdai zhengzhishi, pp. 50-1.
131
Peter Jackson and Laurence Lockhart (ed.), The Cambridge History of Iran VI
(Cambridge, 1986), p. 366;
Laurence
29-32.
Lockhart, The Fall of the Safavi Dynasty and the Afghan occupation of Persia
(Cambridge, 1958), pp. 17,
132
ZZhTJ 254.8237. -,
133
For example, Zhao Lin
*?^,
Yinhua lu @tSi&, in Tang Guoshi bu deng bazhong jMJ&ftf 4?AA.?
ed.
Yang Jialuo $ft?g(DR(Taipei, 1962), 5.35, has a
telling story on the deplorable condition during the Yuanhe TCW
(806-820) era of an
originally prestigious imperial guard office.
to a

398 Sanping Chen
daughter
of a
prominent southern
royal family (in fact the
daughter
of a
puppet Later
Liang $&$. emperor), his patronage of the Southern Buddhist churches, his love of almost
everything connected to the south which before the final unification was seen even by
many in the north as the site of
himself being
an
extraordinarily
the
"legitimate" Chinese
talented man of Chinese
regime,
and last but not least
literacy134 - all of this certainly
had figured in his successful contention for the throne against his elder brother and the
Confucian
dizhang
rule of succession.
Li Shimin's bid for the throne is a very old and thoroughly studied topic. Many factors
contributing to his triumph over his elder brother have been proposed: his unmatched
military deeds in solidifying the dynasty, his command of a large group of talented
followers, his ability
to control the crucial Xuanwu Gate, and his preemptive strike, etc.135
But the issue of sinicization has not attracted
enough
attention in the context of his eventual
command of political legitimacy and its role in the defeat of his brothers. In fact, Li Shimin
set up an "academy" ($JM&) m his official residence to patronize classical literature and
scholarship
as
early
as 621 while the
regime
was
yet
to be solidified.136 One has to admit
that Li Shimin had remarkable political foresight and long-term planning in the struggle
for the succession.
The situation is perhaps best illuminated by the aforementioned quote of Emperor
Gaozu himself when Li Shimin began
to show his independence. It not only indicated the
Li clan's ethnic self identity
as discussed earlier, but also the notion that Li Shimin's political
aspirations were partly prompted by his sinicization - being "instigated by educated Han".
The case of Prince Chengqian
can be studied in the same context. The
hapless
Prince was
known to have
frequently "ignored
his
learning",
whereas his main rival Li Tai was a
diligent student and, following in the footsteps of his father, set up a
literary "academy" to
patronize
classical
scholarship,
In the
spring
of 642, he even
sponsorship.137
and his
presented
On the other hand, his
courtyard
was thus "as crowded as a marketplace".
the court with a major
reported
love of hunting
work authored under his
and war
games aside, there
were two
important
interesting
national
episodes
affair in
indicative
the Central
of Prince
Kingdom
Chengqian's
attitude towards the most
?
agriculture. First, he himself was
accused of having "interrupted farming".138 Secondly, he was credited with saving the life
of an attendant of the
imperial manor from Emperor Taizong's
order of execution.139 The
attendant's offence?
Ignoring
the maintenance of the
"agriculture garden".
It should also be
noted that to
judge by his surname Mu |g, this hapless attendant was
undoubtedly of
Xianbei descent.140
Proceeding
from the cases we have reviewed here, a general
rule in the succession
struggles
can be summarized as follows. The elder sons: Prince
Yong,
Prince
Jiangchen
and
Prince Chengqian (and the Tuoba Prince Xun in some sense) tended to be upstaged by their
more "cultured" and more "sinicized" younger brothers
Emperor Yangdi, Emperor
134
Read for example Wright, "Sui Yang-Ti", pp. 161-2.
135
See, e.g. Hu Rulei, op. cit., pp. 78-9.
136
ZZhTf 189.5931-2.
137
ZZhTf 195.6150, 196.6174.
138
ZZhTf 196.6168.
139
THY 4.44. The incident itself showed Chengqian's decent quality, something the official history
was
reluctant to admit.
140 Yao Weijuan tyffiot, Beichao huxing kao Jt|^iWJS# (Beijing, 1958), pp. 25-8.

Taizong
younger
Succession Struggle and the Ethnic Identity of the Tang Imperial House 399
and Prince Tai
respectively.141
The
paradox
sons had to overcome was the very Chinese
is that what these more "sinicized"
primogeniture
succession
principle.
Among other things, it created an acute dilemma for Han ministers and officials in taking
sides. The famous courtier Wei
Zheng
was a
good example.
This
deep
self-conflict could be
observed throughout Wei's life, a fact on which most biographies including Howard
Wechsler's fine
This
study
politico-cultural
failed to elaborate.142
aspect
of Turco-Mongol conquest regimes, namely
that of
enhancing their legitimacy by patronizing the "native" culture and religion at the expense
of their own ethnic
heritage,
was also
dominions in the Middle East, at least up
amply
demonstrated
by,
to the fourteenth century,
inter alia, almost all Turkic
as observed
by Frye
Sayili.143
Similar endeavour would be observed in the
Qing's
succession
struggles, exempli
fied by, say, the miserable end of both Chen Menglei Btt^Hf , the compiler of the
voluminous Gujin tushu jicheng ii^MttMlfc , and his immediate royal patron.144
However, if the "sinicization" or patronage of Han culture
dimension of the tanistry struggle,
then the military
dimension
represented
was
equally
the political
if not more
important. For the support of the imperial guards who would actually carry out the dirty
and often bloody job, military aptitude and valour on the part of the contender were also
crucial,
as attested
by
the cases of Sui Yangdi, Tang Taizong
and Xuanzong.
and
Failure in the
military aspect and/or an over-kill in "sinicization" would lead to disastrous results to the
contenders, best shown by the case of Prince Tai whose manoeuvre led to the downfall of
the Crown Prince Chengqian but fell short of Tai's ultimate objective of replacing him. As
one would note, similar effort in patronizing
traditional Chinese
scholarship by Emperor
Gaozong's
heir apparent
Li Xian ^K,
no doubt also to enhance his status in the face of his
mother Wu Zetian's blatant political challenge
to him and the entrenched Han patriarchy
tradition,
was
counterproductive
The strong opposition
to Li Tai
our view the existence of a political
Tai's overt
a complex
too.145
as a replacement
force at the Tang
"sinicizing" tendency. This, among other
and slow process and was not
that the choice of the new crown
political compromise
prince
of Prince
Chengqian
court that was not
things,
always going "one-way".
also suggests
sympathetic
shows that "sinicization" was
Few authors have noted
Li Zhi, later Emperor Gaozong, represented
in this ethnico-cultural context.146 This can be
clearly
in
to
seen from the
141
It is interesting
to note several cases of elder and more "militaristic" sons
being passed
over for succession in
the Turco-Iranian realm which have puzzled early Muslim and later authors (Bosworth, The Ghaznavids, pp. 45
and 228). Bosworth has suggested they might be a reminiscence of the steppe custom of
ultimogeniture.
However, as many Mongolists have pointed out, the Mongol practice of ochigin regarding the inheritance of
property might not necessarily apply
to
political power (khanship).
In the Sarbo-Chinese connection, the cultural
aspect of these cases would also seem an interesting topic. One notes contrast with the militarist elder brother Mas'ud, the younger
that, among the Ghaznavids for example,
in
brother Muhammad's tastes "were predominately
literary and studious" (Bosworth, Later Ghaznavids, p. 18).
142
Wei Zheng
was noted for his loyalty
to
Emperor Gaozu's original heir apparent Prince Jiancheng. As
shown by
an
interesting episode in JTSh 71.2559, he was still quite unwavering
on the
dizhang succession rule after
many years of service under Emperor Taizong, whose accession to the throne represented
a breach of this very
principle.
143
Richard N.
Frye and Aydin M. Sayili, "Turks in the Middle East before the
Saljuqs", JAOS, LXII (1943),
pp. 194-207.
144
Hummel, op. cit., p. 94.
145
Li Xian's famed annotation of Hou Han shu has since become an
integral part of that dynastic history.
146
Sun Guodong J^JgM^C, "Tang Zhenguan Yonghui jian dangzheng shishi jSjffV&&ftM3X9*&UW'', Xinya
Shuyuan xueshu niankan ffiffifrffi^Pfflffi^ 1, VII (1965), pp. 39?49, did point out that Zhangsun Wuji, Li Zhi's
a

400 Sanping Chen
fact that Li Zhi was close to Li Yuanchang ^5?fl, a follower of Prince Chengqian in the
Li clan, and Zhangsun Wuji was once a friend of Hou Junji 4$|ff ?|, another important
member of the Chengqian clique.147 Indeed Zhangsun Wuji carried the Tuoba Xianbei
legacy both by descent and in deeds. Instead of being stripped off by the "basically sinified"
former tribesmen,148 the "barbarian" felt hat he wore soon set the fashion, and was later
labelled by the Confucian historians as "devilish" as quoted earlier, another proof that
Prince
Chengqian's
"neurosis" was
hardly
an
anomaly
in the
ruling aristocracy. However,
the clearest indication of this being
a
compromise in the context of "sinicization" was Li
Zhi himself. In addition to ordering Turk companions for his sons as mentioned earlier, he
proved himself the ultimate northern boy by marrying his father's wife as well as allowing
her to dominate the court.
"Sinicization" (or patronage
The Issue of Legitimacy and the Role of Religion
of "native"cultures in the general Turco-Mongol political
sphere) as the political arm of the tanistry struggle leads to the issue of the political
legitimacy of the Turco-Xianbei regimes, the Sui and the (early) Tang included, in China.
In his resourceful and often stimulating book on China's frontiers, Thomas Barfield
contends that as a universal rule the steppe
and
taking
nomadic
regimes
over the Chinese heartland.149 However, this principle
were uninterested in settling
would seem at times
dependent
on
disavowing
the steppe identity
of the nomads as soon as
they
crossed to the
south of the Great Wall, for they would soon
develop
a strong interest afterwards in doing
exactly that. The Tuobas might have originated
in the Xing'an Hf?c Mountains, but the
strong Turkic elements, both linguistic and political,
in them and their successors were hard
to
ignore.
The Turco-Xianbei rulers in northern China
aspiring
certainly
did not show much hesitation in
to become a true son-of-heaven for all those under heaven. As was mentioned
earlier, the Tuoba Wei's wholesale sinification and move of its capital were viewed
many as
prompted by such an aim. It is remarkable that the Niizhen &Jk (Jurchen)
Emperor Wanyan Liang S^ilKft: also went through an extensive process of sinicization
prior
to his disastrous military expedition
to
unify
China.150 However, since the
collapse
of
the Western Jin, political and cultural "legitimacy" had always been regarded, by people
in
both the south and north, as residing with the Southern dynasties.151 The Sui/Tang regime
spared
no effort in overcoming
ours was to present
example,
exactly
various
themselves
post-Islam-Conquest
the same so as to achieve
this politico-cultural
as having
obstacle. One of their crucial endeav
been Han Chinese all along. We note that, for
Iranian
dynasties
political legitimacy.152
of native and Turkic
The appearance
origins
of the sinicization
decisive backer, was not a man of letters while three major supporters of Prince Wei, namely Liu Ji 8|^,
Cen
Wenben ^?3t$ and Cui Renshi ^tHZfifl all were, and all had a rather miserable end.
147
See ZZhTJ 197.6195 and Quan Tang wen 161.1645.
148
Paraphrasing the title of David Honey's essay on sinicization. See note 8.
149
Thomas J. Barfield, The Perilous Frontier (Cambridge, Mass., 1989), p. 197.
150
Tao, Jing-Shen, Jwrc/ien in Twelfth-Century China (Seattle, 1976), pp. 42-7.
151
Li Baiyao ^"S*SI,
Bei Qi shu (Beijing, 1972), 24.347-48.
152
Shahrokh Meskoob, Iranian Nationality and the Persian Language (Translation of Millilyat
va zaban) (Wash
ington, D.C., 1992), pp. 36-7.
in
by
did

"bandwagon"
Succession Struggle and the Ethnic Identity of the Tang Imperial House 401
in
tanistry struggles
extension of this conscious effort.
Few authors have
Zhou domain, just prior
paid
of the Tuoba's
Sui/Tang
attention to the acute ethnic strife, especially
to the Sui/Tang
unification of the country.
successors is thus a natural
An
in the Northern
important
was the cover-up and
fence-mending
efforts
by
these two
regimes.
A
good example
was
the sack by the Northern Zhou forces of the city of Jiangling ?E8? in 554, the temporary
capital of the southern Liang $& house and an established cultural centre. The brutality,
horror, and in particular
the
large
Hans, all social strata included, would certainly
mitted
obvious
by
the Manchus in conquering
reason
scale and indiscriminate enslavement of ethnic Southern
southern China
have made the atrocities
political reasons, not least the fact that the fathers and
Sui/Tang ruling clique
were active
participants
pale
in this most
in comparison.153
allegedly
grandfathers
com
But for
of the
savage feat, only sporadic
pieces
of evidence of the actual atrocity
were
preserved,
while deeds hard to
gloss
over like
the burning of the Liang royal library
were
conveniently blamed on the victims' own
acts.154
It is simply
unbelievable that memories of such atrocities would be
of a few decades when the Sui, followed
can understand the obsession these two
images
struggles.
of theirs, as well as the utility
For instance, Emperor Yangdi's
therefore in our view not caused
by
by
regimes
the Tang,
of the bandwagon
some
forgotten
in a matter
came to power. In this context we
had with the "politically
correct" ethnic
of "sinicization" in succession
famed aversion to the character hu #i155 was
arrogant sinocentricism as most authors have
alleged, but was dictated by the need for political legitimacy as perceived by Yangdi, and
the ethnic skeletons in the Yang's family's
closet.
It is also
interesting
to examine the role of
religion
in this context.
Many
authors
including Wright have noted that Emperor Yangdi's patronage of the southern Buddhist
schools was
politically
motivated. But few seem to have
recognized
the ethnic factor here:
the Sui was
evidently using Buddhism to
help bridge the ethnic divide, a feeling that must
have been very strong after the Jiangling atrocity whose
Yangdi's own grandfather Yang (Puliuru) Zhong li(1SF7\?n),&.
The
utility
of
religion
short-lived Sui, the early Tang emperors
that failed to prolong
political legitimacy,
their
major perpetrators included
became even more evident in the house of Tang.
predecessors'
the Tang imperial
After the
were no great patrons of this "foreign" religion
mandate of heaven. For the urgent need of
house found an even better solution than the
southern Buddhism schools to mask the clan's non-Han
origin:
to
identify
descendants of Li Er 2?If (Laozi ^rf-), the alleged founder of Taoism.
themselves
A striking parallel
can be found among the Safavids, the founder of the most
splendid
post-Turkmen dynasty in Iran that was
largely responsible for the now entrenched Shf 1
heritage in that country. With a questionable claim to native Iranian (or rather Kurdish)
153 Wang Zhongluo 3E'fW?> Wei-fin nanbeichao shi fJHFltttbttJtl (Shanghai, 1980), ii, p. 983, discusses
briefly the savageries of the Northern Zhou army based on some scattered data.
154 ZZhTf
165.5121.
155 See for example Ciyuan Kfffi, (Beijing, 1988), p. 82 quoting Shuofu R&JP.
as the

origin,156
the clan's
family language
402 Sanping Chen
was nevertheless Turkic Azeri.157 For obvious
politico
religious considerations, particularly an authentic Shn origin, the family falsified a gene
alogy
from one of the Twelver Imams, and
"systematically destroyed any evidence" that
would imply otherwise.158 In this regard the Mongols certainly had the fewest worldly
obstacles in portraying Chinggis Khan as the incarnation of a Buddhist universal emperor
to legitimize their rule of a world empire.159
This political dimension of religion has important bearings on the issue of tanistry
struggles, the case between Li Shimin and Crown Prince Jiangcheng in particular. Failure in
recognizing
this
political and, as we shall
insightful
see, ethnic aspect led to
Wright160
question
the
observation by Tang Yongtong $Jfflli? 161 that in this struggle the Buddhists
were on the side of the elder brother (Wright's erroneous notion has been all but refuted,
albeit implicitly, by Stanley Weinstein).162
As a "foreign" religion, the Buddhist establishment in China had a vested interest in the
imperial
house's
acknowledgement
of its non-Han
origin
and
heritage.
The church must
have actively countered any opposite move. In fact a famous monk Falin ffitt openly
"slandered" the
imperial family's
ethnic
origin, claiming
its descent from the Tuoba.163
Proceeding
from our examination of the role of "sinicization" in succession
struggles,
it
was
quite natural that the Buddhist church was behind the Crown Prince as
Tang has
observed. Meanwhile the Taoists rallied to the challenger Li Shimin, as shown by the cases
of two Taoist priests.164 It was
revealing that when Fang Xuanling ^fetfc and Du Ruhui
it#n#?, the two most trusted followers of Li Shimin, sneaked back to the capital
participate
Taoists.165
in the Xuanwu Gate coup d'etat, they reportedly disguised
Emperor Taizong's
edict of 637 elevating
Taoism to a higher
themselves
status than
Buddhism166 was a most natural reward for this support. Similarly,
this could also
explain
Empress Wu Zetian's patronage of "foreign" Buddhism in her bid to succeed her husband
and become an
unprecedented
woman
emperor against every Han Chinese custom and
principle.
156 Zeki Velidi Togan, "Sur l'origine des Safavides", in Melanges Louis Massignon (Damascus, 1957), iii,
pp. 345-57; Michael M. Mazzaoui, The Origins of the Safawids: Si ism, Sufism, and the

We have in this
Succession Struggle and the Ethnic Identity of the Tang Imperial House 403
characteristics of the period,
dynasty,
Concluding
study, through examining
was in fact a regime with heavy
remarks
the incessant succession
demonstrated that the early Tang,
far from
struggles
being
and other
a "native"
Turco-Xianbei traits, which hence may be more
aptly termed a Sarbo-Chinese regime. But in pursuit of political legitimacy, the imperial
family had made enormous efforts to present itself as a bona fide Han house, and to make
sure that no
compromising
evidence was left in any records.
Two historical factors have contributed to the
Tang's
near success in
maintaining
this
image throughout history: (a) it being the all-dominating polity in the vast East-Central
Asian continent and the sole custodian of historiography, with hardly any independent
cultural entity
of time.
in existence to provide
an alternative view or perspective,
and
(b)
the passing
Not
only
in exclusive control of traditional
historiography,
the Confucian scholar
officials all but monopolized every genre of writing in classical East Asia.167 This is
reflected in the fact that in all three major Altaic languages the early words for "writing"
and "books"are
effective
script,169
generally
let alone a body
considered cognates
was no rival to the tomes of Chinese literature (or "cultural
Ian Watt's
words170)
of the Chinese character bi %? ,168 Without
of existing literature, nomadic oral tradition, albeit rich,
accumulated since the archaic age.
repertoire"
in Jack Goody
On the other hand, what may have been even more critical were the deliberate efforts
the Northern autocratic families. For the
specific
Sui
objective
of
conquering
an
and
the South
where
"legitimate"
Chinese
dynasties
had been maintained, and for the
general
need for
political legitimacy all through
to the early Tang,
a
particular task of the Sui/Tang ruling
class was to mask and obscure their "non-Han" steppe background
by
and connections. A
particular
case in
point
is the contrast between Zhou shu MW, the official
Tang-authorized
history of the victorious Zhou regime in the struggle
to
unify Northern China, and that of
the loser, Bei Qi shu JfcJlFflF. Because the early Tang ruling clique
was dominated by the
descendants of the Northern Zhou
aristocracy,
the chief
compiler Linghu
Defen ^"5MI@t
himself being in fact the grandson of the Zhou Grand General Linghu Zheng ^*JCS ,171
Zhou shu could hardly afford any real impartiality
or
objectivity. This was noticed even by
the Tang historian Liu Zhiji J[|?n|S| .172 Meanwhile, the ruling families of the Qi fared quite
167
See Edwin Pulleyblank, "The Chinese and their neighbours
in
prehistoric
and early historic times", in The
Origins of Chinese Civilization, ed. David N.
Keichtley (Berkeley, 1983), pp. 411-66, in particular pp. 414?15.
168
Middle Chinese pronunciation
bit. For a
dissenting view, see Denis Sinor, "Altaica and Uralica", in Studies
in Finno-Ugric Linguistics
in Honor of
Alo Raun, ed. D. Sinor (Bloomington, 1978), pp. 319-32.
169
Long-speculated
existence of a Xianbei script (see, for example, Victor H. Mair, "Buddhism and the rise of
the written vernacular in East Asia: the making of national languages", J AS, LIII [1994], pp. 707?51) can be
largely disregarded after the discovery of the Xianbei cavern and even more because of other equally
rich
collections of archaeological findings, especially
tomb inscriptions and other artifacts of this period. See Zhou
Yiliang JH?&, Wei-Jin nanbeichao shi lunji xubian
9L^S^^t^$.^MH^ (Beijing, 1991), pp. 179-80.
170
Jack Goody and Ian Watt, "The consequence of literacy", Comparative Studies in Society and History,V
(1963), pp. 304-45.
171
Zhou shu 36.643.
172
Shitong $jj&,
edition used: Zhao Liifu
pp. 457, 960.
j|?.Bm, Shitong xin jiaozhu ^feiSSfR^ (Chongqing, 1990),

miserably
after
forfeiting
404 Sanping Chen
their state. Some former
royal household members even ended
up peddling candles,173 and the Northern Zhou aristocracy's distrust of the people of the
former Qi domain persisted well into Tang times.174 The narration in Bei Qiu shu was
therefore much less inhibited, if not deliberately negative, in describing the Qi regime. As
an
example
of issues
pertinent
to this
study,
one can
get significantly
more information on
the Turco-Xianbei cultural and political traits of the Qi ruling clan than of that of the
Zhou,175 whereas in
reality
the latter was much more "Xianbei-ized" than the former.176 In
view of the "politically correct" ethnic images in Zhou shu, one should not be surprised by
the even less
frequent
Northern traits in the
Tang
records.
On the second factor of time, after the death in exile of the hapless Buddhist monk Falin,
it was not until the Southern Song that the kind of political atmosphere first emerged in
which questions would be raised on this long-established historical image. By that time,
few solid records still remained to allow concrete and detailed examination of the real
origin
This
phy without
and characteristics of the Tang imperial
study
also tries to show the inherent
critically examining
the very
house.
problem
of
relying
source of these records:
on traditional
just
historiogra
like the Arabo
Persian authors who would not write about their ruler's Turkic culture and heritage,
without the attentive interference of a court obsessed with its historical
image,
Confucian
historians would still hardly have found the many Turco-Xianbei aspects of the regime
they were
serving a worthy topic. Repeated rewriting and editing of the histories for the
sake of elegance
even worse.
and concision, if not for some less honest purpose, had made the situation
We have also studied several aspects
relationship
with so-called "sinicization". We show that, contrary
that conquerors would soon melt in the sea of Chinese
steppe people
in the Central
Kingdom
was a long
of the Tang tanistry struggles, especially
populace,
and painful process. The
even
to the conventional view
the "sinicization" of
strong
reactions
to the Tuoba
Emperor
Xiaowen's forced wholesale sinicization, culminating
in the
Six-Garrison /\?& Rebellion and the downfall of the Northern Wei, could still cause
ripples
Another
in early Tang
seemingly
times.
entrenched notion we
challenge through
this
study
is the classifi
cation of Chinese
dynasties
into "native" and
"conquest" regimes.
Even in his otherwise
enlightening book, Barfield sticks to the conventional view that "the collapse of the
[Northern] Wei marked the end of Manchurian rule in China".177 Yet we have shown that
in many aspects the (early) Tang bore a
striking similarity to the Qing dynasty. In addition
to the succession
struggles,
the institution of heir
apparent,
the advance into Central Asia, and even the fate of the "national
the frontier and ethnic
language",
policy,
we may add,
173
ZZhTJ 173.5382.
174
This distrust might have even contributed to the cause of the An Lushan rebellion. See Edwin Pulleyblank,
The Background of the Rebellion of An Lu-shan (London,
175
See for example Albert E. Dien, "Yen Chih-T'ui
1955), pp. 75?81.
(531-91 +): His Life and Thought", Diss. (University of
California, 1962), p. 12, and Jennifer Holmgren, "Politics of the inner court under the Hou-chu (Last Lord) of the
Northern Ch'i (ca. 565-73)", in State and Society in Early Medieval China, ed. A. E. Dien (Stanford, 1990),
pp. 269?330, though both have used the standard "exemplary history" arguments
176
Wolfram Eberhard, A History of China, 4th
to
explain this contrast.
ed. (London, 1977), p. 152.
177
Barfield, op. cit., p. 126.
its

Succession Struggle and the Ethnic Identity of the Tang Imperial House 405
with details to be elaborated in separate studies, that the
Tang
and the
Qing
were the
only
two Chinese dynasties during which provincial governors held enormous
prestige and
power. At the very least, these facts suggest that the distinction between a "native" and a
"conquest" dynasty
is at best a grey notion.