Paper for the Conference on Middle Period China, 800–1400

“Before we all become barbarians”:

Han Yu’s rhetoric of barbarization and the (accidental) invention of “culturalism”

Shao-yun Yang
(University of California, Berkeley/Denison University)

Sixty years ago, Chen Yinke wrote an essay that sought to explicate Han Yu’s 韓愈
importance to Tang cultural history in terms of six aspects. Of these, two aspects related to the
reasons for Han Yu’s well-known opposition to Buddhism. In Chen’s view, one reason had to do
with the social and economic costs that Buddhist monasteries and Daoist temples purportedly
imposed on the state. The other had to do with the idea that Buddhism was a religion of foreign
and thus “barbarian” origin. It is the latter reason that will concern us here. Chen Yinke saw it as
the key to understanding not only Han Yu’s thinking but also the origins of the so-called Ancient
Style (Guwen 文) literary movement to which he was believed to belong. Referring to Han Yu
by his well-known style name Tuizhi 退之, Chen wrote:

What I would now like to argue is that the Tang Ancient Style movement was, in fact,
triggered by the rebellion of An Lushan and Shi Siming and the situation of autonomous
military provinces. An Lushan and Shi Siming were Western Hu (胡, i.e., Sogdians) of
mixed parentage, and moreover the [soldiers in the] military provinces were Hu (i.e.,
northern nomads or Sogdians) or Hu-acculturated (Huhua 胡 ) Han people…, so all the
most outstanding literati of the time consciously or unconsciously had in their minds the
image of the barbarians (Yi) of the four quarters invading the Zhou in distant antiquity
and the five kinds of Hu bringing disorder to the Chinese in more recent history. That is
why “respecting the king and repelling the barbarians” (zunwang rangyi 尊王攘夷) was
the central idea of the Ancient Style movement. Although the Ancient Style writers
slightly before Tuizhi—such as Xiao Yingshi, Li Hua, Dugu Ji, and Liang Su—as well as
other Ancient Style writers from Tuizhi’s generation—such as Liu Zongyuan, Liu Yuxi,
Yuan Zhen, and Bai Juyi—shared this subconscious thinking, they all suffered from an
insufficiently clear understanding and an insufficiently thorough position. Thus they
dared not and could not say or do what Tuizhi did in digging up Buddhism’s roots and
fervently and vehemently rejecting it because the Buddha was a barbarian (Yi-Di) and
 Buddhism a barbarian teaching. This is precisely the reason why Tuizhi became the
leader of the Ancient Style movement.1

As I shall show later in this paper, we have good reason to question Chen Yinke’s theory
that Han Yu and his peers saw the autonomy of the Hebei provinces as a barbarian invasion.
Nonetheless, Chen’s interpretation of Han Yu’s motivations came to be widely accepted and
quoted by historians of the Tang, some of whom (most notably Fu Lecheng2) went further and
interpreted the early ninth century as a time of increasing xenophobia among the entire Tang
elite, not just “the most outstanding literati of the time.” This interpretation remains the
mainstream in Chinese-language historiography today 3 , and was also dominant in Western-
language scholarship until relatively recently. In its simplest form, it takes the An Lushan
rebellion of 755–763 as a turning point that destroyed the “cosmopolitan” character of the early
Tang and replaced it with a sense of ethnocentric intolerance that extended to “foreign” religions
like Buddhism. Five years after Chen’s essay, for example, Arthur Wright made the following
claims in his classic account of the history of Chinese Buddhism:

The [An Lushan] rebellion and its aftermath weakened T’ang self-confidence, and the
cosmopolitanism of the early days of the dynasty gave way to a cultural defensiveness
that occasionally turned into xenophobia…. The old and oft-repeated attacks on
Buddhism now had a more receptive hearing than heretofore…. The upshot was the great
suppression of Buddhism between 842 and 845….

1
Chen Yinke 陳寅 , Jinmingguan conggao chubian 金明館叢 初編 (Beijing: Sanlian shudian, 2001), 329. This
essay, titled “Lun Han Yu” 論韓愈 (On Han Yu), was originally published in Lishi yanjiu 史研究 1954(2). A
similar argument can be found in Chen’s Yuan Bai shi jianzheng gao 元 詩箋證 (Beijing: Sanlian shudian, 2001
[1955]), 150.
2
See Fu, Han Tang shi lunji, 214–226, 361–367.
3
For recent examples, see Li Hongbin 李鴻 , Tangchao zhongyang jiquan yu minxu guanxi—yi beifang quyu wei
xiansuo 唐朝中央集權 民族關係里以 方 域爲綫索 (Beijing: Minzu chubanshe, 2003), 192–200 and Sui
Tang Wudai zhu wenti yanjiu 隋唐五代諸問題研究 (Beijing: Zhongyang minzu daxue chubanshe, 2006), 155–157.
Ge Zhaoguang points out that for ideological and political reasons, Chen Yinke’s interpretation of Han Yu came
under attack within China during the 1950s and 1960s. In the 1960s, even Zhang Shizhao—who was in many
respects a traditionalist like Chen Yinke—also issued a biting critique of Chen’s interpretation in his monumental
Liuwen zhiyao 柳文指要, accusing Chen of endorsing an irrationally xenophobic attitude toward Buddhism by
attributing such an attitude to Han Yu. Since Chen’s posthumous rehabilitation in the 1980s, however, the authority
of his theories has become nearly unshakeable in mainland Chinese academia. In Taiwan and Hong Kong, moreover,
Chen Yinke’s influence has consistently remained strong: the Taiwanese scholar Lo Lien-t’ien’s tactfully expressed
disagreement with Chen’s interpretation of the “Ancient Style movement” as anti-foreign appears to be a notable
exception. See Ge Zhaoguang 葛兆 , Zhongguo sixiangshi 中國思想史, vol. 2: Qi zhi shijiu shiji Zhongguo de
zhishi, sixiang yu xinyang 七至 九世紀中國的知識 思想 信仰 (Shanghai: Fudan daxue chubanshe, 2000),
223; Zhang Shizhao 章士釗, Liuwen zhiyao 柳文指要 (Beijing: Zhonghua shuju, 1971), 758–761; Lo Lien-t’ien 羅
聯添, Han Yu yanjiu 韓愈研究 (Tianjin: Tianjin jiaoyu chubanshe, 2012), 202–203. Note that Lo’s study was first
published in Taiwan in 1977 and revised in 1981.
As for Han Yu in particular, Wright argued:

In the years of the An Lu-shan catastrophe and its aftermath, men of learning and
conscience turned with a new seriousness to the Confucian canon…. These men, whose
gropings anticipate the full-scale Confucian revival, were generally not opposed to
Buddhism…. It was Han Yü—a brilliant polemicist and an ardent xenophobe—who
pulled together the criticisms made by his older contemporaries and laid down the
formula for cultural renaissance: Purge Chinese tradition of all the noxious accretions of
the years of Buddhist dominance; return directly to the immortal truths laid down by the
Chinese sages; rally all men of goodwill and build a new order on these truths.4

Similarly, Jacques Gernet’s influential Le Monde Chinois (1972) asserted the following:

A big change in direction in the intellectual life of China began around 800. In essentials
this change consisted of a deep desire on the part of some people to go back to the ancient
sources of the Chinese tradition combined with an attitude of hostility to the foreign
influences which had permeated China so widely since the end of the Han period. This
reaction, which followed a period when the court and the upper classes had been
particularly friendly to foreigners and to exotic fashions and products, seems to be largely
explained by the aspect of a national defeat assumed by the rebellion of An Lu-shan and
by the change of atmosphere which followed those tragic events.

The two examples that Gernet supplies as evidence for this anti-foreign reaction are, likewise,
the “‘ancient style’ movement” (particularly Han Yu, “a notorious anti-Buddhist”) and the
persecution of Buddhism and other foreign religions under Wuzong 武 (r. 840–846).5

For decades now, the xenophobic or ethnocentric turn that Chen, Wright, and Gernet
claimed to see in Han Yu and his contemporaries has figured in narratives of Tang history and
also influenced the study of late Tang literary culture.6 Nonetheless, it is no more than a chimera,

4
Arthur F. Wright, Buddhism in Chinese History (Stanford: Stanford University Press, 1959), 83, 87–88.
5
I have followed the 1982 English translation of Gernet’s work by J.R. Foster and Charles Hartman: Jacques Gernet,
A History of Chinese Civilization, Second Edition (Cambridge: Cambridge University Press, 1996), 291–292.
6
Stephen Owen’s groundbreaking study of the poetry of Han Yu and Meng Jiao 孟郊 (751–814) even began by
tracing the origins of the anti-foreign turn further back to the first half of the eighth century, a period usually seen as
the heyday of Tang “cosmopolitanism”: “In reaction against the cosmopolitan aspects of T’ang civilization, there
was a growing commitment on the part of many intellectuals of the period to the ‘pure’ Chinese tradition. Although
a mythical by-product of another figment of the modern historical imagination: namely, the
idealized and romanticized image of an exceptional golden age of “cosmopolitanism” under the
early Tang.7 Of the relatively small number of literary texts and incidents that have been used as
evidence that the Tang empire stopped welcoming foreigners and their cultures after 755, nearly
all should be reinterpreted as indications of the exact opposite. There were thousands of foreign
(mostly Arab and Persian) merchants for rapacious imperial troops or rebels to rob and kill in
Yangzhou 揚州 (in 760) and Guangzhou 廣州 (in 879) only because both the imperial court and
the local populace not only tolerated their presence but even welcomed and encouraged it. It was
the merchants’ great wealth, not their foreign origin, that made them particularly vulnerable to
violence when these cities were sacked.8 In 809–811, Yuan Zhen 元稹 (779–831), Bai Juyi 居
易 (772–846), and Chen Hong 陳鴻 (fl. 805–829) used works of poetry or fiction to satirize the
Chang’an elite’s love for “Western” (Hu 胡) music and fashions, but only because that proclivity
for exoticism was still prevalent in the capital in spite of the fact that the Tibetans had all but
severed Chang’an’s connections with Central Asia.9 In fact, our sources suggest that the number

their writings show pride in T’ang military power and international success, the intellectuals were just as often
disturbed by the incursion of foreign elements into popular culture, elements which they felt presented a threat to the
purity and continuity of their own tradition. This ambivalent feeling often emerged as an aggressive cultural
confidence expressed in the perennial theme of the golden age of antiquity, an ideal of humane civilization set in the
remote past.” See Stephen Owen, The Poetry of Meng Chiao and Han Yü (New Haven: Yale University Press, 1975),
1. A more recent version of this argument, tracing the anti-foreign turn to a “patriarchal” backlash against Wu
Zhao’s reign, can be found in N. Harry Rothschild, Wu Zhao: China’s Only Woman Emperor (New York: Pearson
Longman, 2008), 207.
7
An article by Charles Holcombe, the only sustained critique of the late Tang xenophobia theory to date, still
accepts the image of early Tang cosmopolitanism relatively uncritically. See Charles Holcombe, “Immigrants and
Strangers: From Cosmopolitanism to Confucian Universalism in Tang China,” T’ang Studies 20–21 (2002–03), 72–
73.
8
The Yangzhou massacre was perpetrated by imperial troops under Tian Shengong (d. 773), who were
campaigning against a rebel governor but seized an opportunity to plunder the region’s wealthiest commercial city.
Schafer incorrectly described them as “the hordes of the rebel T’ien Shen-kung,” and Gernet repeated this error by
calling them “insurgent bands.” Lee Chamney’s recent interpretation of this massacre as “a symbolic massacre of
the Other” via “the selective extermination of foreign merchants” rests on the flawed assumption that since the
relevant records mention only the killing of foreign merchants, this constitutes evidence that no Chinese civilians
were killed in the looting. Instead, the court historians’ emphasis on foreign victims in Yangzhou probably reflects
the fact that the wanton slaughter of Chinese civilians was too commonplace during the An Lushan Rebellion to be
worth mentioning. The Guangzhou massacre occurred after Huang Chao’s 黃巢 (835–884) rebel army captured the
city and is known only from Arabic sources; the reason for the Arab authors’ emphasis on non-Chinese victims is
even more obvious without the need to ascribe xenophobic motivations to Huang Chao and his followers. Jiu
Tangshu 唐書 (Zhonghua shuju edition [hereafter JTS]), 110.3313, 124.3533; Xin Tangshu 新唐書 (Zhonghua
shuju edition [hereafter XTS]), 141.4655, 144.4702; Edward H. Schafer, The Golden Peaches of Samarkand: A
Study of T’ang Exotics (Berkeley: University of California Press, 1963), 18; Gernet, A History of Chinese
Civilization, 292; Lee Chamney, “The An Shi Rebellion and Rejection of the Other in Tang China, 618–763” (M.A.
thesis, University of Alberta, 2012), 91–93; Hyunhee Park, Mapping the Chinese and Islamic Worlds: Cross-
Cultural Exchange in Pre-modern Asia (Cambridge: Cambridge University Press, 2012), 69–70.
9
Yuan Zhen and Chen Hong also allude disapprovingly to marriages between Chinese women and foreign men in
Chang’an, but the way in which they do so actually demonstrates that such marriages were common and widely
accepted after the An Lushan rebellion. For past interpretations of Yuan Zhen, Bai Juyi, and Chen Hong as
reflecting a general turn toward xenophobia, see Chen, Yuan Bai shi jianzheng gao, 148–150; Florence Hu-Sterk,
“Entre fascination et repulsion: Regards des poetes des Tang sur les ‘barbares,’” Monumenta Serica 48 (2000), 26–
of “Westerners” living in Chang’an increased after the An Lushan Rebellion precisely because
the Tibetan conquest of the Gansu Corridor left diplomatic embassies from the Western Regions
stranded in the Tang capital. These Western envoys established thriving émigré communities
subsidized by the Tang court; they were also joined by enterprising Sogdian traders who passed
themselves off as Uyghurs in order to gain commercial advantage from the new Tang alliance
with the Uyghur khaganate.10 In 831 and 836, the court had to issue edicts restricting private
commercial transactions and social interactions between Chinese and foreigners—but only
because so many Chinese in Chang’an were taking loans from these foreign merchants and
defaulting on the debt.11

This paper aims to build on a different reading of Han Yu and his millieu that has
emerged in recent Western-language scholarship. Such scholarship does not assume anti-foreign
angst and ethnocentrism to be the defining features of late Tang intellectual history: instead,
Peter Bol and Anthony DeBlasi argue that the An Lushan rebellion produced a “crisis of faith in
the ability of culture to influence human behavior,” or a “loss of certainty that the cultural
tradition—the chain of texts that stretched back through history to antiquity—could act as a
reliable guide for men in the present.” According to Bol and DeBlasi, responses to this crisis
primarily comprised efforts to revive a kind of literary culture believed to possess the morally

27, 35–38; Robert Joe Cutter, “History and ‘The Old Man of the Eastern Wall,’” Journal of the American Oriental
Society 106.3 (1986); 520 n. 125. Schafer’s characterization of Yuan Zhen as “exotically anti-exotic” deserves
further consideration: see Schafer, The Golden Peaches of Samarkand, 28, 33.
10
See Xiang Da , Tangdai Chang’an yu xiyu wenming 唐代長 西域文明 (Shijiazhuang: Hebei jiaoyu
chubanshe, 2001 [1933]), 8–9, 36–37.
11
The term zhuse ren 諸色人, which occurs in the 831 edict, has been a source of some confusion in Western
scholarship. It literally means “people of various colors,” but “various colors” was a standard expression meaning
“various types.” In the edict, the term refers to the various categories of Tang subjects—literati, eunuchs, merchants,
clergy, commoners, and so on—to whom the new restriction applied. Schafer rendered the zhuse ren as “various
colored peoples” and mistakenly conflated it with the “foreigners” (waihua ren 外 人) referenced in the 836 edict;
Gernet elaborated on Schafer’s misreading by claiming that the 836 edict referred to “‘people of color’, a term that
denoted foreigners from the regions beyond the Pamirs or from South-East Asia”; recently Mark Lewis,
paraphrasing Schafer’s interpretation, changed “various colored peoples” to the more politically correct “various
dark peoples.” Also, Schafer claimed that the edicts of 831 and 836 reflected growing popular resentment toward the
“arrogance” and “insufferable haughtiness” of “Uighur usurers” in Chang’an, and that this resentment was a major
cause of Wuzong’s anti-Buddhist persecution, which he calls an “outburst of xenophobia.” Neither assertion is
supported by the available evidence. Schafer’s notion that there were Uyghur usurers in Chang’an is based solely on
a superficial reading of Xiang Da: Xiang (following Kuwabara Jitsuzō 桑原隲藏) actually argued that most of the
‘Uyghurs’ were Sogdians and other ‘Westerners.’ Moreover, the 831 edict clearly holds the Chinese debtors
responsible for not paying their debts and thus depriving foreign merchants of the capital that they needed for trade;
it does not fault the merchants for engaging in usury. Nonetheless, an interpretation of the 831 and 836 edicts as
xenophobic seems to have become entrenched—even Holcombe, who is generally skeptical toward the theory of
late Tang xenophobia, reads them as examples of “hints of greater wariness, distaste, and suspicion” toward
foreigners. Ibid.; CFYG 999.11562–11563; Ge, Tangyun Huyin yu walai wenming, 69–70; Schafer, The Golden
Peaches of Samarkand, 20; Gernet, A History of Chinese Civilization, 294; Mark Edward Lewis, China’s
Cosmopolitan Empire: The Tang Dynasty (Cambridge, MA: Harvard University Press, 2009), 170; Holcombe,
“Immigrants and Strangers,” 109.
and sociopolitically transformative power needed for the restoration of good governance. 12 In
other words, the antiquity to which the late Tang literati called for a return was not defined by an
absence of foreign influences, but rather by literary forms that were fully grounded in moral
virtue and thus capable of instilling that virtue in others.13 DeBlasi argues that most late Tang
literati understood good writing to be based on broad, balanced learning and the ability to imitate
a wide range of literary models, and thus characterized by a high level of adaptability. They “co-
opted” the rhetoric of the “return to antiquity” (fugu 復 ), which “had a long history that
predated the Tang,” but only as “a rhetorical means for signaling one’s moral seriousness.” Han
Yu and his followers, on the other hand, took the ideal of antiquity seriously by identifying it
with a narrowly defined set of moral values called the “Way of the Sages” (shengren zhidao 聖
人之 ). To them, learning to write like the ancients was a means of “literary self-cultivation”
that enabled a literatus to rediscover the Way of the Sages and thereby become moral.14

The newer scholarship also recognizes Han Yu as anything but typical of the late Tang
elite. Han Yu’s skill as a prose stylist did gain him admirers throughout the ninth century:
according to Zhao Lin 趙璘 (fl. 834–853), younger literati began imitating his prose during the
Yuanhe 元和 era (806–820) and “literary style underwent a great change” 文體大變 as a
result.15 However, very few of these literary imitators appear to have espoused Han’s argument
that Buddhism was antithetical to the Way of the Sages and partly responsible for its decline, and
none are known to have followed him in arguing that the same was true of Daoism. 16 Nor did
Han’s ideology have any direct influence on Wuzong’s religious persecution, which was purely
Daoist in inspiration—a point that Kenneth Ch’en was already noting in a history of Chinese
Buddhism published just five years after Wright’s.17 Indeed, Wuzong had thousands of Buddhist

12
Peter K. Bol, This Culture of Ours: Intellectual Transitions in T’ang and Sung China (Stanford: Stanford
University Press, 1992), 109; Anthony DeBlasi, Reform in the Balance: The Defense of Literary Culture in Mid-
Tang China (Albany, NY: State University of New York Press, 2002), 117–118.
13
Although neither Bol nor DeBlasi engaged in cross-cultural comparison, their understanding of late Tang ideals
about literary culture reminds one of how Petrarch, a lover of “antiquity” in a different time and culture, idealized
Cicero as representing the perfect union of eloquence and moral wisdom. See Charles G. Nauert, Humanism and the
Culture of Renaissance Europe, Second Edition (Cambridge: Cambridge University Press, 2006), 23.
14
DeBlasi, Reform in the Balance, 5, 115–145.
15
Zhao Lin 趙璘, Yinhua lu 因話錄 (Shanghai: Shanghai guji chubanshe, 1979), 3.82.
16
David McMullen notes: “Only a small number of Han Yu’s followers, themselves obscurely placed enough for
their attitudes to be of little political consequence, remained actively anti-Buddhist.” We shall see that even his anti-
Buddhist followers did not criticize Daoism. Ge Zhaoguang, too, recognizes that Han Yu’s anti-Buddhist sentiment
was not widely accepted during his time. David McMullen, State and Scholars in T’ang China (Cambridge:
Cambridge University Press, 1988), 111; Ge, Zhongguo sixiang shi, vol. 2, 212.
17
Kenneth Ch’en, Buddhism in China: A Historical Survey (Princeton, NJ: Princeton University Press, 1964), 226.
See also the detailed analyses of Wuzong’s persecution in Stanley Weinstein, Buddhism under the T’ang
(Cambridge: Cambridge University Press, 1987), 115–136; Ge, Zhongguo sixiang shi, vol. 2, 230–256. Some
residual influence from Chen Yinke can still be seen in Ge’s analysis (242–243) of Wuzong’s chief minister Li
Deyu 李德裕 (787–850), however.
monasteries to close down and hundreds of thousands of monks and nuns to laicize in the 840s
precisely because Han Yu’s polemics had not diminished Buddhism’s appeal to the Chinese in
any way. Moreover, Wuzong’s death was quickly followed by a return to what Mark Halperin
calls “the generally benign and salubrious climate enjoyed by the [Buddhist] church in the late
T’ang and Five Dynasties.” 18 Interpretations that identify both Han Yu’s exclusivism and
Wuzong’s persecution as manifestations of a society-wide response to the An Lushan rebellion
thus underestimate their radicalism and overestimate their popularity. As DeBlasi, Edwin
Pulleyblank, and Jan De Meyer have argued, Han Yu’s insistence on ideological exclusivity and
purity was actually highly atypical and marginal in a pluralist, eclectic elite culture that generally
favored complementarity, balance, accommodation, and interconnectedness between the so-
called “three teachings” (sanjiao 教）of Classicism, Buddhism, and Daoism.19

Building on work by Bol, DeBlasi has also shown that nearly all of the other literati
whom Chen Yinke identified as members of the “Ancient Style movement” actually belonged to
the eclectic intellectual “mainstream.” Most of these men combined their literary pursuits with an
embrace of the Buddhist faith, and saw no incongruity in doing so—a mentality characteristic of
the pluralist mainstream but unacceptable to Han Yu. 20 Even Chen Yinke acknowledged the
difference between these other literati and Han Yu when he criticized the former for
“insufficiently clear understanding,” but his criticism reflects an ideological bias
anachronistically rooted in the Ming-Qing literati’s greater concern with Classicist orthodoxy, as
does Arthur Wright’s suggestion that these men, being “generally not opposed to Buddhism,”
could manage little more than “gropings” in the direction of a “Confucian revival.” I therefore
18
Mark Halperin, Out of the Cloister: Literati Perspectives on Buddhism in Sung China, 960–1279 (Cambridge, MA:
Harvard University Press, 2006), 31.
19
De Meyer characterizes Tang “intellectual eclecticism” as “an amazing variety of attempts to shape answers to the
great questions of the times by making use of elements derived from Confucianism, Daoism and Buddhism.” He
points out that this was not unique to the Tang and was rather the “continuation of a process which found its earliest
expression during the final centuries of the pre-imperial era, and which even the final triumph of Neo-Confucian
orthodoxy never entirely managed to stifle.” In a study of Korea from the tenth century to the twelfth, Remco
Breuker defines the pluralism of the Goryeo elite as “an ideology that allows the existence of contradictions and
inconsistencies between its constituent parts… accepts the alternative or simultaneous presence or use of
contradictory and incommensurable approaches… [and] maintains an aggregative instead of synthesized worldview,
offering an outlook that exists [sic] of simultaneously present, though not necessarily simultaneously used, partial
worldviews.” I would argue that the eclecticism or pluralism that De Meyer and Breuker describe was the norm in
Chinese, Korean, and Vietnamese elite culture until the ascendancy of Daoxue 學 (i.e., De Meyer’s “Neo-
Confucian orthodoxy”) in the fourteenth and fifteenth centuries made ideological correctness a requirement for
membership in the literati. Breuker’s theory is that Goryeo’s incorporation into the Yuan empire caused a “radical
shift away from a pluralist Weltanschaung,” but it is much more likely that the Joseon court’s embrace of Daoxue
ideology was responsible for this shift. Edwin G. Pulleyblank, “Neo-Confucianism and Neo-Legalism in T’ang
Intellectual Life, 755–805,” in A.F. Wright ed., The Confucian Persuasion (Stanford: Stanford University Press,
1960), 78, 95; DeBlasi, Reform in the Balance, 2–3, 102–113; De Meyer, Wu Yun’s Way, 104–105; Remco E.
Breuker, Establishing a Pluralist Society in Medieval Korea, 918–1170: History, Ideology and Identity in the Koryŏ
Dynasty (Leiden: Brill, 2010), 259, 309–310.

DeBlasi, Reform in the Balance, 23, 25–29, 41–42, 102–103. On Buddhism’s influence on Li Hua, Liang Su, and
20

Dugu Ji, see also David W. Tien, “Discursive Resources and Collapsing Polarities: The Religious Thought of Tang
Dynasty Scholar-Officials” (Ph.D. dissertation, University of Michigan, 2009).
agree with Bol and DeBlasi that we should differentiate the widely-shared literary ideal of
“returning to antiquity” (fugu) from Han Yu’s Ancient Style ideal, the latter being an extreme
agenda of ideological purity that also included a literary component of learning to write like the
ancients in order to acquire their moral values. Since this agenda was limited to Han Yu and a
handful of his students and associates (which arguably did not even include the strongly pro-
Buddhist Liu Zongyuan 柳 元), it is probably an overstatement to speak of him as joining and
then leading a distinct Ancient Style “movement” in the late Tang.21

Taking these reinterpretations of the late Tang intellectual context as a starting point, this
paper reassesses the significance of two related ideas that originated from Han Yu’s most
influential essay, the “Yuandao” 原 (“Finding the Source of the Way”): first, an idea that the
most essential difference between Chinese and barbarians was not geopolitical or ethnic, but
“ritual” (li 禮); second, an idea that this meant Chinese people who practiced a “barbarian (Yi-Di)
religion” 夷 狄 之 法 like Buddhism would thereby become barbarians. 22 Modern Chinese
scholars tend to credit Han Yu with distilling the essence of a fluid Chinese-barbarian dichotomy
that was always defined along cultural (or ritual, or moral) rather than racial (or ethnic)
boundaries. However, there was actually no recorded precedent for applying such a definition of
the dichotomy to contemporary (as opposed to historical) reality. In other words, Han Yu is the
first Chinese person known to have suggested that the Chinese of his own day could turn into
barbarians by continuously practicing a religion or ritual that was only meant for barbarians. This
assertion effectively made Han Yu’s concept of orthodoxy (or orthopraxy, in the case of ritual) a
precondition for Chinese identity, in an ingenious rhetorical strategy for delegitimating
Buddhism’s preceding four to five centuries of remarkable success in gaining Chinese adherents.

Among Western scholars, Charles Hartman and Marc Abramson have emphasized the
originality of Han Yu’s attempt at associating the practice of Buddhism with barbarization.
Hartman argues that Han Yu’s intention was “to demarcate the cultural boundaries between Hua
(i.e., Chinese) and Hu (i.e., foreigners),” to “[lay] down a cultural orthodoxy upon which to craft
the restored political unity of the future,” and to provide a “basic solution to the T’ang

21
Lo Lien-t’ien has pointed out that the term “Ancient Style movement” (Guwen yundong 文運動) originated
with Hu Shih 胡适 in 1928, when China was replete with political, cultural, literary, and social movements of every
kind. Hu Shih himself was a leading figure in the May Fourth and New Culture movements. Lo therefore warns that
using the modern neologism “movement” to describe a late Tang development produces a misleading impression of
highly coordinated action by an organized group. Zhang Shizhao’s Liuwen zhiyao also criticized Chen Yinke’s use
of the term “Ancient Style movement” as anachronistic. Unfortunately, mainland Chinese scholarship has largely
ignored these arguments, although Lo’s argument has gained the support of the Japanese scholar Higashi Hidetoshi.
See Lo Lien-t’ien 羅聯添, Tangdai wenxue lunji 唐代文學論集 (Taipei: Taiwan xuesheng shuju, 1989), 16; Zhang,
Liuwen zhiyao, 760; Higashi Hidetoshi 東英壽 (trans. Wang Zhenyu 王振 and Li Li 李 , et al.), Fugu yu
chuangxin — Ouyang Xiu sanwen yu guwen fuxing 復 創新里歐陽修散文 文復 (Shanghai: Shanghi guji
chubanshe, 2005), 107–111.

On the validity of translating fa 法 as “religion” in this context, see Robert Ford Campany, “On the Very Idea of
22

Religions (In the Modern West and In Early Medieval China),” History of Religions 42.4 (2003), 305–306.
dichotomy between Hua and Hu.”23 Likewise, Abramson argues that “the heart of [Han Yu’s]
discourse” was a notion that “the distinction between Self and Other is not a question of ethnicity
or origins per se but rather is dependent on abstract cultural values based on timeless norms,”
and that Han “was ahead of his time in his strident attempts to define boundaries between Han
and non-Han, Chinese and non-Chinese.”24 Unlike Hartman and Abramson, I would argue that
Han Yu’s ultimate concern was not one of defining Chineseness as a cultural identity. Neither
the Chinese-barbarian dichotomy nor the rejection of foreign religions was truly central to his
ideology. Han Yu’s aim was to redefine the Classicist identity in a way that made ideological
eclecticism unacceptable: the Way of the Sages was founded on Classicist principles alone and
thus could only be fatally compromised by accepting any other teaching, be it foreign or Chinese.
His occasional decisions to equate Classicist identity with Chineseness and to equate heterodoxy
with barbarism were only a rhetorical means to that end. Therefore, besides distinguishing
between the broadly shared fugu agenda and Han Yu’s more radical Ancient Style ideology, we
should also draw a distinction between Han’s ideological agenda and the rhetorical devices that
he utilized (indeed invented, in some cases) for advancing it. The novel idea of a highly
permeable Chinese-barbarian boundary belongs to the latter category.

Let us begin with a close reading of Han Yu’s argument. In the relevant section of the
“Yuandao,” he argues:

When Confucius wrote the Chunqiu, if [any of] the feudal lords used barbarian (Yi) ritual,
then [Confucius] regarded him as a barbarian, and if barbarians were promoted to the
level of the Central Lands, then he regarded them as [part of the] Central Lands. The
Classic [of the Analects] says, “The barbarians (Yi-Di) have rulers but are still not equal
to Chinese states that do not.” The ode says, “He attacked the Rong barbarians and Di
barbarians; he punished [the southern barbarian states of] Jing (i.e., Chu 楚) and Shu.”
But now we are elevating a barbarian (Yi-Di) religion above the teachings of the sage-
kings. How much longer [can this go on] before we all become barbarians (Yi)?
孔子之作 春秋 也 諸 用夷禮則夷之 夷而進於中國則中國之 經曰 夷狄之
有君 不如諸夏之亡 詩 曰 狄是膺 荊舒是懲 也 夷狄之
25
法 而 之 王之教之 幾何 不胥而爲夷也？

23
Charles Hartman, Han Yü and the T’ang Search for Unity (Princeton, NJ: Princeton University Press, 1986), 146,
151, 158.
24
Marc Abramson, Ethnic Identity in Tang China (Philadelphia: University of Pennsylvania Press, 2008), 67, 183.
25
Ma Qichang 馬 昶 ed., Han Changli wenji jiaozhu 韓昌黎文集校註 (Shanghai: Shanghai guji chubanshe, 1986),
17. Some Song-period and most modern editions of Han Yu’s works do not have the characters 夷而 in the first line.
Fan Wenli has demonstrated that these characters do appear in many Song, Yuan, Ming, and Qing texts that
anthologize or quote the “Yuandao,” and makes a strong case that they were in the original text. Fan is mistaken,
however, in claiming that while the Song-period Ge 閣, Hang 杭, and Shu 蜀 editions and Fang Songqing’s 方菘卿
(1135–1194) Hanji juzheng 韓集 正 all included these characters, Zhu Xi’s Changli xiansheng ji kaoyi 昌黎 生
集考異 followed another unspecified edition in omitting them from the main text and indicated in a note that some
editions included them. Liu Zhenlun’s study of the only extant Southern Song print edition of the Hanji juzheng
(now in Japan), which Fan did not consult, shows that almost the exact opposite is true: Fang Songqing followed the
Ge and Hang editions in omitting these two characters, while noting that the Shu edition included them; Zhu Xi then
The second quotation in this passage is from the ode “Bigong” 閟宮. It is clear from this ode’s
context that its subject is the reigning Lord of Lu 魯, “descendant of the Duke of Zhou and son
of Lord Zhuang” 周公之孫 公之子, which would make him either Lord Min 公 (r. 661–
660 BC) or Lord Xi 僖公 (r. 659–627 BC). However, Mencius supposedly interpreted the
“Bigong” as an ode about the Duke of Zhou himself. A well-known chapter of the Mencius
portrays its protagonist denigrating Xu Xing 許行 (n.d.), an ‘agrarianist’ philosopher who had
come to the Central Lands (i.e., the north Chinese states) from the southern kingdom of Chu 楚,
as a “shrike-tongued Man barbarian whose way is not that of the sage-kings” 南蠻鴃舌之人
非 王之 . Mencius then quotes these lines from the “Bigong” to argue that Xu Xing’s
Chinese followers were becoming disciples of the very same barbarians whom the Duke of Zhou
had fought as enemies. Later in the same chapter, Mencius uses the same lines to respond to
criticisms for being “fond of disputation” 好辯—that is, fond of engaging in aggressive polemics
against rival schools of thought. Mencius argues that since the followers of Yang Zhu and Mozi
are similar to the barbarians in “denying their fathers and rulers” 無 無君, he is simply carrying
on the Duke of Zhou’s mission by opposing them.26

Han Yu’s love of reading the Mencius as a young man had a formative influence on his
understanding of Classicist identity, and quite likely on his literary and rhetorical preferences as
well.27 When these passages from the Mencius are read alongside the “Yuandao,” it becomes
evident that the polemical sections of the Mencius served as a model for the use of the Chinese-
barbarian dichotomy in the “Yuandao.” Han Yu opposed Buddhism not because it was not
Chinese, but simply because it was not Classicism; his agenda was about refuting heterodoxy,
not repelling barbarians. Nonetheless, his anti-Buddhist polemics emphasized the Chinese-
barbarian dichotomy because he imagined himself an intellectual successor to Mencius. In the
Mencius, labelling a rival philosophy as barbaric is an easy way to make it look morally inferior
and strange without having to explain where that inferiority or strangeness actually lies.
Similarly, denigrating Buddhism as a barbarian religion was much easier than refuting it on a
philosophical level. This strategy was not even new to anti-Buddhist polemics. In the “Mouzi
lihuo lun” 牟子理惑論, possibly the earliest extant example of the Buddhist apologetic genre,
Mouzi’s challenger quotes Analects 3:5 (“the barbarians have rulers but are still not equal to

followed Fang Songqing’s lead. Fan Wenli 樊文禮, Rujia minzu sixiang yanjiu—xianqin zhi Sui-Tang 儒家民族思
想研究里 秦至隋唐 (Jinan: Qilu shushe, 2011), 218–222; Liu Zhenlun 劉真倫, Hanji juzheng huijiao 韓集 正
彙校 (Nanjing: Fenghuang chubanshe, 2007), 195; Zhu Jieren 朱傑人, Yan Zuozhi 嚴佐之, and Liu Yongxiang 劉
永翔 eds., Zhuzi quanshu 朱子全書 (Shanghai and Hefei: Shanghai guji chubanshe/Anhui jiaoyu chubanshe, 2002),
vol. 19, 450. See also the list of other Song-period editions that contain the characters 夷而 in Qu Shouyuan 屈 元
and Chang Sichun 常思春 eds., Han Yu quanji jiaozhu 韓愈全集校註 (Chengdu: Sichuan daxue chubanshe, 1996),
2682.
26
Mencius 3A:4, 3B:14.
27
Han Yu reveals in his writings that the Mencius was already his favorite classical text as a youth, and that he read
it long before reading the Xunzi or Yang Xiong’s Fayan. As a result, he developed a belief that Mencius was the
only true inheritor of Confucius’s Way. Ma ed., Han Changli wenji jiaozhu, 36–37, 261.
Chinese states that have none” 夷狄之有君 不如諸夏之亡) as one of his arguments against
the idea of Chinese people following a foreign creed.28

We also know from numerous other extant works of Buddhist apologetics that the
Daoists had been playing the ‘barbarism card’ against their Buddhist rivals for as many as five
centuries by Han Yu’s time, beginning with the Laozi huahu jing 老子 胡經 (Scripture on
Laozi Transforming the Westerners) of ca. 300. By late Eastern Han times, it was already widely
believed that Laozi became the Buddha after leaving the Central Lands and vanishing into the far
west. The Laozi huahu jing and its later derivatives used this myth to allege that because Laozi
designed Buddhism as a way of restraining the innately inhumane and immoral natures of
Westerners (Hu), its teachings could hardly be applicable to the innately good Chinese. 29 I think
we should be careful of assuming such arguments to be evidence that Daoists were particularly
ethnocentric or xenophobic. 30 Rather, the Daoist religion was never Buddhism’s equal in
philosophical sophistication, and Daoist polemicists recognized their creed’s indigenous origins
to be its only clear advantage over Buddhism. They were therefore determined to maximize that
advantage by claiming that a teaching from and for the barbarians had to be inferior to Daoism
and unsuited to the Chinese. Similarly, Han Yu had a limited understanding of Buddhist
philosophy compared to many of his literati peers, and he apparently never made an effort to

28
Hongming ji 弘明集 (T52:2102), 5.
29
Some versions modify the story by identifying the Buddha with Yin Xi 尹喜, the magistrate of Hangu Pass 函谷
關, whom Laozi enlightened before going west. In these versions, Yin Xi goes west with Laozi. The intent was
presumably to diminish Buddhism’s prestige further by making its founder a disciple of Laozi, rather than Laozi
himself. Kristofer Schipper has argued that stories about Laozi’s (or Yin Xi’s) conversion of the Westerners to
Buddhism were originally intended to represent Buddhism and Daoism “as complementary equals” and only
acquired an anti-foreign edge during the Tang, when Daoism “became part of Chinese national identity” and
“conceived of itself as a national ideology.” I believe this intepretation greatly underestimates the intensity of
Buddhist-Daoist rivalry prior to the Tang, as well as the pre-Tang “conversion” stories’ emphasis on the Westerners’
moral inferiority to the Chinese. Schipper also downplays the significance of the Laozi huahu jing story by noting
that “in the works preserved today, this story is never the main subject of any chapter, much less of a whole
scripture,” with the exception of the Laozi huahu jing itself, which has not been preserved and moreover “is quoted
practically nowhere.” This seems to overlook the real reason for the Laozi huahu jing’s ostensible lack of influence:
in 1281, the Yuan court ordered the destruction of all anti-Buddhist texts in the Daoist canon, including the Laozi
huahu jing. As a result, only quotations from these texts have survived, ironically, in Buddhist apologetic texts.
Schipper calls the Taishang dadao yuqing jing 大 玉清經 of ca. 750 “the most violently anti-Buddhist text
Taoism ever produced,” but it would probably be more accurate to call it one of the few overtly anti-Buddhist Daoist
texts that escaped Yuan censorship, to be restored to the Daoist canon under the Ming. There were many more such
texts before 1281, as shown by a list of the titles proscribed by the Yuan court. See Kristofer Schipper, “Purity and
Strangers: Shifting Boundaries in Medieval Taoism,” T’oung Pao LXXX (1994), 61–81. On the proscription of
1281, see De Meyer, Wu Yun’s Way, 128–129; Li Xiaorong 李 榮, ‘Hongming ji’ ‘Guang Hongming ji’ shulun
gao 弘明集 廣弘明集 述論 (Chengdu: Bashu shushe, 2005), 238–239. Chapter 3 of Li Xiaorong’s book
is a useful survey and analysis of the various known versions of the “Laozi transforming the Westerners” myth, as
well as earlier scholarly literature on the subject.
30
In my opinion, Li Xiaorong and Jan De Meyer are both overly confident in interpreting the use of ethnocentric
tropes in Daoist anti-Buddhist polemic as a sign of xenophobia and ethnic conflict. See Li, ‘Hongming ji’ ‘Guang
Hongming ji’ shulun gao, 256–260, 264–265, 336; De Meyer, Wu Yun’s Way, 138, 145.
address that weakness.31 He therefore borrowed the Daoists’ strategic use of ethnocentrism and
merged it with the Mencian polemical model to advance a new ideology of Classicist
exclusivism that treated both Buddhism and Daoism as the enemy.

Besides Classicist exclusivism, the other truly original aspect of Han Yu’s use of the
barbarism card against Buddhism was the argument that the Chinese would become barbarians
as a result of following a “barbarian religion,” because it implied that the basis of Chineseness
was adherence to a Chinese ethical code, “the teaching of the sage-kings.” This argument’s
cleverness lay in the likelihood that the literati were more prepared psychologically to respect a
“barbarian” teaching than to be labeled as barbarians because of it. Taken to its logical
conclusion, Han Yu’s formula meant that a Chinese who practiced Buddhism or any other
barbarian religion would literally become a barbarian, while a barbarian who practiced Classicist
(but presumably not Daoist) ritual and ethics could become Chinese. As Hartman phrases it, “to
behave like a Hua [Chinese] is to be a Hua, to behave like a barbarian is to be a barbarian.”32
This formula equated ideological convictions and ethical norms with Chinese identity in an
unprecedentedly absolute manner—or, to put it differently, it made Chinese identity
unprecedentedly relative and conditional on ideological and ethical purity. Mencius had accused
one of Xu Xing’s Chinese followers of being “changed by a barbarian (Yi)” 變於夷, but not of
having changed into a barbarian.33 Even the Daoists, who criticized their Chinese Buddhist rivals
for serving a foreign god and adopting foreign habits like sitting cross-legged, did not claim that
the Buddhists had thereby become barbarians.

The idea of literal “barbarization” thus appears to have been Han Yu’s invention. The
communis opinio since Han times, and possibly much earlier, had been that barbarians were born,
not made. In Analects 14:17, Confucius famously credited the statesmanship of Guan Zhong 管
仲 (ca. 720–645 BC) with saving the Central Lands from falling to the Di barbarians or to Chu:
“If not for Guan Zhong, we would now be wearing our hair untied and fastening our robes on the
left side” 微管仲 吾 被髮 衽矣. But this remark is ambiguous: one cannot assume that for
Confucius, dressing like a barbarian was equivalent to becoming a barbarian, especially if such
dress had been forcibly imposed by barbarian conquerors. Likewise, acculturation to “barbarian
(Yi) customs” 夷俗 among Chinese communities on the frontier was far from unheard of, but
was never described in terms of becoming barbarians. The closest precedent that I have found is
Fu Xuan’s 傅玄 (217–278) argument that if the Central Lands should lose the “teachings of
ritual and moral duty” 禮義之教 by which it has been able to keep the barbarians at bay, “then

31
Many ommentators on Han Yu’s anti-Buddhist polemics have pointed out his superficial understanding of
Buddhism, one of the earliest being Mao Kun 茅坤 (1512–1601). Lo, Han Yu yanjiu, 299.
32
Hartman also interprets the “inescapable” logic of Han Yu’s argument as being that “those who ‘today elevate
barbarian practices’ are essentially barbarians” (italics original to the text). Hartman’s use of the word “essentially”
should not be read as diluting Han Yu’s idea of literal barbarization by Buddhism; it does not mean “figuratively” or
“almost.” Rather, it implies that in this formula, the essence of barbarism is ritual and ethics, not ethnicity or place
of origin. Hartman, Han Yü and the T’ang Search for Unity, 157–158.
33
Mencius 3A:4.
[the Chinese] would be like the barbarians (Yi-Di)” 則 乎夷狄.34 But even this is a warning
about becoming as bad as barbarians, not about becoming barbarians. The distinction between
the two is subtle but important, since the latter makes a claim about the contingent nature of
Chinese identity that the former does not.

Han Yu’s choice to cite the Chunqiu as a basis for the idea of “barbarization” was equally
unprecedented. Much has been written about the fluidity of Chinese-barbarian dichotomy as
found in the Chunqiu, but most of it anachronistically projects the influence of Han Yu’s formula
onto earlier periods, as though he merely devised the most succinct way to express a time-
honored tenet of Chinese identity. In fact, we do not find any reference to the Chunqiu
commentaries’ language of barbarization and promotion in Chinese political or philosophical
discourse prior to the ninth century. This strongly suggests that such language was generally
regarded as an idiosyncrasy unique to the Chunqiu, rather than a principle that should apply to
the Chinese-barbarian dichotomy in general.35 The clearest expression of this purely exegetical
perception is in the Chunqiu fanlu 春秋繁露, a product of the Gongyang commentarial school
that is traditionally (but problematically) attributed to Dong Zhongshu. The author of the
Chunqiu fanlu notes: “The normal language of the Chunqiu credits the Central Lands, not the
barbarians (Yi-Di), with behaving in accordance with ritual” 春秋 之常辭也 不予夷狄而
予中國爲禮. But, the author goes on, there are significant exceptions in which the principle of
demotions and promotions takes precedence: “The Chunqiu does not have a uniform language
throughout; it shifts according to changes [in the circumstances]” 春秋 無通辭 從變而
移.36 The Chunqiu commentaries and subcommentaries and texts like the Chunqiu fanlu thus
provided some good material for a philosophical reinterpretation of the meanings of Chineseness
and barbarism, but that material’s potential had remained unexplored before Han Yu wrote the
“Yuandao.” Han Yu’s innovation lay in transforming the Chunqiu commentaries’ idea of
“barbarizing” demotion into a rhetorical strategy in the entirely separate genre of anti-Buddhist
polemics, and in claiming that “barbarization” was a real-world consequence (i.e., becoming a
barbarian) rather than just a figure of speech (i.e., being called a barbarian for behaving like a
barbarian). In other words, Han Yu’s rhetoric gave the idea of barbarization new implications for
the identity of those being “regarded as barbarians.”

While the Gongyang and Guliang commentaries to the Chunqiu are well known for
semantically demoting feudal lords to the level of barbarians, they never do so over matters of
ritual. On the surface, then, the only basis for Han Yu’s emphasis on “barbarian ritual” as a

34
Fu Xuan 傅玄, Fuzi pingzhu 傅子評註 (Tianjin: Tianjin guji chubanshe, 2010), 56.
35
While it is true that many pre-Tang texts have been lost, I believe it is significant that even surviving quotations or
fragments from lost pre-Tang texts do not include the use of the Chunqiu to make arguments about Chinese turning
into barbarians or barbarians turning into Chinese.
36
Strangely, the two examples of “barbarizing” demotion cited in the Chunqiu fanlu—namely, the semantic
demotions of Jin at the Battle of Bi (597 BC) and of Zheng during its attack on Xŭ 許 (588 BC)—are actually
not interpreted as such in the Gongyang commentary. See Su Yu 蘇輿 (Zhong Zhe 鍾哲 ed.), Chunqiu fanlu yizheng
春秋繁露義證 (Beijing: Zhonghua shuju, 1992 [hereafter CQFL]), 2.46, 2.63.
barbarizing factor would seem to be the Zuozhuan interpretation of the Qĭ rulers’ demotions,
especially as elaborated by Du Yu. The Zuozhuan attributes the semantic demotion given to Lord
Huan of Qĭ 桓公 (r. 637–567 BC) in 633 BC was attributed to his use of “barbarian (Yi)
ritual” 夷禮 when visiting the court of Lu. The Yi “barbarians” in this case were presumably the
indigenous peoples of the Shandong peninsula, most of whom had been subjugated or reduced to
vassalage by the Western Zhou. A slightly earlier passage had explained the posthumous
demotion of Lord Huan’s predecessor, Lord Cheng 公 (r. 654–637 BC), in 637 BC by simply
saying, “Qĭ was of the barbarians (Yi)” 夷也, even though the Qĭ ruling house was believed
to be descended from the kings of the Xia dynasty. Du Yu’s 杜預 (222–285) commentary to the
Zuozhuan interpreted this line to mean that Confucius demoted Lord Cheng upon his death
because he had begun using barbarian ritual and had continued doing so until he died.37 But
when a later Qĭ ruler, Lord Wen 文公 (r. 550–536 BC), received an identical demotion in 544
BC, the Zuozhuan explained it as a gesture “to show his inferiority” 賤之 without making any
reference to barbarians or barbarian ritual. Du Yu’s commentary attempted to harmonize the
discrepancy by stating that Lord Wen’s inferiority was due to his reversion to barbarian ritual.
According to Du Yu, Lord Huan had “given up using barbarian ritual” 捨夷禮 by 615 BC, as
seen from his reversion to the title “Earl of Qĭ” 伯 in the Chunqiu record.

Besides the Zuozhuan precedent, it is quite possible that Han Yu was also alluding to a
new trend in Chunqiu exegesis that had originated with Dan Zhu 啖助 (724–770) and Zhao
Kuang 趙匡 (fl. 770–775) and gained popularity through the influence of Lu Chun 陸淳 (d.
805).38 The works of Dan Zhu and Zhao Kuang have been lost, but many quotations from them
are preserved in Lu Chun’s extant works, which essentially collate, synthesize, and systematize
interpretations made by Dan and Zhao. Of particular interest to us is a passage in Lu’s Chunqiu
jizhuan zuanli 春秋集傳纂例 that outlines the principles for “barbarizing” demotion in the
Chunqiu:

When feudal lords behave like barbarians (Yi-Di), then they are written of as barbarians
(Di) [The same language is used for rulers and their ministers; they are referred to only

37
Li Xueqin 李學勤 et al ed., Chunqiu Zuozhuan zhengyi 春秋 傳正義 (Beijing: Beijing daxue chubanshe, 1999
[hereafter ZZZY]), 15.408, 16.435. On the purported Xia ancestry of the Qĭ rulers, see SJ 36.1583. Li Wai-yee has
recently made a persuasive argument that geopolitical conflict between Lu and Qĭ, not ethnocultural differences,
best explains why the Zuozhuan (which reflects a Lu perspective) sometimes represents Qĭ as barbaric or inferior. In
other words, the language of the Zuozhuan sometimes reflects Lu propaganda meant to assert superiority over Qĭ.
Indeed, in 633 BC Lu used the Qĭ ruler’s “disrespectful” 不共 (恭) performance of diplomatic ritual as a pretext to
invade Qĭ. Li Wai-yee 李 儀, “Hua-Yi zhi bian yu yizu tonghun” 華夷之辨 異族通婚, in Chiao Chien 喬健,
Chiu Tien-chu 助, and Luo Hsiao-nan 羅曉南 eds., Tanqing shuoyi: Qing, hunyin ji yi wenhua de kuajie lunshu
談情說異 情 婚姻暨異文 的跨界論述 (Taipei: Center for the Study of Foreign Cultures, Shih Hsin University,
2012), 45–63.
38
There is now quite an extensive body of literature on this new exegetical school. Ge Huanli 葛煥禮, Zunjing
zhongyi: Tangdai zhongye zhi Beisong monian de xin “Chunqiu” xue 尊經 義 唐代中葉至 宋 的新 春
秋 學 (Jinan: Shandong daxue chubanshe, 2011), 87–121, is particularly useful.
 by the name (of their state)]. Examples of this include the “man of Jin” defeating the Qin
army at Xiao [The Zuozhuan says that the men of Qin wished to station troops in Zheng
in order to plan a surprise attack on Jin, but the people of Jin defeated them at Xiao 39],
Zheng attacking Xŭ 40 , and Jin attacking the Xianyu [The Zuozhuan says that the Jin
[minister] Xun Wu claimed falsely to be heading for a meeting with the Qi army and
obtained permission from the Xianyu to pass through their territory. He then conquered
the state of Fei. That winter, Jin attacked the Xianyu from Fei41]. The same applies to the
“man of Jing (i.e., Chu)” coming to the court [of Lu]42, the “man of Chu” besieging [the
capital city of the state of] Song [At that time, Lord Wen of Jin had just become the
hegemon, but these feudal lords followed the Viscount of Chu in besieging Song. That is
why the Gongyang commentary says, “The Viscount of Chu is called a ‘man’ in order to
call these feudal lords ‘men’”43], and the “man of Chu” sending Yishen to present war
booty [to Lu]44.

39
The Zuozhuan account of the events leading to the Battle of Xiao actually does not support this interpretation.
Note that Lu Chun follows Dan Zhu in reading the reference to Qin as “the Qin army” as a demotion of the Qin ruler.
Dan Zhu did not accept the Gongyang and Guliang interpretation that the Chunqiu record of the Battle of Xiao also
calls the Jin ruler the “man of Jin” in order to convey moral disapproval of his decision to lead his army into battle
instead of seeing to the burial of his newly deceased predecessor. In his view, the Chunqiu was actually covering up
and thus excusing the Jin ruler’s violation of ritual norms for the sake of defending his state. Lu Chun evidently
adopts this position as well, as his annotation places no blame on the Jin ruler. ZZZY 17.470–471, 473–475; Zhong
Qianjun 鍾謙鈞 ed., Gu jingjie huihan 經解彙函, vol. 2 (Yangzhou: Guangling shushe, 2012), 1075, 1171, 1236.
40
This took place in 588 BC. Neither the Gongyang commentary nor the Guliang has any comment on this event,
but the Chunqiu fanlu (which belongs to the Gongyang tradition) argues that Confucius regarded the Zheng ruler as
a barbarian because he broke a newly-made covenant by attacking Xŭ. He Xiu’s Gongyang subcommentary has a
different interpretation: that Confucius regarded the Zheng ruler as a barbarian because he had aligned himself with
Chu in attacking other Chinese states. Lu Chun does not indicate which interpretation he accepts. CQFL 2.63; Li
Xueqin 李學勤 et al ed., Chunqiu Gongyang zhuan zhushu 春秋公羊傳註疏 (Beijing: Beijing daxue chubanshe,
1999 [hereafter GYZS]), 17.379.
41
The Guliang commentary’s interpretation of the Jin case was that the Jin ruler was regarded as a barbarian (Di 狄)
for “attacking the Central Lands in concert with barbarians (Yi-Di)” 夷狄交 中國, these barbarians being the
state of Chu (which was then attacking the state of Xú 徐). This seems to overlook the fact that the Xianyu were
themselves Di barbarians who had established a state known as Zhongshan 中 山 . Fan Ning’s Guliang
subcommentary tries to resolve this problem by claiming that since Zhongshan was on the North China Plain, it
counted as part of the Central Lands; note also that unlike Lu Chun, Fan Ning did not consider Xú, the target of the
Chu attack, to be a barbarian or barbarized state. Dan Zhu proposed an alternative interpretation—which Lu Chun
accepted—in which the Jin ruler’s offense lay in lying to the Xianyu and then launching a surprise attack on them (a
detail reported by the Zuozhuan). In Dan’s opinion, this was “something that [only] barbarians (Yi-Di) do” 夷狄之
所爲也; if “a lord and hegemon of the Central Lands practiced deceit on barbarians (Yi-Di)” 中國之 伯反行詐於
夷狄 and not vice versa, then he deserved to be spoken of as a barbarian. Zhong ed., Gu jingjie huihan, vol. 2, 1187;
Li Xueqin 李學勤 et al ed., Chunqiu Guliang zhuan zhushu 春秋 梁傳註疏 (Beijing: Beijing daxue chubanshe,
2000 [hereafter GLZS]), 17.332.
42
As mentioned above, the Gongyang and Guliang commentaries interpreted this as a slight promotion of the Chu
ruler. However, Lu Chun (following Dan Zhu) did not agree, for reasons that will be explained in a later footnote.
43
This is actually the Guliang commentary’s interpretation; the Gongyang interpretation is that only the Chu ruler
was demoted, to condemn him for taking the Song ruler captive at a multi-state conference in 639 BC. In another
work, the Chunqiu weizhi 春秋微旨, Lu Chun attributed the Guliang interpretation correctly and stated that Dan
 諸 有夷狄之行者,則以狄書之[君臣 辭 但 而已] 晉人敗秦師于殽[ 氏 秦人因
鄭而謀襲晉晉人敗之于殽…] 鄭 許 晉 鮮虞 [… 氏 晉荀吳僞會 師 假 於鮮
虞 遂滅肥 冬 晉 鮮虞 因肥也]之類是也 荊人來聘 楚人圍宋[… 時晉文新霸 諸
45
從楚子圍宋 故公羊曰 人楚子 所以人諸 也 ] 楚人使宜 來獻捷亦

“Behave like barbarians” is an ambiguous phrase that one also encounters in the
Gongyang commentator He Xiu’s 何 (129–182) interpretation of the cryptic phrase “new
barbarians” 新夷狄, found in the Gongyang commentary’s passage on the Battle of Jifu 雞 /雞
甫 in 519 BC. This interpretation seems to have been occasioned by language that seemed to
demote the Chu state’s Chinese vassals to the same level as their barbarian opponent at Jifu, the
ruler of Wu. He Xiu’s subcommentary offers the following explanation:

The Central Lands are different from the barbarians (Yi-Di) by virtue of their ability to
respect those deserving of respect. The royal dynasty was in disorder, yet none [of the
feudal lords] was willing to save it; relations between rulers above and ministers below
were in a state of ruin. Thus they, too, had newly begun behaving like barbarians. That is
why [Confucius] did not allow [Chu’s Chinese vassals] to have superiority over [Wu].
中國之所以異乎夷狄也 以 能尊尊也 王室亂 莫肯救 君臣 壞敗 亦新有
46
夷狄之行 故不使主之

Lu Chun’s examples of feudal lords behaving like barbarians also indicate that to him, barbaric
behavior by a Chinese lord could include acting perfidiously or opportunistically against other
states, as well as allying with a barbarian state against a Chinese state. This reflects the
Gongyang and Guliang commentarial traditions’ approach to “barbarizing” demotion.

However, Lu Chun also incorporates the Zuozhuan interpretation of “barbarizing”

demotion by placing as much emphasis on differences in ritual as on moral standards and
geopolitical allegiances:

When feudal lords use purely barbarian ritual, they are referred to only by the name of
their state. Examples of this include Jing (i.e., Chu), Wu, Xú 47 , and Yue…. Those

Zhu agreed with it. The siege of Song took place in 633 BC. GLZs 9.170; GYZS 12.255; Zhong ed., Gu jingjie
huihan, vol. 2, 1113, 1169.
44
This took place in 639 BC, shortly after the Chu ruler committed the devious act of taking the Song ruler captive
at a conference. The war booty consisted of the Song ruler’s possessions. Lu Chun follows the Gongyang
interpretation, which is that the Chu ruler is semantically demoted from viscount to “man” in order to criticize him
for taking the Song ruler captive. GYZS 11.243.
45
I have quoted and translated (in square brackets) only the most relevant annotations by Lu Chun. These are
differentiated from main text by the use of italics in the English translation and smaller font in the original Chinese.
Zhong ed., Gu jingjie huihan, vol. 2, 1108.
46
GYZS 24.517–518.
47
Little is known about this state, which was conquered by Wu in 512 BC and should not be confused with the
similar-sounding Xŭ. On the chronologically confused body of lore surrounding the most famous Xú ruler, the
 [vassals of Lu] who mix barbarian (Yi) ritual [with Chinese ritual] are referred to as “a
man,” for example, “a man of Zhu, a man of Mou, and a man of Ge came to court.”
諸 純用夷禮者 以國稱之 荊 吳 徐 越之類是也 … 雜用夷禮者 以人稱
之 邾人 牟人 葛人來朝 是也 48

Lu Chun’s interpretation of the Zhu, Mou, and Ge case as a demotion for using barbarian ritual
follows Zhao Kuang, but his interpretation of the Qĭ rulers’ demotions is highly ambiguous,
perhaps reflecting a reluctance to choose between the Zuozhuan interpretation and that of Zhao
Kuang.49 Zhao Kuang had asserted, modifying an interpretation by the Guliang subcommentator
Fan Ning 寧 (339–401), that the Qĭ demotions originated from the hegemons of the time and
had nothing to do with barbarian ritual or with Confucius.50 Unlike Zhao, Lu Chun stated that
“subtle messages” 微旨 from Confucius were to be found in such changes in title:

[The rulers of] Teng, Xue, and Qĭ were initially all referred to as marquises. After Lord
Zhuang [of Lu], [the ruler of] Teng is referred to as a viscount [In Lord Huan Year 2 (710
BC), he is called a viscount, but that is because he was in mourning at the time; that was
not his actual title51]. As for [the rulers of] Xue and Qĭ, after Lord Zhuang they are
sometimes referred to as marquises, sometimes as earls, and sometimes as viscounts. The
meanings of these [changes] contain subtle messages. They were originally not
barbarians (Yi-Di), nor were they vassals [of Lu], so they are not referred to as “a man”
[This means that they were originally feudal lords of the Central Lands, not barbarians;
hence [Confucius] passed judgment on them].
滕 薛 初皆書 公之後 滕則稱子[桓公 稱子 時爲在喪 非正 也] 薛
則自 公之後或稱 或稱伯 或稱子者 義 微旨 非夷狄 又非附庸 故
52
不書人也[言 中國諸 非夷狄 所以 之也]

It is unclear whether this was Lu Chun’s own interpretation or one that came from Dan Zhu:
Dan’s works have been lost, and Lu does not quote his interpretations of the Teng, Xue, and Qĭ

humane but excessively pacifistic King Yan 偃王, see Robin McNeal, “Returning to the Canon: A Review of
Michael Nylan’s The Five ‘Confucian’ Classics,” Early China 29 (2004), 278–294; Xu Yulong 徐玉龍, “Shichuan
yu minjian gushi zhong de Xu Yanwang gushi yanjiu” 史傳 民 故 中的徐偃王故 研究, Donghua Zhongguo
wenxue yanjiu 東華中國文學研究 7 (2009), 49–64.
48
Zhong ed., Gu jingjie huihan, vol. 2, 1108.
49
Lu Chun quotes Zhao Kuang’s interpretation of the Zhu, Mou, and Ge case at ibid., 1113, 1235.
50
Ibid., 1232–1233; GLZS 3.41, 9.166, 11.202, 16.311.
51
Here, Lu Chun agrees with an intepretation by Zhao Kuang. Whereas Fan Ning interpreted the Teng demotion of
710 BC as an act by the reigning king, Zhao Kuang attributed such demotions to hegemons. Since there was no
hegemon yet in 710 BC, Zhao inferred that the Teng ruler’s “demotion” for that year was part of the mourning ritual
for his predecessor. The Zuozhuan mentions (in another context) that feudal lords referred to themselves as zi 子
(meaning “viscount” but also “son”) when in mourning for their fathers. Zhong ed., Gu jingjie huihan, vol. 2, 1206;
ZZZY 13.356.
52
Zhong ed., Gu jingjie huihan, vol. 2, 1108.
cases. In any case, there is a great deal of ambiguity in Lu’s use of the word “originally” (ben )
and the phrase “passed judgment on them” (yizhi 之), not to mention the cryptic reference to
“subtle messages.” Does the phrase “originally not barbarians” mean that the Teng, Xue, and Qĭ
rulers had become barbarians at some point? Was Lu Chun suggesting that they belonged to a
special category of Chinese feudal lords who used a mix of barbarian and Chinese ritual but were
not vassals of Lu, unlike the rulers of Zhu, Mou, and Ge? If so, did Lu believe that the changing
titles of the Teng, Xue, and Qĭ rulers corresponded to changes in their use of ritual, as Du Yu had
surmised in the case of Qĭ?

Lu Chun’s interpretation of the principles behind the semantic promotion of barbarian

rulers would seem to be highly relevant to these questions. The section on “barbarians” (Yi-Di)
in the Chunqiu jizhuan zuanli outlines these principles as follows:

Barbarian (Yi-Di) rulers and ministers are referred to only by the name of their state,
examples being the Di barbarians, Jing (i.e., Chu), Wu, Xú, and Yue. When they visit the
court [of Lu] or attend a conference [of the states], the word ren (“a man of”) is typically
added [to the name of the state]. This is because one cannot say [for grammatical reasons],
“Jing came to court”53 or “Wu and a man of Zeng.”54 The same language is used for
rulers and their ministers, unlike the Central Lands.… Jing, Wu, and Yue all adopted
barbarian (Yi) customs, so they are referred to according to the conventions for
barbarians (Yi-Di). The labels “Rong” and “Di” are not added [to the names of these
states] in order to clarify that they were originally not barbarians. The Chu people were
initially following barbarian (Yi) customs, so they were referred to as “Jing” to label
them with the name of a province, as if saying that they were the barbarians of Jingzhou
[province]. Later, they could communicate [better with the Central Lands] and were no
longer like the barbarians, so they were [labeled] using the same conventions as the
Central Lands (i.e., as “Chu”). [The ruler of] Wu disputed [with Jin] over seniority at the
Huangchi conference and accepted a lower title as a result, so he is referred to by the
feudal title [of viscount] there [Previously, he had mostly used the title of king55]; at the

53
In the Chunqiu jizhuan bianyi 春秋集傳辯疑, Lu Chun quotes Dan Zhu’s explanation for this: “If we say, ‘Jing
came to court,’ then it would sound like the whole province came. So the word ren (‘a man of’) is added to complete
the meaning of the text, that is all. There is no other meaning to it” 若言 荊來聘 則似 州皆來 故
人 字以 文義爾 無他義. Dan was thus rejecting the Guliang and Gongyang interpretation that the Chu ruler
was promoted slightly in 671 BC. Lu Chun makes a similar argument with regard to the supposed promotion of the
Di ruler in 642 BC. Note, however, that in the Chunqiu jizhuan zuanli, Dan Zhu is quoted as using the 671 BC
example to argue, “Whenever a barbarian (Yi-Di) visits the court [of Lu], he should be called ‘a man,’ and the same
language should be used for rulers and ministers” 凡夷狄朝聘當稱人 君臣 辭. Ibid., 1053, 1220, 1230.
54
This example comes from the Gongyang commentary for Lord Xiang Year 5 襄公五 (568 BC): GYZS 19.421.

55
The royal titles used by the Wu and Chu rulers were considered a usurpation of the Zhou king’s authority; the
Chunqiu thus refers to them only by the feudal title of viscount when it even states their titles at all. There are
various accounts of what happened at the Huangchi conference of 482 BC. According to the Guoyu 國語, the Jin
ruler conceded seniority and precedence to the Wu ruler on condition that the Wu ruler give up his royal title and
refer to himself as the Duke of Wu 吳公. According to the Guliang commentary, the Wu ruler gave up his royal title
and used the title of viscount in order to show respect and submission to the Zhou king; Confucius therefore
promoted him semantically by referring to him as a viscount in the corresponding Chunqiu record. The Gongyang
commentary claims that while Confucius promoted the Wu ruler in order to acknowledge that he “presided over”
 Battle of Boju, the people of Cai employed him [as an ally against Chu], thus increasing
the esteem he was accorded, so his feudal title is written there; his ambassador [Ji] Zha
visited [Lu] and conformed to ritual, so he is referred to by his feudal title there [To
commend his ability to send an embassy to the Central States56]. In all other cases, he is
referred to according to the conventions for barbarians (Yi-Di).
夷狄之君臣皆書 國而已 若狄 荊 吳 徐 越之類是也 朝聘列會 例 人
字 不可言 荊來聘 不可連言 吳 鄫人 故也 但君臣 辭 異於中國
耳 … 荊 吳 越之輩全 夷俗 故依夷狄書之 不標 狄 之字 明
非夷狄也 楚人初 夷俗 則書曰 荊 以州言之 如曰 荊州之夷
後自通 又不 夷 則遂 中國之例 吳唯黃池之盟以爭長降號 [已前多稱王] 故稱
柏 之戰 以蔡人之 也 故書 使 來聘 以 禮故稱 [善 能聘中國
57
也] 餘則皆依夷狄例

This passage shows that Lu Chun differed from the Gongyang and Guliang commentaries by not
taking the barbarian status of Wu and Chu as given. Instead, he identified Wu, Chu, and Yue as
states whose originally Chinese rulers had become barbarians over time by adopting barbarian
customs and ritual. During Confucius’s time, the rulers of Wu were already claiming kinship
with the Zhou royal house through the legend of Taibo 伯, a senior uncle of King Wen who
had gone south and founded the state of Wu in order to let the Zhou kingship pass to King Wen’s
line. 58 Sima Qian’s Shiji gave the Taibo legend its final form and also ascribed sage-king
ancestors to the Chu and Yue rulers.59 These myths of Chinese ancestry had already entered the
sphere of Chunqiu exegesis through Fan Ning’s Guliang subcommentary, which followed the
Shiji in claiming that the first Chu rulers were descended from Zhurong 融, a minister under
the sage-king Ku 喾. According to Fan Ning, because “their state was close to the southern Man

(zhu 主) the Huangchi conference, he still mentioned the Jin ruler before the Wu ruler (thus giving the Jin ruler
precedence semantically) in order to show that he disapproved of “barbarians (Yi-Di) presiding over the Central
Lands” 夷狄之主中國. The Zuozhuan claims that the Wu ruler finally conceded precedence to Jin at Huangchi and
therefore did not preside over the conference at all. The Shiji contains two contradictory accounts, one following the
Zuozhuan and another following the Guoyu. Lu Chun evidently accepts a combination of the Guoyu and Guliang
versions, whereas quotations from Zhao Kuang in the Chunqiu jizhuan bianyi shows that Zhao rejected all existing
versions of the story. Zhao argued instead that the Jin and Wu rulers presided jointly over the Huangchi conference
using the “ritual protocol of equality” 敵禮. Guoyu, “Wuyu” 吳語 chapter; GLZS 20.396–397; GYZS 28.614–616;
ZZZY 59.1670–1671; Sima Qian 司馬遷, Shiji 史記 (Zhonghua shuju edition [hereafter SJ]), 31.1473–1474,
39.1685; Zhong ed., Gu jingjie huihan, vol. 2, 1264–1265
56
The Guliang commentary interprets this case of semantic promotion as the Chunqiu commending the Wu ruler for
employing the morally worthy Ji Zha as an ambassador to Lu in 544 BC, while the Gongyang claims that the one
being commended in this case was Ji Zha. Lu Chun presents his own interpretation, which resembles the Gongyang
interpretation of the Chu ruler’s slight promotion to “a man of Jing” in 671 BC. GLZS 16.311; GYZS 21.464–466.
57
I have quoted and translated (in square brackets) only the most relevant annotations by Lu Chun. These are
differentiated from main text by the use of italics in the English translation and smaller font in the original Chinese.
Zhong ed., Gu jingjie huihan, vol. 2, 1109.
58
In fact, the Taibo legend was the basis of the Wu ruler’s attempt to claim seniority over the Jin ruler at the
Huangchi conference.
59
See SJ 31.1445, 40.1689, 41.1739.
barbarians and gradually adopted their customs” 國近南蠻 遂漸 俗, the people of Chu were
“rejected and regarded as barbarians (Yi)” 棄而夷之. But Confucius eventually “promoted” 進
them back to the level of the Chinese once they “knew how to submit to the Central Lands and
also became great and powerful” 知内附中國 亦轉強大. Fan Ning also claims elsewhere that
the promotion of Chu had to do with its reversion to the use of Chinese diplomatic ritual.60

Like Fan Ning, Lu Chun inferred from the example of Chu that the process of
barbarization was reversible through increased contact with Chinese states. 61 Fan Ning’s
interpretation of Chu came close to treating barbarization in the Chunqiu as something that
actually happened to Chinese people who adopted barbarian modes of behavior, rather than just a
metaphor by which Confucius condemned certain acts as morally unacceptable (hence
“barbaric”). But his use of the phrase “rejected and regarded as barbarians” remains ambiguous:
had the Chu people literally become barbarians, or had they only become similar enough to
barbarians to be taken for barbarians themselves? Lu Chun appears to remove this ambiguity by
stating that Chu, Wu, and Yue “were originally not barbarians,” thus implying that they did later
become barbarians as a result of adopting barbarian customs. One can infer that Lu interpreted
Confucius’s apparent decision not to demote the Teng, Xue, and Qĭ rulers to the level of “man”
as a sign that these rulers were not as far along the process of barbarization.

As we have seen, Han Yu’s “Yuandao” applied a similarly literal and ritual-oriented
interpretation of barbarization in the Chunqiu to the contemporary context of Buddhism’s
influence on the Chinese. Does this indicate that Lu Chun’s Chunqiu exegesis had a strong
influence on Han Yu’s ideas, as some historians have claimed?62 It seems clear from Han Yu’s
writings that he was, at best, uninterested in Lu Chun’s scholarship, probably because he saw it
as tainted by association with the disgraced Wang Shuwen 王叔文 faction.63 Han Yu’s animosity
60
GLZS 6.102, 11.200.
61
This idea of barbarized Chinese states that could return to being Chinese suggests that Fan Wenli is wrong to
argue that the meaning of Han Yu’s phrase, “if barbarians (Yi) were promoted to the level of the Central Lands, then
he regarded them as [part of the] Central Lands” 夷而進於中國則中國之, changes significantly when the first two
characters 夷而 are omitted. After all, this omission merely implies that the subjects of promotion are the same
“feudal lords [who] used barbarian ritual” referenced in the preceding phrase. Fan, Rujia minzu sixiang, 222–223.
62
Hartman disagrees with Pulleyblank’s “more conventional” statement that Han Yu “seems to have concerned
himself very little with [the new Chunqiu scholarship],” claiming instead that Han “had strong sympathies and
sometimes direct contact” with the new school. But he cites no evidence for this reinterpretation. Hartman, Han Yü
and the T’ang Search for Unity, 174, 336 n. 5; Pulleyblank, “Neo-Confucianism and Neo-Legalism,” 112. See also
the discussion of Saiki Tetsuro’s theories below.
63
In 811 Han Yu wrote a poem of apology to the reclusive Luoyang poet and scholar Lu Tong 盧仝 (d. 835), after
Lu protested that Han (who was then magistrate of Luoyang) was being too harsh in applying the death penalty to a
group of voyeuristic ruffians against whom Lu had lodged a complaint. Two lines of the poem suggest that Lu
Tong’s style of Chunqiu exegesis was very similar to Lu Chun’s:

The [scrolls for the] three commentaries to the Chunqiu are left tied up on the shelf,

While you cradle the ancient Classic alone, investigating it from beginning to end.

春秋 傳 高閣
toward the Wang Shuwen faction predated its political downfall and was not dictated by mere
political expediency. Indeed, he suspected that the faction’s leaders had played at least an
indirect role in his demotion to the magistracy of Yangshan 陽山 county in early 804, possibly
because they had learned of his hostile attitude toward them from Liu Zongyuan and Liu Yuxi.64
It is worth noting that in the Veritable Records for Shunzong’s reign, which Han Yu compiled
for the imperial court in 813–815, Lu Chun’s name appears at the head of the list of “famous
men of the time who opportunistically sought rapid advancement” 當時 欲僥幸而速進者 by
aligning themselves with Wang Shuwen.65

Han Yu’s studied indifference toward Lu Chun’s scholarship is particularly evident when
compared to his warm praise in 818 for Yin You’s 殷 侑 (767–838) new Chunqiu
subcommentary, which was limited to expounding on the Gongyang commentary and therefore
subscribed to the traditional mode of Chunqiu exegesis. 66 Yin You’s commitment to the

獨抱遺經究終始

However, Han Yu describes Lu Tong’s exegetical style not in order to endorse it, but rather as part of a flattering
depiction of Lu as a high-minded eccentric who rejects all societal norms. Pulleyblank is therefore correct to argue
that Han Yu “is mainly interested in [Lu Tong] as a recluse and a poet, and his comments on his scholarship,
although appreciative, are made from the outside and show nothing of Liu Tsung-yüan’s fervor [for Lu Chun’s
scholarship].” That said, there is no evidence whatsoever for Pulleyblank’s description of Lu Tong as “a follower of
the Tan Chu school of criticism of the Spring and Autumn Annals.” While Lu Tong’s poems have survived, his
works of Chunqiu exegesis have not, and nothing is known of their connections (if any) with those of Dan Zhu,
Zhao Kuang, or Lu Chun. Pulleyblank, “Neo-Confucianism and Neo-Legalism,” 112; Qu and Chang eds., Han Yu
quanji jiaozhu, 540–541; Zhang Qinghua 張清華, Han Yu nianpu huizheng 韓愈 譜匯證 (Nanjing: Jiangsu jiaoyu
chubanshe, 1998), 272–273. On Lu Tong’s Chunqiu scholarship see Zhao Boxiong 趙伯雄, Chunqiu xue shi 春秋學
史 (Jinan: Shandong jiaoyu chubanshe, 2004), 414.
64
For various interpretations as to whether Han Yu’s suspicion was correct, see Zhang, Han Yu nianpu huizheng,
170–173; Lo, Han Yu yanjiu, 53–60; Wu Zaiqing 吳在慶, Tingtao zhai zhonggu wenshi lungao 聼濤 中 文史論
(Hefei: Huangshan shushe, 2011), 158–169; Hartman, Han Yü and the T’ang Search for Unity, 51–57.
65
In the Veritable Records, Lu Chun’s name is given as Lu Zhi 陸質, the new name that he adopted in 804/805. It is
interesting that Liu Yuxi and Liu Zongyuan are named at the very end of this list, perhaps showing that Han Yu held
them in higher regard and therefore downplayed their involvement in the faction. Saiki Tetsuro’s claim that Han Yu
held Lu Chun in high regard is based solely on misreading one of Shunzong’s edicts (quoted in the Veritable
Records) as reflecting Han’s own assessment of Lu. Saiki’s argument that the language of the Veritable Records
reflects Lu Chun’s influence on Han Yu is equally flawed, since its only evidence is the fact that both the Veritable
Records and Lu Chun’s exegetical writings use the phrase renqing 人情, which Saiki identifies as a “key concept” in
Lu Chun’s exegesis. It seems obvious from the examples Saiki cites that Han Yu does not even use the phrase in the
same way as Lu Chun: renqing means “public opinion” in the Veritable Records, while in Lu Chun’s writings it
means “common sense” or “ritual etiquette.” Qu and Chang eds., Han Yu quanji jiaozhu, 2876–2877, 2980; Ma ed.,
Han Changli wenji jiaozhu, 721; Saiki Tetsuro 木哲郎, “Kan Yu to ‘Shunjū’— Eitei kakushin wo megutte” 韓愈
と 春秋 永貞革新をめぐって, Chūgoku tetsugaku 中国哲学 35 (2007), 149–151. On the question of
whether the extant version of the Veritable Records for Shunzong’s reign is Han Yu’s work, see Denis Twitchett,
The Writing of Official History Under the T’ang (Cambridge: Cambridge University Press, 1992), 145–151, which
essentially answers in the affirmative.
66
The letter from Han Yu to Yin You, dated 818, shows that Yin had shown Han the subcommentary and explained
its general arguments to him some time before. Since another preface by Han Yu shows that Yin You was on a
diplomatic mission to the Uighurs for much of 817, the subcommentary was probably complete by the winter of 816,
when Han successfully recommended Yin for an appointment to the Censorate and the Department of Rites. Han’s
traditional mode of exegesis was such that he memorialized the throne in 822 to request that a
new examination on the three traditional Chunqiu commentaries be instituted in order to reverse
a decline in the study of these commentaries.67 We should not, however, read Han Yu’s letter as
evidence that he, too, had a strong preference for the traditional commentaries. Rather, Han
admits to Yin You that he has neglected the study of the Classics since passing the civil service
examinations, and laments that official duties and sheer laziness have kept him from making an
effort to learn more about Chunqiu exegesis from Yin You. He is ashamed to compare himself to
“men of learning and true Classicists” 學士真儒; only the fact that his “prose is close to that of
antiquity” 辭章近 has led Yin You to ask him to write a preface to the new subcommentary
despite his ignorance about its contents.68

Some of the language in this letter may be polite self-deprecation, but given the frank and
self-assured tone with which Han Yu speaks of his accomplishments as a prose writer in other
letters, the sense of embarrassment about his lack of familiarity with Chunqiu exegesis seems
quite genuine. The Chunqiu was clearly not Han Yu’s favorite classical text, perhaps because its
laconic prose was unattractive to him. I would therefore hypothesize that if he did derive the
concept of ritual barbarization from Lu Chun’s Chunqiu jizhuan zuanli, his interest in that text
was limited to searching for interpretations that he might use to turn Lu’s admirers against
Buddhism. Moreover, since Liu Zongyuan is the only contemporary of Han Yu known to have
been equally interested in Lu Chun’s exegesis and the practice of Buddhism, it is plausible that
this section of the “Yuandao,” at least, was written with Liu Zongyuan in mind.69

Let us turn to another of Han Yu’s famous polemical works: not his famous 819
memorial on the Buddha relic, which says nothing at all about barbarization, but rather his letter
to Meng Jian 孟簡 (d. 823) denying rumors of his conversion to Buddhism. In late 819, the Tang
court reassigned Han Yu from Chaozhou 潮州, where he had been exiled over the Buddha relic
memorial, to the slightly more northerly prefecture of Yuanzhou 袁州 (modern Yichun 宜春).
He arrived in Yuanzhou to take up the post of prefect in early 820. In the autumn of 820, one of
Han Yu’s subordinates was passing through Jizhou 州 (modern Ji’an ) while returning to
Yuanzhou from an errand when Meng Jian, who was the local deputy prefect 司馬, handed this

recommendation letter states that Yin “is familiar with all three [Chunqiu] commentaries and also versed in the other
Classics, and has gained his own insights beyond the commentaries and subcommentaries” 兼通 傳 傍習諸經
註疏之外 自有所得. Unfortunately, Yin You’s subcommentary has not survived. Ma ed., Han Changli wenji
jiaozhu, 208–209, 272–273, 603; Zhang, Han Yu nianpu huizheng, 340–341, 349–350, 359; Zhao, Chunqiu xue shi,
412–413.
67
THY 76.1398.
68
Ma ed., Han Changli wenji jiaozhu, 209.
69
Among the other members of the Wang Shuwen faction, Liu Yuxi was also interested in Buddhism but does not
seem to have been interested in Chunqiu exegesis, while Lü Wen 呂溫 (772–811) was a student of Lu Chun but
showed much less interest in Buddhism than Liu Yuxi and Liu Zongyuan did. On Lü Wen and Buddhism, see Chen
Jinping 陳津萍, “Lü Wen shengping jiqi zuopin yanjiu” 呂溫生 及 作品研究 (MA thesis, National Chung
Cheng University, 2002), 90–96.
official a letter addressed to Han.70 Meng had recently been demoted to Jizhou after being found
guilty of making large bribes to a powerful eunuch during Xianzong’s reign. 71 He was an
exceptionally devout Buddhist and, having heard rumors that the monk Dadian 大顛 (732–824)
had converted Han Yu to Buddhism in Chaozhou, was writing to Han to ask if they were true.72
Upon reading Meng Jian’s letter, Han Yu promptly wrote a reply to deny the rumors. He
admitted to having been friends with Dadian but maintained that he had no reason to “leave the
Way of the Sages, abandon the doctrines of the sage-kings, and follow a barbarian (Yi-Di)
teaching in pursuit of blessings and gain” 去聖人之 捨 王之法 而從夷狄之教以求福
73
利.

The second part of Han Yu’s letter to Meng Jian explained his reasons for opposing
Buddhism, presumably because Meng’s letter had represented his past anti-Buddhist polemics as
a mistake. Rather than pointing out Buddhism’s faults as he had in the past, Han Yu now argued
that he was simply trying to follow in Mencius’s footsteps by opposing Buddhism and Daoism.
Han Yu may have decided to focus on Mencius because Meng Jian was considered a descendant
of that philosopher. Han’s letter provides further evidence that the Mencian model was an
important source for his rhetoric of barbarization. Han Yu writes:

Moreover, there is also a good reason why I, Han Yu, do not support Buddhism and
instead attack it. Mencius said, “Now in all under heaven, those who do not go to Yang
Zhu’s [teachings] will go to Mozi’s. The teachings of Yang Zhu and Mozi together create
chaos, and the Way of the sages and worthies is no longer clearly understood. Thus the

70
Lo Lien-t’ien and Zhang Qinghua misread Han Yu’s reply letter as indicating that Han himself passed through
Jizhou and received the letter from Meng Jian. They also infer that this took place when Han traveled from
Chaozhou to take up his new post at Yuanzhou. This interpretation overlooks five points: first, Meng would not
have needed to write to Han if he could meet Han in person in Jizhou; second, Meng was only exiled to Jizhou two
or three months after Han’s arrival in Yuanzhou; third, Han clearly wrote his reply letter in autumn, and there is
little reason for him to have waited so long if he received Meng’s letter in spring; fourth, Hu Sanxing’s Zizhi
tongjian commentary states that the term xingguan 行 refers to a category of local official whose job was to run
long-distance errands for governors and prefects; fifth, the phrase “returned from the south” 自南迴 shows that
whoever received the letter was returning to Yuanzhou, not traveling there for the first time. Ma ed., Han Changli
wenji jiaozhu, 211–212; Sima Guang 司馬 et al. (with commentary by Hu Sanxing 胡 省), Zizhi tongjian 資治
通鋻 (Zhonghua shuju edition [hereafter ZZTJ]), 223.7162–7163; Lo, Han Yu yanjiu, 99; Zhang, Han Yu nianpu
huizheng, 386–387.
71
Meng Jian’s bribery of the eunuch Tutu Chengcui 突承璀 (d. 820) occurred in 818–819 and turned out to be
badly timed: Tutu was executed upon Muzong’s accession because he had favored another of Xianzong’s sons to
succeed him. But the case against Meng only arose because he arranged the murder of a trusted subordinate who
tried using the evidence of bribery to blackmail him. The murdered man’s family sought redress from the imperial
court, and Meng Jian’s bribes were discovered as a result. Ironically, Meng Jian was previously exiled in 811 for
criticizing Tutu Chengcui; the bribes were probably aimed at mending fences with Tutu. JTS 16.477, 163.4258;
XTS 160.4968.
72
Meng knew Sanskrit and participated in a court-sponsored sutra translation project in 811. His Jiu Tangshu
biography states that his Buddhist fervor was a target of criticism from Classicists, possibly a reference to Han Yu.
Meng’s Xin Tangshu biography implies that the criticism occurred in his last years. JTS 163.4257–4258; XTS
160.4969.
73
Ma ed., Han Changli wenji jiaozhu, 212.
 three bonds [of ruler-subject, father-son, and husband-wife] have declined and the nine
methods of governance are in ruin; the rites and ritual music have collapsed and the
barbarians (Yi-Di) wreak violence unrestrained. How much longer can this go on before
we become birds and beasts?” That is why he said, “Whoever can use his words to resist
the teachings of Yang Zhu and Mozi is a disciple of the sages.”
且愈不助 氏而排之者 亦有說 孟子 不之楊則之墨 楊墨交亂
而聖賢之 不明 則 綱 而九法斁 禮樂崩而夷狄橫 幾何 不爲禽獸也？
74
故曰 能言拒楊墨者 皆聖人之徒也

The second quotation in this passage does appear in the Mencius, but the first is only a very loose
paraphrase. The original text reads:

The words of Yang Zhu and Mo Di fill all under heaven; of the words spoken in all under
heaven, those that do not come from Yang will come from Mo. Yang’s teaching is based
on self-interest, which denies [one’s duty to] one’s ruler. Mo’s teaching is based on
universal love, which denies [one’s special affection to] one’s father. To deny one’s
father and ruler is to be a bird or a beast.
楊朱 墨翟之言盈 之言 不 楊 則 墨 楊氏爲 是無君也 墨氏
75
兼愛 是無 也 無 無君 是禽獸也

In this dialogue, Mencius apparently does not intend the “bird or beast” analogy to be taken
literally. He goes on to warn that if the teachings of Yang Zhu and Mozi continue unchecked,
wild animals and even starving human beings will eventually be eating the corpses of people
who have died of hunger. But he does not suggest that the human beings themselves will have
become wild animals by denying their fathers and rulers. Han Yu, on the other hand, amplifies
Mencius’s rhetoric of bestialization into an echo of the barbarization rhetoric from the
“Yuandao”: the phrasing of the question attributed to Mencius, “How much longer can this go on
before we become birds and beasts” 幾何 不爲禽獸也, is nearly identical to that of “How
much longer can this go on before we all become barbarians” 幾何 不胥而爲夷也.

The human-animal distinction and the Chinese-barbarian dichotomy are thus

interchangeable in Han Yu’s rhetoric. The Mencius, too, implicitly conflates barbarism and
animality by having Mencius claim that just as the Duke of Zhou attacked barbarians who
“denied their fathers and rulers,” he had no choice but to criticize any teaching that would lead
people to do the same.76 In other words, birds, beasts, barbarians, and the followers of Yang Zhu
and Mozi are alike in lacking filiality and loyalty. Moreover, the Duke of Zhou analogy also
involves another conflation of military action against foreign enemies with ideological warfare
against heterodoxies. Han Yu’s letter further reinforces both conflations by arguing that “we
would all be fastening our robes on the left side and babbling in a foreign tongue if there had
been no Mencius” 無孟氏 則皆服 衽而言侏離矣, since it was thanks to Mencius’s

74
Ibid., 214.
75
Mencius 3B:14.
76
Ibid.
teaching that the Chinese of Tang times—despite having lost the Way of the Sages—still knew
how to revere Confucius, to esteem humaneness and moral duty, and to respect legitimate rulers
rather than hegemons. 77 With the allusion to the fastening of robes, Han Yu is comparing
Mencius to Guan Zhong, whose political acumen Confucius credited with saving the Chinese
from a fate of “wearing our hair untied and fastening our robes on the left side” 被髮 衽.78
Whereas Confucius claimed that Guan Zhong had protected the Chinese states from literal
barbarians by military means, Han Yu claims that Mencius protected the Chinese from moral
barbarism by defending the Way of the Sages with his polemics against Yang Zhu and Mozi.

Yet the letter to Meng Jian does not make this analogy of barbarism and heterodoxy
explicit. Instead, the very next sentence of the letter claims that Mencius’s achievement was as
great as that of the sage-king Yu, who saved the Chinese world from a great flood.79 Nor does
Han Yu attempt to make the Guan Zhong analogy more coherent by mentioning the Chu
‘agrarianist’ philosopher Xu Xing, whom Mencius did label as a barbarian, in addition to Yang
Zhu and Mozi. These inconsistent, muddled, or incomplete analogies reflect a general pattern in
Han Yu’s writing: rhetorical effect takes precedence over intellectual coherence. To his mind,
animality and barbarism were both just metaphors for expressing an image of moral decline due
to ideological error. The real issue at stake for him was not the preservation or definition of
Chinese ethnocultural identity, even though his denigration of Buddhism as a barbarian teaching
may give us that impression. Rather, the issue was the urgent need for a renewed commitment to
Classicist values for the sake of rescuing the Way of the Sages from final oblivion. In the letter
to Meng Jian, this urgency finds expression in Han Yu’s stirring (if somewhat histrionic)
summation of his ideological agenda:

Since the Han dynasty, the many Classicists have merely been able to patch up [the
Way’s] hundreds of holes and thousands of wounds. Meanwhile, it has continued to
suffer distortion and loss. Its peril is like that of a single hair holding up a thousand jun
[i.e., eighteen metric tons], stretching long and unbroken but dwindling almost to
extinction. At such a time, to sing the praises of Buddhism and Daoism and encourage
everyone under heaven to follow them would, alas, be the height of inhumanity! The
harmful effects of Buddhism and Daoism are greater than those of the teachings of Yang
Zhu and Mozi, yet I, Han Yu, am not as worthy as Mencius. Mencius could not save [the
Way] when it was not yet lost, yet I, Han Yu, wish to preserve it when it has already been
corrupted. Alas, I am overreaching myself and blind to my own danger, and cannot save
myself from certain death! Even so, if this Way can be passed on in some imperfect form
because of me, I will have absolutely no regrets though I suffer destruction and death!
漢氏已來 羣儒 修補 孔千瘡 隨亂隨失 危如一髮引千鈞 緜緜
寖以微滅 於是時也 而唱 老於 鼓 之眾而從之 嗚呼 亦不仁甚
矣！ 老之害過於楊墨 韓愈之賢不及孟子 孟子不能救之於 亡之前 而韓愈乃

77
Ma ed., Han Changli wenji jiaozhu, 214.
78
Analects 14:17.
79
Ma ed., Han Changli wenji jiaozhu, 214.
 欲全之於已壞之後 嗚呼！ 亦不量 力且見 身之危 莫之救以死也！雖然 使
愈而粗傳 雖滅死萬萬無 ！80

Here, Han Yu is doing nothing less than presenting himself as Classicism’s long-awaited
savior—a role that Zhang Ji 張籍 (ca. 767–ca. 830) had urged upon him without success more
than twenty years before.81

For Han Yu, as for Mencius, it was rhetorically expedient to accentuate Classicism’s
importance to his audience by presenting either animality or moral barbarism as its only
alternatives. In other words, Han represents being human, being Chinese, and being an orthodox
Classicist as one and the same thing, probably not because he really believes the three categories
to be equivalent, but because that is the best way to emphasize how essential orthodoxy is.
Admittedly, Han does suggest in one of his shorter essays, the “Yuanren” 原人 (Finding the
Source of Humanity), that living beings are divided into three categories: human beings,
barbarians (Yi-Di), and animals. The three can collectively be called “humanity” (ren 人), Han
argues, because humans are meant to be masters over barbarians and animals.82 This system of
categorization effectively equates full humanity with Chineseness, although Han Yu does not
bother to explain why barbarians are less than human. The implications of the equation are not
explored in any of Han Yu’s extant writings, and it was only centuries later that the Daoxue
philosophers took these implications up and attempted to inject some coherence into them.83

Charles Holcombe has gone further than any other historian in challenging the notion of
late Tang xenophobia (albeit in highly tentative terms) and cautioning against the assumption
that Han Yu’s anti-Buddhist polemic “marks the beginning of some general anti-foreign shift in
Chinese attitudes.” Nonetheless, even Holcombe suggests that Han Yu’s insistence on orthodoxy
was indicative of a growing general intolerance of cultural diversity:

[There was] a subtle yet significant shift from the cosmopolitan openness of the early
Tang dynasty towards a less tolerant late-imperial Confucian universalism. The
pluralistic world of early Tang China, where invidious distinctions were (allegedly)
simply not made between Chinese (civilization) and foreign (barbarism), gave way in
mid- and late Tang to a more narrowly judgmental definition of civilized behavior—one
that was still supposedly ‘universal,’ in the sense of being theoretically open to anyone of
sufficiently elegant achievement, but which was now an orthodoxy whose Tao did indeed
have constant forms and names…. Cultural expectations throughout the Tang may have

80
Ibid., 215.
81
Dong Gao 董誥 ed., Quan Tangwen 全唐文 (Zhonghua shuju edition [hereafter QTW]), 684.7007–7009; Ma ed.,
Han Changli wenji jiaozhu, 131–136.
82
Ma ed., Han Changli wenji jiaozhu, 25–26.
83
Beginning with Cheng Hao 程顥: see Cheng Hao and Cheng Yi 程頤, Er Cheng ji 程集 (Beijing: Zhonghua
shuju, 1981), 122.
 gradually tended to become somewhat more uniform, with a corresponding loss of
tolerance for diversity.84

Holcombe’s addition of the word “allegedly” does suggest some skepticism about the
widespread notion that the Chinese-barbarian dichotomy did not matter in the early Tang. Such
skepticism is well-founded. As I have shown elsewhere, “invidious distinctions” did play an
important role in some rhetorical contexts during the early Tang, including both anti-
expansionist and pro-war rhetoric.85 These contexts all rested on an assumption that the boundary
between Chineseness and barbarism was both immutable and coterminous with the empire’s
present frontiers. Han Yu’s innovation lay in subverting that assumption and using the Chunqiu
to argue that the Chinese were susceptible to ritual, moral, and ideological barbarization when
they forsook the Way of the Sages. The reasons for this move were literary and rhetorical rather
than political or philosophical: Han Yu’s emulation of Mencius’s polemics, his attempt to exploit
new trends in Chunqiu exegesis against the Buddhists, and his predilection for playing with
dichotomies. If, as Holcombe argues, the upshot of Han Yu’s concept of Classicist orthodoxy
was a narrowing of the definition of civilization, the side-effect was a broadening of the
rhetorical uses to which notions of barbarism could be put. Han Yu’s rhetoric thus made the
concept of barbarization available to Chinese polemicists for use as a metaphor for ideological
heterodoxy or moral bankruptcy.

Bol, De Meyer, Holcombe, and Abramson have all argued that Han Yu’s formula took
him as far as claiming that Daoism, being unclassical, was as barbaric as Buddhism.86 While
such a claim would indeed be the formula’s logical conclusion, there is actually no indication of
it in the “Yuandao” or any of Han’s other writings. While Han Yu frequently paired Buddhism
and Daoism together as the major heterodoxies of his day, he never suggested that both teachings
were not Chinese. This suggests that Han was not interested in thinking through his formula’s
theoretical implications and was content to use it for anti-Buddhist polemic alone. Had Han Yu
tried to explore or elucidate the philosophical ramifications of his ideology-centered redefinition
of the Chinese-barbarian dichotomy, he could quite conceivably have arrived at the argument
that Daoism was barbaric. But there is no indication that he ever did so.87 It is thus likely that his

84
Elsewhere, Holcombe concedes that Wuzong’s proscription of foreign religions “may be a representative episode
of post-An Lushan xenophobia” (albeit a “notable belated and fleeting spasm”); follows Robert Joe Cutter in reading
Chen Hong’s “Dongcheng laofu zhuan” 東城老 傳 as evidence of “a certain amount of corroboration of ‘nascent
sinocentric xenophobia’”; and interprets certain prohibitions as “hints of greater wariness, distaste and suspicion”
toward foreigners. These concessions signal a reluctance to break completely from the old model of a xenophobic
turn. Holcombe, “Immigrants and Strangers,” 96, 101, 109, 173.

Shao-yun Yang, “What Do Barbarians Know of Gratitude?—The Stereotype of Barbarian Perfidy and its Uses in
85

Tang Foreign Policy Rhetoric,” Tang Studies 31 (2013), 28–74.

Bol, This Culture of Ours, 130; De Meyer, Wu Yun’s Way, 104; Holcombe, “Immigrants and Strangers,” 104;
86

Abramson, Ethnic Identity in Tang China, 67.

87
Some of Han Yu’s Northern Song admirers—notably Liu Kai 柳開 (947–1000) and Shi Jie 石 (1005–1045)—
would, however, make such a rhetorical move, although they did so via the disingenuous claim that Laozi was a
barbarian. See Quan Songwen 全宋文 121.292–294; Shi Jie 石 , Culai Shi xiansheng wenji 徂徠石 生文集
(Beijing: Zhonghua shuju, 1984), 10.116–117.
redefinition was a rhetorical means to an end, not an end in itself. Han Yu’s lack of interest in
arriving at a reading of the Chinese-barbarian dichotomy that had true philosophical coherence
and depth is very much in keeping with the “alternative picture” that David McMullen has drawn
of a man who never developed a “consistent and systematic philosophy” and was indeed
“temperamentally unsympathetic” to such a goal.88 He was at heart a rhetorician and a polemicist,
not a philosopher. Han’s ideas of a lost orthodoxy and barbarization appear in only two of his
writings, namely the “Yuandao” and the letter to Meng Jian. A comprehensive view of the full
corpus of his work, therefore, shows that whereas his literary creativity was almost limitless, his
attempts at developing a concept of orthodoxy were sporadic and never advanced beyond the
radical but simple idea that the Way of the Sages had to be saved from Buddhism and Daoism.
To borrow an analogy that Isaiah Berlin first applied to Tolstoy, Han Yu was a literary fox (who
knows many things) trying—rather halfheartedly, one might add—to be a philosophical
hedgehog (who knows one big thing).89

Thus far, I have argued that Han Yu did not propose to redefine the Chinese-barbarian
dichotomy as a supra-ethnic distinction between any human being who abided by Classicist
standards of morality and any human being who did not. I would also argue, however, that some
of his late Tang imitators came very close to doing so: Chen An’s 陳黯 (ca. 805–871) “Huaxin”
華心 (The Chinese Heart) and Cheng Yan’s 程晏 (n.d.) “Neiyi xi” 内夷檄 (A Call to Arms
against the Inner Barbarian) come to mind.90 The idea of barbarization proved so useful as a
rhetorical device that these admirers of Han Yu’s prose incorporated it into their own writing,
applying it to topics beyond the sphere of anti-Buddhist polemic.91 That said, the various modern

88
David McMullen, “Han Yu: An Alternative Picture,” Harvard Journal of Asiatic Studies 49.2 (1989), 606, 657.
89
Isaiah Berlin, The Hedgehog and the Fox: An Essay on Tolstoy’s View of History (Princeton, NJ: Princeton
University Press, 2013 [1953]), 1–5.
90
For the “Huaxin” see QTW 767.7986; Li Fang 李昉 et al ed., Wenyuan yinghua 文苑英華 364.8a–8b (Siku
quanshu edition). For the “Neiyi xi” see QTW 821.8650; TWC 49.545. There are two full translations of the
“Huaxin” into English, by Hartman and Abramson; Abramson has also made a full translation of the “Neiyi xi.”
Unfortunately, all three translations contain errors that affect the interpretation of the text, and Abramson’s
translation of the “Huaxin” shifts erratically between “Chinese” and “Han” as translations for Hua 華—a practice
that often imposes a concept of “Han” ethnicity artificially onto that essay’s use of the Chinese-barbarian dichotomy.
See Hartman, Han Yü and the T’ang Search for Unity, 158–159; Abramson, Ethnic Identity in Tang China, 180–181.
91
I would suggest that both the “Huaxin” and the “Neiyi xi” are inspired not only by the “Yuandao” but also by
another undated essay by Han Yu, usually referred to as Miscellaneous Discourse #3 雜説 . This essay uses myths
that the ancient sages had animal-like physical features to argue that it was better to look like an animal than to
behave like one:

A sage of the past had a head like that of a bull, another had a body like that of a snake, another had a beak
like a bird’s, and another had a face as ugly as an exorcist’s mask. They resembled [these things] in
outward appearance but not in their hearts. Can one say that they were not human? Some people have
excellent physiques and complexions, with faces as ruddy as cinnabar; they are handsome but vicious. Such
people have the appearance of human beings but the hearts of birds and beasts; how can they be called
human? In that case, judging people by the rightness or wrongness of their appearance is not as reliable as
judging them by their hearts and deeds.
historians who have identified Chen An and Cheng Yan as embodiments of the late Tang
zeitgeist or weltanschaung have greatly overestimated the amount of attention their works
received during their lifetime and, indeed, for more than a thousand years afterward. 92 If the
“Huaxin” and the “Neiyi xi” “probably represented the views of growing numbers of literate
elites in the late Tang,” as Abramson argues, it is strange that we see no other examples of such
arguments before the Song.93 The enduring influence that the idea of barbarization had on later
discourses about Chineseness owes much more to the fact that the “Yuandao”—not the “Huaxin”
or the “Neiyi xi”—became exceptionally influential among Northern Song literati. By the same
token, the “subtle yet significant shift” from ideological pluralism toward orthodoxy that
Holcombe traces to the ninth century actually did not begin until the eleventh.

The target of Han Yu’s attacks was mainstream Tang literati culture’s ideological
pluralism and eclecticism, not the Tang empire’s ethnic diversity. The idea of ethnic exclusivism
was irrelevant to Han Yu’s ideology: he neither advocated nor opposed it. Holcombe has argued
that because the “Yuandao” does not speak of the Chinese-barbarian dichotomy “in the familiar
modern terms of ethnicity, race, or nationality,” Han Yu saw the Way of the Sages as a universal
civilization, “theoretically open to anyone.” That may be so, but there is no evidence that Han
was interested in exploring this theoretical implication. 94 The objects of his ideological

昔之聖者 首有若牛者 形有若蛇者 喙有若鳥者 貌有若蒙倛者 彼皆貌似而心不

焉 可謂之非人耶？即有 脅曼膚 顏如渥丹 美而很者 貌則人 心則禽獸 又 可謂之人
耶？然則觀貌之是非 不若論 心 行 之可否爲何不失也

Just as the Miscellaneous Discourse claims that a sage with an animal’s head or body is more human than a human
being with an animal’s heart, the “Huaxin” and “Neiyi xi” assert that a barbarian who honors Chinese moral values
has more Chineseness in his heart than a Chinese person who dishonors them. Clearly, playing with discrepancies
between appearances (or names) and realities was one of the favorite polemical strategies used by Han Yu and his
imitators, and the Chinese-barbarian dichotomy was only one of several that were seen as open to such play. Note
also that the Miscellaneous Discourse’s argument is very similar to one found in the Liezi 列子, suggesting that Han
Yu was not above borrowing ideas from a Daoist philosophical text. Ma ed., Han Changli wenji jiaozhu, 34–35; Qu
and Chang eds., Han Yu quanji jiaozhu, 2709, 2714.
92
While Yuri Pines has claimed that in imperial China, “[s]upporters of the culturalist paradigm would eagerly quote
Chen An (ninth century),” I have not encountered a single text written before the twentieth century that quotes either
the “Huaxin” or the “Neiyi xi.” The first Chinese historian to cite the “Huaxin” seems to have been Chen Yuan in
1923, although he used it as evidence for the “sinification” of foreigners rather than an example of Chinese
“culturalism.” John Fincher was the first scholar to note the “unusual clarity” of the “culturalist” position in the
“Huaxin,” although he mis-dated the text to the early Song. Xie Haiping appears to have been the first historian to
notice the “Neiyi xi”; in 1978, he cited it and the “Huaxin” as examples of Tang “culturalism.” Pines, “Beasts or
Humans,” 60; Chen Yuan 陳垣, Yuan xiyu ren huahua kao 元西域人華 考 (Shanghai: Shanghai guji chubanshe,
2000 [1923]), 4–5; John Fincher, “China as a Race, Culture, and Nation: Notes on Fang Hsiao-ju’s Discussion of
Dynastic Legitimacy,” in David C. Buxbaum and Frederick W. Mote eds., Transition and Permanence: Chinese
History and Culture (Hong Kong: Cathay Press, 1972), 63–64; Xie Haiping 謝海 , Tangdai liuhua waiguoren
shenghuo kaoshu 唐代留華外國人生活考述 (Taipei: Taiwan shangwu yinshuguan, 1978), 8–9.
93
Abramson, Ethnic Identity in Tang China, 184.
94
Holcombe’s citing of the “Huaxin” in support of this assertion is misleading, since Han Yu never wrote anything
remotely similar to that essay: “Immigrants and Strangers,” 104–105.
proselytizing remained always the Chinese literati or literatus-like Chinese monks. While non-
Chinese individuals—whether foreign Buddhist monks or foreign jinshi candidates like the Li
Yansheng praised in Chen An’s “Huaxin”—are conspicuously absent from the literary output
produced by Han Yu’s social interactions, this was hardly exceptional for a late Tang literatus
and should not be read as evidence that he avoided contact with foreigners.95 Chen Sanping has
recently attempted to show that Bai Juyi had a strong affinity to Central Asia, and argued that
Han Yu’s “strong sinocentric stance” as a “fierce defender of Chinese tradition against all
foreign intrusions” was the reason for the “striking lack of close relation and interaction”
between him and Bai.96 Leaving aside the problem of whether Chen’s interpretations of Bai and
Han are correct, it should be noted that whereas Bai and Han did exchange a few poems in 821–
82297, not one of Bai’s many extant poems is addressed to a foreign individual or even mentions
interacting with a foreigner.

As for the question of whether Han Yu was exclusive or intolerant in cultural matters, as
Holcombe seems to imply, that would depend entirely on how one chooses to define the
nebulous concept of culture. It is clear that Han Yu did not understand Classicism or the Way of
the Sages in the way that we would understand “Chinese culture” today, as the sum of all cultural
forms produced by the Chinese throughout history. To him only certain Chinese cultural forms—
those that were created by the sages—were worthy of preservation; all other forms had to be
rejected as threats to the purity of the culture inherited from antiquity. The boundaries of his
normative “culture” were thus not ethnic but ideological; his cultural exclusivity was predicated
on a sense of Classicist ideological exclusivity that regarded foreign creeds and Chinese
heterodoxies as equally unacceptable. Buddhism’s popularity with the Chinese was, to him, only
the latest symptom of a deeper cultural or ethical malaise that began long before that religion’s
introduction to the Central Lands.

95
Chen Yinke has cited a jueju 絕 (quatrain) poem attributed to Han Yu in the Quan Tangshi 全唐詩, titled “To a
Sutra-translating Monk” 贈 經僧, as evidence that his anti-Buddhist stance was based on anti-foreign sentiments:

萬 言 路 Don’t talk about the ten-thousand-li road you’ve traveled—

有誰教汝度流沙？Who asked you to cross the shifting sands?
隻 中國方多 The Central Lands are now troubled enough;
不用無端更亂華！No need or reason for you to bring more disorder to the Chinese!

The references to the Central Lands and to the Chinese imply that this poem was addressed to either a specific
foreign monk or a generic foreign monk representing those who had introduced Buddhism to the Chinese centuries
before. The Quan Tangshi traces the poem to a twelfth-century anthology of Tang jueju, but it is absent from the
various Song-period editions of Han Yu’s collected works. Moreover, its plain, direct, and colloquial language
differs greatly from his poetic style. Two modern editions of Han Yu’s poetry have therefore expressed doubt as to
its authenticity. The poem may have been composed in the Song period and falsely ascribed to Han, the most
famous anti-Buddhist literatus of the Tang period, in order to raise its prestige. Chen, Jinmingguan conggao chubian,
329; Qu and Chang eds., Han Yu quanji jiaozhu, 3029; Qian Zhonglian 錢仲聯, Han Changli shi xinian jishi 韓昌
黎詩繫 集 (Shanghai: Shanghai guji chubanshe, 1984), 15.
96
Sanping Chen, Multicultural China in the Early Middle Ages (Philadelphia: University of Pennsylvania Press,
2012), 181.
97
Lo, Han Yu yanjiu, 181–182.
 An attitude of caution with the use of the word “culture” is especially germane because
the “late-imperial Confucian universalism” to which Holcombe refers is essentially the same
discourse of Chinese identity that Joseph Levenson and many later scholars of Chinese
nationalism have labeled as “culturalism.”98 As my preceding arguments in this paper will have
made clear, I do not accept the conventional assumption that a “culturalist” interpretation of the
Chinese-barbarian dichotomy originated with Confucius and has existed throughout imperial
Chinese history. Although the Chunqiu exegetical traditions’ notions of demotion and promotion
contained the potential for such an interpretation, it was Han Yu and his imitators who
inadvertently laid the true intellectual foundations for “culturalism” when they began using the
ideas of moral barbarism and barbarization as polemical tropes. At the time, they had no idea that
they were creating a fundamental reinterpretation of what it meant to be Chinese or to be
barbaric; their concern was only with immediate rhetorical agendas—agendas that were at first
anti-Buddhist, then much more varied. But the effect of such rhetorical innovation and imitation
was to create a new way of talking about the Chinese-barbarian dichotomy, which in turn
eventually enabled a new way of thinking about it in the Song and later periods. In other words,
“culturalism” as rhetoric preceded “culturalism” as philosophy or ideology.

Ironically, whereas Levenson posited “culturalism” and modern Chinese nationalism as

distinct opposites in terms of the object of one’s loyalty, Gernet seems to have seen “culturalism”
as a form of Chinese protonationalism that first appeared in the late Tang:

The term “nationalism” would be an anachronism, yet it was certainly reactions

analogous to nationalism that took vague shape after An Lushan’s rebellion and that were
to become evident again on other occasions in China’s history. This attachment to an
authentic tradition supposed to have been corrupted by foreign elements, this desire to
return to the pure—and imaginary—sources of orthodox thought and morality are
difficult to sum up, since they do not relate to the quite recent idea of a nation, but to the
idea of culture. If we wanted one word for them, we should have to invent the barbarous
term “culturalism.”99

Similarly, Stephen Owen has spoken of Han Yu as representing a “‘cultural nationalism,’ if we

may call it that,” in which “some Mid-Tang intellectuals… gave their allegiance more
exclusively to traditionally Chinese social and moral values,” while James Townsend would
probably identify Han’s anti-Buddhist polemics as a perfect example of “‘culturalism as
movement’, in which conscious argument and action become necessary to defend a culture under
threat.”100 I hope to have proven by now, however, that this emphasis on a distinction between

98
On the concept of “culturalism” see especially James Townsend, “Chinese Nationalism,” The Australian Journal
of Chinese Affairs 27 (1992), 97–130.
99
Gernet seems strangely unaware of Levenson’s prior invention of the term “culturalism” in the 1950s, as well as
the irony in his use of the word “barbarous.” Gernet, A History of Chinese Civilization, 293.

More recently, Ge Zhaoguang has also characterized Han Yu’s ideology as one of “cultural nationalism” (wenhua
100

minzu zhuyi 文 民族主義). See Ge, Zhongguo sixiangshi, vol. 2, 211–212; Owen, The Poetry of Meng Chiao and
Han Yü, 3; Townsend, “Chinese Nationalism,” 105–106.
foreign and Chinese cultures is deeply misleading: the intellectual origins of “culturalism” had
nothing to do with xenophobia. I would also suggest that the closest analogues to Han Yu’s
ideology are not modern nationalist movements but instead Renaissance humanism and the
Protestant Reformation, both of which were founded on a new historical consciousness that saw
most of Western Christendom’s prior history as a long age of degeneration. In this I build on
Stephen Owen’s brief but insightful observation, in a later work on the literary culture of Han
Yu’s day, that the Reformation is the most obvious European analogy to the “remarkable
abrogation of continuous history [by which] Han Yu declared himself and his moment a turning
point in Chinese culture, a leap across more than a millennium to resume the Confucian tradition
that had fallen into error and corruption after Mencius.”101

Charles Nauert, building on an influential article by Theodore Mommsen and modifying

Jakob Burkhardt’s seminal The Civilization of the Renaissance in Italy, observes that Petrarch
was the first European to develop a conception of history as divided into three periods:

[Petrarch] was in effect declaring something like this: “In addition to the ancient age of
light, of high civilization, and the modern [we would say medieval] 102 age of darkness
and barbarism, there is a third age, a new age of light, and it begins with me!” This sense
of standing at a turning-point in human history, and of reviving lost civilization, explains
what often seems an exaggerated sense of self-importance, an outlook that Jakob
Burkhardt labelled ‘individualism’ and demonstrated largely with reference to the
humanists’ thirst for fame…. Burkhardt made the mistake of putting a secondary
characteristic, a heightened sense of individualism, in place of the truly primary
characteristic, the new historical consciousness that emerged in the thought of Petrarch.
This sense of being engaged in the restoration of true civilization after many centuries of
barbarian darkness finds its first clear statement in the works of Petrarch, and some such
claim is common to virtually all of those writers (Salutati, Valla, Ficino, Erasmus) whom
historians identify as the crucial figures in the history of humanism.103

Clearly, these observations are highly applicable also to Han Yu’s reinterpretation of the history
of Chinese civilization.104

101
Shortly afterwards, however, Owen reverts to the traditional image of late Tang xenophobia, writing of “a China
that is, for the first time, conceived in times of excluding the foreign, as is proposed in Han Yu’s famous ‘Memorial
on the Buddha Bone.’” Owen, The End of the Chinese ‘Middle Ages,’ 9, 16.
102
Note original to the text.

Nauert, Humanism and the Culture of the Reniassance in Europe, 20–21. For Mommsen’s argument see
103

Theodore E. Mommsen, “Petrarch’s Conception of the ‘Dark Ages,’” Speculum 17.2 (1942), 226–242.
104
Interestingly, Owen argues that the unprecedentedly individualistic use of writing and claims to interpretive
authority by Han Yu and some of his peers heralded the “end of the Chinese ‘Middle Ages’” (hence the title of his
book), but does not credit Han Yu with inventing the concept of a “Middle Age” in Chinese history. Note also that
Owen (unlike Burkhardt) argues that the challenges to received textual authority seen in the Renaissance and
Reformation were not ultimately driven by individualism. Owen, The End of the Chinese ‘Middle Ages’, 55–57.
 It may be particularly interesting to think of Han Yu in comparison not only to Petrach
but also to Desiderius Erasmus and Martin Luther. In a manner somewhat similar to Han Yu’s
fulmination against Buddhism, Erasmus condemned scholastic theology and Medieval Latin as
the intellectual and literary sides of “barbarism”—products of a cultural decline from classical
(that is, Roman) standards of learning and eloquence that in turn had brought about the Christian
world’s moral degeneration under the weight of ignorance, corruption, and superstition. Indeed,
Erasmus’s most concerted attempt at a narrative of this decline bears the title Antibarbari,
“Against the Barbarians.”105 At the same time, Han Yu’s claim to be engaged in reviving a long-
lost Classicist Way bears some resemblance—as Owen noted—to Luther’s equally radical claim
(which Erasmus rejected as sheer arrogance 106) to have discovered the truth of the Christian
gospel, obscured by thirteen centuries of erroneous church tradition and teaching. Han Yu’s
Buddha relic memorial, which rejects the then-widespread belief in the spiritual potency of relics,
can perhaps also be compared to Luther’s attack on the sale of indulgences in the Ninety-five
Theses. Moreover, Han Yu’s reinterpretation of Classicist identity as an exclusive ideological
commitment was as fundamental as Luther’s reinterpretation of Christian soteriology and
arguably more influential in the long run, since it came to be accepted by all who identified
themselves as Classicists.

These striking similarities are accompanied by key differences, of course. Unlike

Petrarch, Erasmus, and other Renaissance humanists, who dreamt of reviving the classical
civilization of ancient Rome but did not think one had to reject Christianity in order to do so,
Han Yu argued that the Way of the Sages could not be restored as long as Daoism and Buddhism
continued to confuse and mislead the Chinese. Whereas Luther’s excommunication and
condemnation as a heretic lasted to the end of his life and beyond, Han Yu’s denigration of the
Buddha was soon forgiven and his exile was relatively brief. Erasmus’s notion of barbarism was
focused on literary style and theological method rather than heterodoxy, whereas Han Yu never
associated parallel style prose with barbarism. The impact of the Ciceronian works rediscovered
by Petrarch, and of Erasmus’s new edition of the New Testament, stands in stark contrast to Han
Yu’s lack of influence on classical scholarship, although one may speculate whether things might
have been different had Han completed his commentary to the Analects.107 Most importantly,
Petrarch, Erasmus, and Luther all started important intellectual or religious movements within
their lifetimes, whereas Han Yu’s intellectual influence in the ninth century was negligible

On Erasmus, I have benefited greatly from István Bejczy’s Erasmus and the Middle Ages: The Historical
105

Consciousness of a Christian Humanist (Leiden: Brill, 2001).

106
Ibid., 170–182.
107
A poem by Zhang Ji, describing the scene at Han Yu’s deathbed, mentions an unfinished Analects commentary;
other sources speak of a ten-juan Analects commentary by Han that was circulated among his students, and Han
himself says in an undated letter that he had once worked on such a commentary. It is thus likely that the ten-juan
work, which was lost after the Northern Song, was only an incomplete draft. Hartman has attempted to identify it
with the two-juan Lunyu bijie 論語筆解 (Jotted Notes on the Analects) attributed to Han Yu and Li Ao since the
Song period. However, I believe that McMullen and Chen Jo-shui are correct to conclude that there is insufficient
evidence for the authenticity of the Lunyu bijie. Lo, Han Yu yanjiu, 115; Ma ed., Han Changli wenji jiaozhu, 3, 728;
Qu and Chang eds., Han Yu quanji jiaozhu, 2760; Hartman, Han Yü and the T’ang Search for Unity, 339–341;
McMullen, “Han Yü: An Alternative Picture,” 640–642; Jo-shui Chen, Liu Tsung-yüan and Intellectual Change in
T’ang China, 773–819 (Cambridge: Cambridge University Press, 1992), 124 n. 92.
compared to his literary reputation.108 Nonetheless, Han Yu’s arguments about the lost Way of
the Sages eventually served as an inspiration to several successive generations of Song literati.
As is well known, some of these literati used the idea of the Way of the Sages as an ideological
foundation for building the philosophical system eventually known as Daoxue or Lixue, which
has had a profound influence on the study and interpretation of the Classics since Song times.

A detailed treatment of these developments after the ninth century would lie beyond the
scope of the present paper. Instead, I will use the rest of this paper to address Chen Yinke’s
influential theory that the Hebei military provinces’ defiance of imperial authority was rooted in
an ethnocultural divide, as well as the accompanying theory that the ‘barbarization’ of Hebei
inspired Han Yu’s anti-Buddhist polemics. Although most recent Chinese interpretations of Han
Yu have accepted both theories, I find it highly unlikely that the late Tang literari saw the Hebei
problem as one of barbarization at all. As we saw at the beginning of this paper, Chen Yinke
noted that the Hebei military originated from surrendering rebel armies whom the court
permitted to remain there after the An Lushan rebellion. These included significant numbers of
foreign (Hu) troops who had previously served in the northeastern frontier armies. Chen Yinke
claimed that the Chinese troops in Hebei were also “Hu-acculturated” (Huhua) due to long
exposure to these foreigners; they had thus lost their sense of cultural affinity to the Tang court
elite and instead identified culturally with the foreign troops. To Chen’s mind, this identification
effectively amounted to a sense of ethnic separatism. Chen further argued that when Han Yu
condemned Buddhism as a barbaric religion, he was motivated by a realization that Hu and “Hu-
acculturated” troops had barbarized a part of the Chinese heartland. In other words, his vehement
opposition to a foreign religion was part of a broader reaction against foreign intruders.109

Despite the authority that Chinese historians routinely accord to Chen Yinke’s theory of
Hebei barbarization, the theory is mainly founded on a kind of ethnocultural essentialism that
equates the difference between civilian (or literati) and military values with the difference
between Chinese and Hu cultures. In other words, Hu culture is assumed to be distinctly warlike
and Chinese culture to be distinctly peace-loving and scholarly, so a preference for martial
activities over scholarly pursuits is interpreted as evidence of Hu origin or “Hu-acculturation.”
Chen made this assumption explicit in his influential Tangdai zhengzhishi shulun gao 唐代政治
史述論 (1943), arguing of late-Tang Hebei society that “it was fond of warfare and had no
esteem for literature and education; in essence, its gradual Hu-acculturation had gone deep while
its Han-acculturation was only superficial.” 110 Edwin Pulleyblank, whose analyses of the An
Lushan rebellion owe much to Chen Yinke’s theories, has traced the supposed barbarization of

108
Erasmus and the other Renaissance humanists have been described as unsystematic thinkers who preferred
rhetoric to philosophical system-building; if this characterization holds true for them, it surely applies even more to
Han Yu. Nauert, Humanism and the Culture of the Renaissance in Europe, 204–205.
109
Chen, Jinmingguan conggao chubian, 329.
110
Chen Yinke 陳寅 , Tangdai zhengzhishi shulun gao 唐代政治史述論 (Taipei: Taiwan shangwu yinshuguan,
1994 [1943]), 29.
the Tang military to the shift from the fubing 府 militia system to permanent, professional
frontier armies in the early eighth century:

With the formal abolition in 737 of the practice of sending militiamen and other
conscripts back and forth on tours of duty and the enlistment instead of permanent
frontier armies the role of non-Chinese elements in the Chinese defense forces became
more firmly entrenched…. And, there is good reason to think that the Chinese in these
[frontier] armies… were also more or less ‘barbarized,’ that is, they had become deeply
imbued with the military ethos of their barbarian comrades in arms.”111

Likewise Charles Hartman, who was influenced by both Chen Yinke and Pulleyblank, has
written that because “the professional soldiers and mercenaries that replaced the conscripts of
early T’ang were largely of foreign origin,” as a result “the early T’ang equilibrium between
literary and military values dissolved into a dichotomy that came to be defined largely along
cultural lines: to be literary was to be Hua, to be military was to be Hu.”112

Contrary to these arguments, there is not a single text from the eighth and ninth centuries
that suggests that the Tang army had become exclusively identified with foreign-born troops;
instead, numerous sources show that the distinction between fan 蕃 (foreign) troops and Han 漢
(Tang/Chinese) troops remained important to the very end of the dynasty. Moreover, any
professional army is likely to develop a “military ethos” by virtue of the nature of its work, and
the Tang frontier armies of the early eighth century were no exception. If they had a distinct
subculture that differed greatly from that of the literati, this was not neccessarily a product of
influence from foreign soldiers. Nor was this military subculture unique to Hebei, although there
is good evidence that it became particularly strong there because of the political dominance
enjoyed by military men. The court elite’s attitude of insensitivity and disdain toward this
subculture probably contributed to the Hebei military’s desire for independence from court
control—why should soldiers, having tasted autonomy and political authority, willingly
surrender it to haughty civilians in the name of loyalty to the emperor? It does not follow from
this, however, that the Chinese soldiers in Hebei no longer saw themselves as Chinese, or that
the court elite regarded them as barbarians rather than just illiterate boors.

Hartman’s full-length study of Han Yu cites three anecdotes as evidence that the court
elite of Han’s day believed Hebei to have been barbarized into “an independent, separatist,
largely Hu culture.”113 The first anecdote comes from Du Mu’s 杜牧 (803–852) epitaph for Lu
Pei 盧霈 (ca. 809–839). The epitaph claims that Lu grew up in the Hebei military province of

111
Ironically, Pulleyblank began this essay by dismissing the hoary myth that premodern China was “a pacifist, or at
least, a very pacific country,” only to then adopt a version of that myth by equating a “military ethos” with
barbarians. E.G. Pulleyblank, “The An Lu-shan Rebellion and the Origins of Chronic Militarism in Late T’ang
China,” in J.C. Perry and B.L. Smith eds., Essays on T’ang Society: The Interplay of Social and Political and
Economic Forces (Leiden: Brill, 1976), 33, 39–40.

In a footnote, Hartman explains that Hua “means racially Chinese.” Hartman, Han Yü and the T’ang Search for
112

Unity, 122–123, 317 n. 11.

113
Ibid., 123, 125–128.
Chengde 德 without ever hearing of the Duke of Zhou and Confucius or of the world outside
Hebei. He knew nothing but horse breeding, hunting, polo, and warfare until the age of nineteen,
when he heard of “the sage-kings’ Way of classical learning” 王儒学之 and the imperial
capitals (Chang’an and Luoyang) from a Classicist 儒者 named Huang Jian 黃建 (n.d.). Huang
told Lu Pei that classical learning would open the door to high office at court, allowing him and
his descendants to live in affluence and peace. Inspired by this revelation, Lu left his family,
spent ten years teaching himself the Classics in a Daoist monastery, and then went to Chang’an
in 838 to take the jinshi examination. He impressed the capital elite with his extensive
knowledge of Hebei’s strategic geography, and claimed that this knowledge could enable
imperial armies to regain control of the region. Unfortunately, he was killed by bandits while
traveling in 839.114

Both Chen Yinke and Hartman use Lu Pei’s epitaph as evidence of a profound cultural
difference between “barbarized” or “Hu-acculturated” Hebei and the rest of the Tang.115 This
implicitly assumes that Lu Pei’s original way of life was objectively less “Chinese” than that of a
classically-educated literatus, and that a literatus like Du Mu would have perceived it as such.
Both assumptions are flawed. Besides the problem of essentializing “Chineseness” by equating it
with literati culture, it would also be overly simplistic to conclude from Lu’s epitaph alone that
the typical Hebei person was totally ignorant of the classical tradition and the outside world.
Even Du Mu acknowledges that in Zhenzhou 鎮州 (modern Zhengding 正定), the capital of
Chengde, there was a Classicist scholar (Huang Jian) who knew about the imperial capitals and
whom “the townspeople respected and addressed as ‘Master’” 鎮人敬之 呼爲 生. 116 Du
clearly recognized that the people of Chengde still held classical scholarship in high regard in
spite of (or perhaps because of) its rarity in their province. Ma Jizhao has pointed to a passage in
the Tang zhiyan 唐摭言 as evidence that the Lu Pei epitaph greatly overstates the decline of
classical learning in Hebei. The Tang zhiyan passage includes a list of the Tang court’s quotas
for examination candidates from various provinces in the year 845, and indicates that Chengde
(here called by its alternative name Zhen-Ji 鎮冀) and the other Hebei provinces were limited to
eleven jinshi candidates and fifteen mingjing 明經 candidates each. Ma argues that these quotas
would have been meaningless if the Hebei provinces did not produce examination candidates or
only produced a mere handful each year.117

114
QTW 755.7824; Wu Zaiqing 吳在慶 ed., Du Mu ji xinian jiaozhu 杜牧集繫 校註 (Beijing: Zhonghua shuju,
2008), 767–768.
115
Chen, Tangdai zhengzhishi shulun gao, 28–29; Hartman, Han Yü and the T’ang Search for Unity, 125.
116
QTW 755.7824; Wu ed., Du Mu ji xinian jiaozhu, 767.
117
Wang Dingbao 王定保, Tang zhiyan 唐摭言 (Beijing: Zhonghua shuju, 1959), 1.2; Ma Jizhao 馬 照, Hebei
Tangshi dili yanjiu 河 唐詩地理研究 (Baoding: Hebei daxue chubanshe, 2012), 63–68; see also Zhang Guogang
張國剛, Tangdai fanzhen yanjiu 唐代藩鎮研究 (Changsha: Hunan jiaoyu chubanshe, 1987), 84–88. Ma and Zhang
are among a very small number of Chinese scholars who have openly disagreed with Chen Yinke’s theory of Hebei
barbarization. Ma emphasizes the cultural affinity between Hebei literati and Chang’an literati during the ninth
century, while Zhang emphasizes the fact that the imperial court continued to exert political influence over Hebei.
 Moreover, Du Mu’s depiction of Hebei in the Lu Pei epitaph should be balanced against
another depiction in one of his essays, the “Zhanlun” 戰論 (Discourse on Warfare):

Hebei has frugal and simple customs and [its people] do not produce clever but useless
works of craftsmanship. They are artless, tenacious, and strong-willed, equally single-
minded as warriors and as farmers. Its famous cities and its strong fortresses tower like a
mountain chain; its tall mountains and wide rivers are interlocked. Add to that its soil,
which abounds in fast steeds for cavalry charges against the enemy, and one can see why
its people are victorious when they march out to war and prosperous when they remain at
home. Without coveting the products of the empire (literally “all under heaven”), they are
able to cultivate whatever they need, just like a peasant household with a large farm does
not need to own pearls in order to be rich. The empire, on the other hand, cannot do
without Hebei. When [the people of] Hebei become caitiffs (lu, i.e., enemies of the state),
then we are left without fine armor, crack troops, sharp swords, strong bows, and fast
steeds. Should the barbarians (Yi-Di) unexpectedly raise alarms on our four frontiers,
encroach on our territory, and get through our external [defenses] to reach the interior
[provinces], what would we use to fend them off?
河 者 俗儉風渾 不生 樸毅堅強 果於戰耕 城堅壘 峉嶭相貫 高
山大河 盤 交鎖 以土息健馬 便於馳敵 是以出則勝 處則饒 不窺 之
產 自可封殖 亦猶大農之家 不待珠璣然後以爲富也 無河 則不可 河
既虜 則精 銳卒利刀良弓健馬無有也 卒然夷狄驚四邊 摩封疆 出表裏 吾何
以禦之？118

In the Lu Pei epitaph, Du Mu frames Lu Pei’s transformation from hunter-herdsman to literatus

in terms of an escape from parochial ignorance and a life of constant war. In the “Zhanlun,”
however, he idealizes the simple, spartan life of the Hebei warrior-farmer, represents Hebei’s
martial subculture as an essential element in the empire’s defense, and maintains a clear
distinction between rebel “caitiffs” and foreign barbarians. Despite the difference in tone
between these two depictions of Hebei, we can infer quite confidently from both that Du Mu
never associated Hebei’s military subculture with barbarism.

The second anecdote that Hartman cited relates to an incident in 821. After the Hebei
military province of Lulong 盧龍 submitted to direct imperial rule, a group of imperial officials
was sent in to reestablish direct court control. Some of them reportedly antagonized the Lulong
troops by denigrating them as “renegade caitiffs” (fanlu 反虜) and mocking their illiteracy.119
Hartman translates the phrase fanlu as “rebel barbarians,” while Pulleyblank earlier translated it
as “rebellious savages.”120 Both translations overlook the fact that in the Tang empire, the label
lu could be applied to any enemy of the state, whether Chinese or foreign—we see this, for
example, in the passage just quoted from the “Zhanlun.”121 Moreover, illiteracy in Tang armies

118
QTW 754.7813; Wu ed., Du Mu ji xinian jiaozhu, 649.
119
JTS 129.3611; XTS 127.4448; ZZTJ 241.7793.
120
Hartman, Han Yü and the T’ang Search for Unity, 126; Pulleyblank, “The An Lu-shan Rebellion,” 50.
was hardly unique to Hebei, nor was illiteracy unique to the military in pre-modern societies.
This example is thus irrelevant to the issue of barbarization.

The third ancedote is more persuasive than the others, but still problematic nonetheless. It
comes from an epitaph written by Liu Yuxi 劉禹錫 (772–842) for the spirit road stele (shendao
bei 碑) of Shi Xiaozhang 史孝章 (800–838).122 Xiaozhang was a military officer’s son who,
like Lu Pei, embraced classical scholarship as a young man, earning him the nickname of
“scholar” 書生 from his friends in the Wei-Bo 魏博 military province.123 In 822, his father Shi
Xiancheng 史憲誠 (d. 829) led the Wei-Bo troops in a mutinous attempt at forcing their
governor Tian Bu 布 (785–822) to reassert the province’s autonomy instead of following the
imperial court’s orders to make war on neighboring Chengde. Tian Bu, who was loyal to the
imperial court and also obliged to avenge his father’s recent murder by the Chengde army, chose
to end his dilemma by falling on his sword. Since the Tang court lacked the means to punish the
Wei-Bo army’s mutiny, it had to confer official approval on the mutineers’ choice of Shi
Xiancheng as their new governor.124

According to Shi Xiaozhang’s epitaph, he eventually decided to persuade his father to

submit to the court for the sake of restoring Hebei’s honor and preserving their family’s future.
Xiaozhang went down on his knees and made the following appeal, addressing his father like a
minister remonstrating with his ruler:

Your subject humbly observes that here to the north of the great river (i.e., the Yellow
River), our land is rich and our troops are strong, yet the worthy men of the empire
(literally “all under heaven”) hold us in contempt and look upon the Heshuo region (i.e.,
Hebei) as though we were barbarians. Why is this so? It must be because most of the men
who control this territory acquired it through strategic opportunism, and not by just and
moral means. Now your subject’s parents have received titles and fiefs [from the court]
and risen to high status; the emperor has extended the fullest measure of grace to our
family. If we do not thoroughly change our ways and improve our conduct to
demonstrate our trustworthiness clearly to the court, we will have no way to silence the
mockery of wise men and recognize our enlightened ruler’s intentions [toward us]. Once
our steadfast loyalty is widely known, our family’s fortunes can be made to last. If we

121
On this see also Christopher I. Beckwith, Empires of the Silk Road: A History of Central Eurasia from the Bronze
Age to the Present (Princeton, NJ: Princeton University Press, 2009), 360. Note that I do not agree with Beckwith’s
larger argument that the translation “barbarian” is not applicable to any Chinese word because none of the generic
labels for foreign peoples in Chinese was pejorative. The label Yi-Di was clearly derogatory in a way that Hu and
Fan were not.
122
According to the Zizhi tongjian, Shi Xiaozhang was originally named Shi Tang 史唐 and changed his name to
Xiaozhang in 829. ZZTJ 244.7864.
123
JTS 181.4686; XTS 148.4790; Qu Tuiyuan 瞿蛻園 ed., Liu Yuxi ji qianzheng 劉禹錫集箋證 (Shanghai:
Shanghai guji chubanshe, 1989 [hereafter LYX]), 3.99.
124
JTS 141.3852–3853, 181.4685–4686; XTS 148.4785–4786, 210.5935; ZZTJ 242.7806–7807.
 seize this timely opportunity, then we can avert a disaster that would otherwise befall us
as swiftly as turning on one’s heels.
臣竊 大河之 地雄 精 而 賢士心 之 目曰河朔 視猶夷狄 何也？
蓋有土者多乘 機際會 非以義取 臣家 母封 爲貴門 君 至矣 非痛
折節礪行 彰信於朝 無以弭識者之譏 寤明君之意 節著於外 福 於家 乘
125
時蹈機 禍不旋踵

The Xin Tangshu biography of Shi Xiaozhang contains a paraphrase of these words and is clearly
based on the epitaph by Liu Yuxi, which was included in Liu’s collected works (the original stele
is lost). 126 Liu Yuxi claims that by the time Shi Xiaozhang finished speaking, tears were
streaming down his face. 127 But the whole conversation during which Xiaozhang supposedly
made this argument may be a product of Liu Yuxi’s imagination.128 The conversation is not
mentioned in another epitaph for Shi Xiaozhang (composed by one Li Jingxian 李景 ) that was
excavated from his tomb in 2004. That excavated epitaph states instead that during the tense
period immediately after Shi Xiancheng’s rise to the governorship, Shi Xiaozhang secretly sent
letters to the imperial court to affirm his loyalty to the emperor, an act for which he was
rewarded with three titles, including that of Deputy Governor 副節度使.129 This contradicts the
epitaph by Liu Yuxi, which states that the imperial court rewarded Shi Xiaozhang with these
titles after the emperor heard of his tearful remonstrance with Shi Xiancheng and “praised him
saying, ‘He (i.e., Shi Xiancheng) has a [good] son indeed” 嘉之 曰 彼真有子 .130 Liu
Yuxi may have felt that Shi Xiaozhang’s secret communications with the court made him
vulnerable to criticism as an unfilial son, and therefore invented the remonstrance episode in
order to create an image of filiality.

125
LYX 3.99–100. The epitaph places this incident between 822 and 828. Shi Xiaozhang’s Xin Tangshu biography
names the emperor at the time as Wenzong 文 (r. 827–840) and goes on to speak of the imperial court’s punitive
expedition against another Hebei governor, Li Tongjie 李 捷 (d. 829), in 827–829. This implies that Xiaozhang’s
remonstrance occurred in 827 or 828. The Zizhi tongjian seems to conflate Shi Xiaozhang’s remonstrance on this
occasion with a slightly later case of remonstrance against Shi Xiancheng’s intent to side with Li Tongjie against the
imperial court. The Xin Tangshu separates the two instances of remonstrance, while Shi Xiaozhang’s epitaph and Jiu
Tangshu biography do not mention the second occasion. See JTS 181.4687; XTS 148.4790; ZZTJ 243.7858.
126
Shi Xiaozhang’s Jiu Tangshu biography does not quote his words to Shi Xiancheng. JTS 181.4687; XTS
148.4790.

The various versions of this account all emphasize Shi Xiaozhang’s weeping as a sign of his sincerity. LYX 3.99–
127

100; JTS 181.4687; XTS 148.4790; ZZTJ 243.7858.

128
Deng Xiaonan’s use of the epitaph to draw inferences about Shi Xiaozhang’s “subjective self-identity” 自 主觀
and understanding of the Chinese-barbarian dichotomy thus seems overly simplistic and unmindful of the
text’s genre. Deng Xiaonan 鄧 南, Zuzong zhi fa—Beisong qianqi zhengzhi shulue 祖 之法里 宋前期政治述
略 (Beijing: Sanlian shudian, 2006), 85–86.

Guo Maoyu 郭 育 and Zhao Zhenhua 趙振華, “Tang ‘Shi Xiaozhang muzhi’ yanjiu” 唐
129
史孝章墓 研究,
Zhongguo bianjiang shidi yanjiu 中國邊疆史地研究 17.4 (2007), 116.
130
LYX 3.99–100 (cf. JTS 181.4687; XTS 148.4790). The Quan Tangwen version of the epitaph changes 有子 to 孝
子, so that the sentence means, “He is truly a filial son”: see QTW 609.6153.
 Moreover, although Shi Xiaozhang’s family certainly had foreign ancestry (their specific
ethnicity is disputed), Liu Yuxi quotes him as blaming unethical behavior, not ethnocultural
differences, for Hebei’s purported image as a quasi-barbarian region.131 Therefore, Liu Yuxi’s
piece alone cannot prove that in the 820s or 830s, there was a general perception of Hebei as
culturally or ethnically “barbarized.” Yet the Shi Xiaozhang epitaph seems to have inspired the
eleventh-century compilers of the Xin Tangshu to begin and end its section on the autonomous
military provinces with claims that the Hebei governors “caused their people to see themselves
as Qiang and Di [barbarians]” 使 人自視 羌狄然 and “made barbarians of their people” 夷狄
人 by resisting imperial authority.132 In fact, these lines go even further than the epitaph by
claiming that the people of Hebei perceived themselves as barbarians and actually became
barbarians. This may be a result of their projecting Northern Song concerns about the Kitan-
ruled Yan 燕 region onto the context of late-Tang Hebei; the phrase “made barbarians of their
people” also suggests some inspiration from the rhetoric of the “Yuandao.” Mou Zhenyu has
recently pointed out that Chen Yinke’s theory of Hebei barbarization was directly inspired by
these passages in the Xin Tangshu.133 In that case, the theory actually originated in the eleventh
century, and Chen Yinke merely reintroduced it to modern historiography. But that does not
make the theory any more historically accurate; the evidence for it still ultimately comes down to
a single line of reported speech in a ninth-century epitaph, probably taken out of context.

There is an irony here that may be worth pointing out. While Chen Yinke’s theory of
Hebei barbarization was indirectly inspired by Han Yu’s rhetoric of barbarization via the
medium of the Xin Tangshu, he was also one of the first modern Chinese historians to apply a
“culturalist” mode of analysis to the question of ethnic identity in premodern Chinese history. An
argument that Chen made in the Tangdai zhengzhishi shulun gao has acquired doctrinal status in
Chinese scholarship on the Northern Dynasties as well as the early Tang:

[W]hen it came to the difference between Han people and Hu people, under the Northern
Dynasties, culture (wenhua 文 ) was more important than race (xuetong 血統, literally

131
Shi Xiancheng’s Jiu Tangshu and Xin Tangshu biographies describe his ethnic origin as Kai/Qay 奚, but it was
long suspected, on the basis of his surname, that he was actually a Sogdian. The excavated tomb epitaph of Shi
Xiaozhang has further complicated the issue, as it suggests that Xiaozhang identified himself with both the Türk
royal clan of Ashina and a certain “Ge clan” 葛氏. Guo Maoyu and Zhao Zhenhua have read the epitaph as evidence
that Shi Xiancheng and Shi Xiaozhang were Türks, Zhang Long has read it as evidence that they were Sogdians
passing off as Kai/Qay, and Yin Yong has read it as evidence that they were Uyghurs passing off as Sogdians. One
ultimately cannot choose between these different theories, due to the impossibility of identifying the “Ge clan” with
any degree of certainty. See Guo and Zhao, “Tang ‘Shi Xiaozhang muzhi’ yanjiu,” 117–118; Zhang Long 張龍,
“Shi Xiancheng fuzi yu fanzhen Wei-Bo—yi ‘Shi Xiaozhang muzhi’ wei xiansuo” 史憲誠 子 藩鎮魏博里以
史孝章墓 爲綫索, Minzushi yanjiu 民族史研究 10 (2011), 32–34; Yin Yong 尹勇, “Tang Wei-Bo jiedushi
Shi Xiancheng zushu zai yanjiu—jianlun ‘fan Sute’ wenti” 唐魏博節度使史憲誠族屬再研究里兼論 泛粟特
問題, Shoudu shifan daxue xuebao 首都師範大學學報 2010(4), 17–22.
132
XTS 210.5921, 213.6021.

Mou Zhenyu 牟振 , “Chen Yinke ‘huhua shuo’ youlai tanxi” 陳寅

133
胡 說 來探析, Xueshu tansuo 學
術探索 2008(2), 95–98.
 “bloodline”). All Han-acculturated (Hanhua 漢 ) people were seen as Han people, and
Hu-acculturated (Huhua 胡 ) people were seen as Hu people, without any regard for
what their race was.134

Yet Chen Yinke himself was unaware that this “culturalist” mode of discourse about
Chineseness did not exist before Han Yu invented it in the “Yuandao.” 135 “Culturalist”
intepretations of the Chunqiu and other Classics had become so pervasive and influential by
Chen’s day that he, like many others, assumed Han Yu was simply restating an age-old belief
about the culturally-determined nature of Chineseness.

134
Chen, Tangdai zhengzhishi shulun gao, 19.
135
Compare Chen’s argument to a very similar claim that Qian Mu made in 1948: “We can thus say that in ancient
thought, there was indeed a different standard [from today’s] for distinguishing between the barbarians and the
Chinese: not race (xuetong) but culture. As they said, ‘When a feudal lord uses barbarian ritual, then he is seen as a
barbarian; when barbarians advance to the level of the Central Lands, they are seen as part of the Central Lands.’
This is clear evidence of taking culture as the difference between Chinese and barbarians. What I call culture here,
when spoken of in concrete terms, is just a way of life and a mode of government.” Note how Qian Mu quotes Han
Yu’s “Yuandao” (without attributing the quotation to Han) because he could find no earlier statement of the
“culturalist” discourse that was anywhere near as coherent as this—which is hardly surprising, since that discourse
did not really exist before the “Yuandao.” Qian Mu 錢穆, Zhongguo wenhuashi daolun 中國文 史 論 (Shanghai:
Sanlian shudian, 1988 [1948]), 35.