On the sixth day of the eighth month in the autumn of the fifteenth year [284 CE.] a Prince from [the Korean state of] Baekje called Araki came before the court and presented two fine horses to the [Japanese] emperor.… This Araki was very good at reading the [Confucian] classics.… Hearing this, the emperor asked Araki, “Do you possess a fine Confucian professor [in Baekje]?” Araki replied, “There is one called Wani, he is excellent.” Arata Wake and Kamunaki Wake were dispatched to Baekje to get Wani. In spring in the second month of the sixteenth year Wani arrived. Prince Uji no Waki Iratsuko took him as his teacher. He learnt various classics from Wani. There were none of them he could not master. Wani became the first keeper of the imperial books.Footnote 1
So began the history of Confucianism in Japan. Confucian professors were fine gifts for princes who could use them as symbols of status and connection. As a prince would show his prowess and exhibit his own status by skillfully riding an especially fine stallion, so too a prince “able to master” all the Confucian classics, taught by a fine teacher, exhibited not only the status of “having” that teacher, but also his own accomplishments in being able to “master” the material.
This passage seems to be the first Japanese historical source narrating Confucianism's arrival in Japan. Whether this occurred in 284, as the source could be literally interpreted, or 402, as Peter Kornicki has suggested, or whether this story is a work of complete fiction from the late seventh or early eighth centuries, when Nihon Shoki was compiled, we will never really know.Footnote 2 What is evident is that Confucianism—both its personnel and texts—were initially perceived primarily as part of the status symbolism of East Asian interstate relations.
The beginnings of Confucianism in Japan were thus closely intertwined with the beginnings of the Japanese state itself. The formation of a single, dominant state in central Japan occurred concurrently with a new wave of importation and institutionalization of political and religious culture from mainland Asia, most notably from states on the peninsula now known as Korea. The impact of, most importantly, Buddhism, but also Confucianism and other religious ideas associated with Daoism, accelerated dramatically during the seventh century, coinciding with the rise of the Tang state (618–907) in China and that state's push into Korea. These religious, cultural, educational, and administrative paradigms provided many of the sociological tools necessary for the construction of a more complex centralized state capable of projecting and holding power over a large area.Footnote 3 Even the core symbols of the Japanese Emperors, the sword and mirror, arrived as part of this transmission of Confucianism.Footnote 4 In Japan, the formulation of Confucianism and the formation of the state were concurrent and symbiotic processes. Confucianism was part of the processes that formed the early Japanese state itself, and conversely, these processes of state formation also helped to shape the particular, early Japanese manifestation of Confucianism.
In this article I argue that the way in which Confucianism was institutionalized and reimagined in the early Japanese state limited its efficacy as a form of imperial legitimation. This state, although otherwise modeled on continental forms, lacked one of the main ritualistic schemes that early Chinese and Korean imperial and kingly states used to legitimize the rule of sovereigns. Through most of history, and certainly from China's Later Han dynasty onward (and therefore throughout the Common Era), Chinese empires positioned Confucian ritual as a central part of the symbolic schemes through which the power of the sovereign was legitimated. The dysfunction of this ritual system in the Japanese context, which this article argues was a result of the nature of cultural transmission and reproduction, is one reason Japan, through most of its premodern history, was not ruled by its emperors and imperial institutions. To put it differently, analyzing the nature of the transmission and reproduction of Confucian culture is one way to understand the choice made by the real (non-emperor) rulers of Japan from the tenth century onward to not bother with usurping the emperorship, and instead rule through alternative titles and institutional frameworks of their own making, like kanpaku and shogun.
We also find here a complex historical case for considering the bifurcation of ritualistic and political influence in sacred kingship more generally. The disengagement of ritual practice from political power in traditionally sacred kingships has been described in a variety of societies in widely different places and periods.Footnote 5 The phenomenon is often explained in terms of local cultural imperatives, or the teleological development of religious or political history. Here I will focus instead on the mechanics of cultural transmission and reproduction as key.
Scholars have long been interested in the role of Confucian ritual in the political legitimation of Chinese imperial states. Recently, however, historians of early China have more lucidly theorized the role of ritual. In rejecting standard late twentieth-century theories like those of Clifford Geertz and Maurice Bloch, which portrayed ritual as either a reflection of social reality or as something outside human experience, Michael Puett has argued that ritual in early China, rather than a reflection of social relations, was a deliberately constructed “as if” world “where one [could] act as if a certain situation was the case” in an attempt to change that reality.Footnote 6
Puett uses examples from the Confucian Book of Rites to suggest that rituals for the worship of imperial ancestors in ancient China were used to imagine or create relationships between members of the ruling family, the ruling class, and the population at large. These rituals, by emphasizing the disjunction between reality and ritual, looked to inculcate appropriate social relations: “The goal of the ritual is to create a harmonious, hierarchical relationship between them of ancestor-father-son. Underlying the ritual, of course, is the clear knowledge that this is not the way the world operates.… The entire realm is thus organized in such a patriarchal form, with Heaven above, and the ruler below as son of Heaven and parent to the populace.”Footnote 7
This vision of ritual's role in early East Asian polities thus sees it not as reflecting reality, even for the participants, but rather as “training one's disposition to act properly” in an attempt to transform social relations by deliberately playing in a mode that is not real.Footnote 8 The acknowledgment of the unreality of ritual by practitioners and commentators in the ancient Chinese context is a vitally important point, and is reflected in the Confucian classics themselves. Confucius Analects is explicit regarding the “as if” nature of rites in the famous passage: “The Master sacrificed to the dead as if they were present. He sacrificed to the spirits, as if the spirits were present.”Footnote 9
This acknowledgment of disjunction from reality was what empowered these rituals to affect or work on the real world. It gave them their political utility. Overt reference to the politically utilitarian nature of the rites can also be found in the classics, for instance in sections from the Rites of Zhou, a text that provided much of the background of the legitimation rites mentioned in the Tang Chinese administrative codes that I will examine in this essay's second halfFootnote 10: “Sacrificing to [worshipping] the Five Emperors [this rite] is the taking in hand of the hundred officials [the nobility bureaucracy] and the utilization of them.”Footnote 11
As we will see, in Japan, by contrast, the ritual schemes did not facilitate acknowledgment of their unreality or political utility. This is significant because such acknowledgment was a large part of what allowed political culture and government practices to survive through the regular processes of dynastic transition and lineage slip that characterized premodern kingly states. The community's and the ritual scheme's granting the unreality of the ritual process in the Chinese imperial context allowed new emperors not related to the previous emperor to still be portrayed as if they were the son of Heaven. This was so even when everyone knew they were not biologically related to the previous progeny of Heaven, and therefore were obviously not actually the son of Heaven. The disjunction between the heavenly and terrestrial realms meant they could still be legitimated as emperor through the ritual of seeing them as the son of Heaven, regardless of whose child they actually were organically on earth. The Confucian ritual form's acknowledgment of its own unreality, as in the Chinese tradition, thereby enhanced its ability to play a transformative political role, including in dynastic transition. This meant that many Chinese dynasties, even new ones, could enjoy a relatively strong continuity in political culture and governing practice, and in the state's personnel, knowledge, and security apparatuses.
The process through which the imperial statutory (ritsuryō) state was formed in Japan, on the other hand, excluded these elements of Confucian ritual and replaced them with non-Confucian Japanese rites (later sometimes identified as “Shinto”).Footnote 12 These “indigenous” emperor rites were themselves new, having probably been formulated in Japan within the context of the formation of the statutory state (after the Jinshin war of the late 600s). Compared to the Confucian rites of imperial legitimation seen in the Tang and other Chinese imperial states, however, these Japanese rites were more organic, positing the emperor as the actual descendant of Heaven.Footnote 13 This made them harder for new rulers to coopt, which meant they had less continuous political tradition at their disposal for legitimating their rule.
This article's first part looks at the significant influence Confucian ideas of civilization and rites had on the conceptualization of external relations in Japanese state-making. It confirms a significant Confucian resonance in state practices, which justified a dominant and exclusive Japanese state in the archipelago. The second, main part of the article examines the significant lack of influence Confucian ritual had in the early Japanese state's internal justification of emperor rule. I employ an analysis comparing early Japanese state codes with their parallel Chinese texts to argue that the weak integration of Confucianism into the Japanese statutory state model meant that core elements that justified the emperor's rule in continental models were missing from the basic models of ritual institutionalization that continued throughout premodern Japanese history.
CONFUCIANISM AND THE STATE'S EXTERIOR: CIVILIZATION AND RITES
There is no doubt Confucianism significantly affected not only how the early Japanese state was conceptualized, but also how that conceptualization was expressed in its first institutionally produced state political texts. An example is the Kojiki, one of the two core early texts legitimizing the Japanese state. Though it is the most linguistically un-Chinese of the two, its introductory section is modeled exactly on the style of the introduction to the Wujing zhengyi 五經正義, a collection of the Confucian classics from the Tang period.Footnote 14 Its introduction thus mimics the form of the most famous Chinese Confucian commentary of the Tang period. The structure of the other core early text, the Nihon Shoki, is unmistakably based on Chinese dynastic histories, which by this time were outliers of the Confucian canon. In this sense, the very form of the Nihon Shoki was highly derivative from Confucian textual tradition.
The content of these texts also reveals how conceptual paradigms at the heart of Confucianism affected Japanese state formation. The conceptualization of a state, particularly the logic behind delineations of outer and inner, and the interaction of this logic with larger questions relating to the place of humans in the natural world, seems to have been derived primarily from Confucian paradigms. The formation of the early Japanese state occurred through a process of military, political, cultural, and religious interaction with Korean kingdoms in the mid-first millennium CE.Footnote 15 They were related to the emergence of enduring state identities, state violence (war), and early codification of principles of governance which, while certainly not “laws” in the modern sense, had enduring effects on the later development of state institutions including legal ones. Positive Confucian influence on this codification needs to be reconciled with the violence inherent in the consolidation of the early state. Thinking about the history of the early Japanese state thus invites reflection upon the contradictions inherent in the utility of Confucian universalism to statecraft at a broader level, not just in early Japan, but also in later periods and in other parts of East Asia and Southeast Asia.
Confucian influence in early Japanese statecraft manifested itself in two opposing directions. On one hand, Confucian ideas were drawn upon to create frameworks for mediation and consensus building in Japanese society, and between Japanese and foreign peoples. On the other, they helped model a hierarchy of societies that justified Japanese state violence against so-called “barbarians.”Footnote 16 Scholars of Chinese, Korean, and Japanese intellectual history often emphasize Confucianism's message of shared human values and respect for the other. Scholars in Japan have repeatedly cited Shōtoku's Seventeen Article Constitution (also called The Seventeen Injunctions) as a strong example of this. This is a list of principles of governance that tradition claims was written in 604 by the imperial regent Prince Shōtoku. In the centuries thereafter, Prince Shōtoku was discursively transformed into a kind of saint.Footnote 17 Various Buddhist traditions claimed him as a bodhisattva, and he has been revered by those of all political colors since. His Constitution has been one of the most heavily referenced treatises in Japanese political history right into the twenty-first century.Footnote 18 This is partly because of its flexible nature: it is hortatory rather than regulatory, consensus driven, and religiously pluralist. Shōtoku's Seventeen Article Constitution is correctly cited both as representative of the Buddhist nature of the early Japanese state and as one of the most clearly Confucian-influenced texts in early Japanese history. This is indicative of the intellectually and religiously pluralist nature of this text, but also of Japanese political culture at this time more generally.
The main claim modern commentators have made about Shōtoku's Seventeen Article Constitution, however, is that it represents a “Japanese” idea of “harmony.” How exclusivist the “Japaneseness” of this harmony is imagined to be has usually depended on the political persuasion of the scholar and the moment in modern Japanese history they worked within. Imperialist and nationalist Japanese scholarship in the mid-twentieth century valorized this as an exclusionary “Japanese value.” Post-World War II Japanese scholarship challenged these readings by emphasizing the comparatively universalist perspective of the work and clear Confucian influence on it. Intriguingly, this later, more politically correct scholarship did not actually challenge the valorization of early Japanese state codes apparent in the earlier nationalist readings, but instead simply relocated the valorized ideas to a transnational Asian discourse rather than an exclusionary national Japanese one. This “Asianizing” of a previously “nationalist” narrative of early state formation or codification allowed that process to be discussed within the same positive value judgment as before.Footnote 19 What is undeniable is that the Constitution's references to harmony come from Confucian sources.
The very first sentence of the first of the seventeen injunctions of this text is a line from the Confucian classic The Book of Rites, which also appears in Confucius Analects: “Harmony is to be valued, and contentiousness avoided.” The injunction continues, “When those above are harmonious and those below are conciliatory and there is concord in the discussion of all matters, the disposition of affairs comes about naturally.”Footnote 20 This injunction is quintessentially Confucian in the sense that it advises for a form of rule in which the use of force is unnecessary. Traditional understandings of Confucianism in both Shōtoku's time and ours have been primarily based on conceptions of Confucian values that came into being in Han dynasty China (206 BCE–220 CE).Footnote 21 Han dynasty China classically defined Confucian values as humanitarian and conciliatory in opposition to the rule- and force-based political values of the Chinese Legalist tradition of thinkers like Han Fei (280–233 BCE), the latter being associated with the brutal and short-lived Qin dynasty (221–207 BCE) that the Han dynasty replaced. This classical interpretation of the Confucian tradition sees it as recommending governance through ritual that conciliates, as opposed to the Legalist tradition that sought to govern through rules backed up with coercive violence. This injunction, therefore, begins Shōtoku's Seventeen Article Constitution by unambiguously identifying it with the Confucian rather than the Legalist tradition and style of governance in East Asia. In other words, the work not only adopts Confucian terminology and quotes Confucian texts, but also uses these in ways that follow standard Confucian political discourse patterns dating from the Han, which claim to have a relatively conciliatory, harmonizing character.
This preference for rites as a preferred method of rule is also emphasized in the fourth injunction. After quoting from another Confucian work, the Classic of Filial Piety, “Rites must be the basis of governance,” it goes on to conclude, “If the common people have rites, then the state will govern itself.”Footnote 22 The emphasis on the centrality of cultivating mediating relationships is represented again in injunction nine, which opens with a quote from Confucius Analects, “Trust is the basis of justice,” and concludes with, “If there is trust between sovereign and vassal then nothing cannot be achieved, if there is no trust then all will be destroyed.”Footnote 23 A similar emphasis on conciliatory human relations, this time between rulers and commoners, is presented in Injunction Sixteen, which opens with the quote from Confucius Analects, “The common people should be employed according to the season.” This reference to the seasons is a warning to rulers not to demand corvée labor from the peasants in times of agricultural labor intensity such as harvest, because to do so would interfere with the peasant's livelihood.Footnote 24
All of these injunctions share a characteristic of warning members of the ruling elite to emphasize conciliation and harmony in their relationships with others, including by being aware of others' needs, even those of peasants. For the rulers, this implied that they should moderate their use of coercion and force in exercising power. These injunctions all plainly take Confucian textual sources as their bases, and make points that could indeed be characterized as representing basic Confucian approaches to social governance.Footnote 25
Confucian universalism had another side, however. The moderating and civilizing aspects of Confucian influence went hand in hand with a Confucian worldview that demarcated different human societies in a distinct hierarchy. By establishing a single, universalist, cultural idea of “civilization” upon which human societies could be comparatively judged, Confucianism recognized the possibility of grading human societies, with a central civilized state at the top and barbaric peripheries at the bottom. This allowed the Confucian idea of “civilization” to be deployed to justify conquests of peripheral “barbarian” peoples and states. Such justifications can be found all though the classic Confucian texts. In the context of the Japanese archipelago, the Yamato state of Prince Shōtoku saw itself as the civilizing center. It used the same paradigms and language as found in the Confucian classics to validate its wars of conquest against other peoples in the archipelago, and the taking and trading of these peoples as slaves. Non-Japanese, which at this time meant any peoples on the archipelago unwilling to submit to the authority of the Yamato sovereign, were referred to in Nihon Shoki as “barbarians,” using the same phrasing employed in the Chinese Confucian classic The Book of Rites.Footnote 26
Early Japanese state documents also mimic Confucian tradition in narrating the world in terms of a unipolar imperial order of civilization. They narrate the conduct of state ceremonies involving subjugated “barbarian” peoples and surrounding states along the lines of this logic.Footnote 27 In this sense, Confucian universalism was used in Japan, as in China, to justify ideas of cultural superiority and military domination. Importantly, this justification was often achieved, as in the example just given, by the use of ritual forms lifted from the Confucian Book of Rites.
The rewriting of Japanese history since World War II, including its highly critical approach to Japanese nationalism and militarism, has not led to much soul searching regarding this role of the Confucian tradition. That is because most of the critique of the earlier modern Japanese nationalist and imperialist historiography has come from the perspective of the postcolonial nation states, most notably Korea. Postwar perspectives on early Japanese history thus emphasize the role of Korean states, the interaction of Japan with Korea, and Korean influence on Japan, and this is seen as sufficient revision of the old nationalist historical outlook. But these new historical perspectives pay little attention to the people who fell between the boundaries of the modern nation states and their premodern forerunners—“barbarians”—who did not belong to premodern Korean, Japanese, or any other states. Postcolonial history has not required much criticism of the Confucian world order in relation to premodern Korea and Japan, because premodern Japanese texts recognized and acknowledged Korean kingdoms by name. “Barbarians,” commonly referred to as emishi in period sources, was a signifier that identified only someone not part of a recognized kingly state, meaning either Japan or Korea.Footnote 28
Most historians agree that the words used to identify Japanese and non-Japanese in the early history of the archipelago, rather than relating to racial or ethnic identity, were primarily political categories. Emishi or “barbarian” in its premodern usage cannot be identified as a signifier of non-Japaneseness, because many non-Japanese, especially subjects of Korean kingdoms, were not called “barbarians.” Neither did “barbarian” signify non-Japanese in a modern ethnic sense, because there are many examples in the historical sources where Japanese, through a change in political affiliation, quickly “became” emishi, or vice-versa. Those submitting to the authority of a kingly court, be it Yamato, Silla, Baekje, or another, were civilized; those not subjugated by one of these courts were barbarians. In the context of the Japanese archipelago, this meant that anyone not submitting to the Yamato court was characterized as what would later be interpreted to mean non-Japanese. Culture and civilization in the Confucian sense, and indeed the idea of harmony, were thus rooted in a brutal politics of submission that transcended (or did not imagine) ethnicity, but also institutionalized and legitimized, as never before, set relations of clan, class, and state power and the violence inherent therein.Footnote 29
Postwar social history in Japan has shown that many people in premodern societies in Japan fell between state boundaries or sat outside of them. Historians like Amino Yoshihiko have provided considerable data and narrative about such people. The classic example given by Amino is of “people of the sea,” who made their living primarily on the sea and often moved around.Footnote 30 But intellectual historians have been slow to think about how the history of these peoples can be reconciled with the rise of Confucian-inspired state structures and political discourses apparent throughout Japanese history. As we have seen, Confucian ritual, and indeed the Book of Rites directly, provided the early Japanese state with examples of how to justify state violence projected outward. What then of Confucian ritual's role in justifying the state itself, and particularly in granting internal legitimacy to its sovereign rulers?
CONFUCIANISM INSIDE THE STATE: THE POVERTY OF ORGANIC RITUAL
Confucianism inside the Japanese state was institutionalized in a particularly passive, peripheral, and ritualistically weak position. This was closely related to the high level of pluralism of Japanese society and its state institutions at this time. Japanese state institutionalization of Confucianism was effected under direct influence from Korea, mainly by drawing on institutional and legal examples from Sui and Tang dynasty China. The Sui dynasty (589–616), and particularly the very successful Tang dynasty that followed it, were especially pluralist Chinese states. Although nearly all periods in Chinese history saw massive religious pluralism, in most periods from the beginning of the Common Era onward Confucianism held a preeminent position in the systems of state ritual practice that constituted the empire's primary ideological apparatus. In the Han period, when many Chinese institutions were formed, the state elevated Confucianism to a position of orthodoxy above other religious traditions.Footnote 31 This superior position of Confucianism, though relinquished to some extent during the Sui and Tang dynasties, was further developed in the Song (960–1276), Yuan (1271–1368), Ming (1368–1644) and Qing (1644–1911) dynasties. Through most of Chinese history, then, particularly in the earlier and later dynastic states, Confucianism held a privileged position in the state. Exceptions were the Sui, which favored Buddhism, and the Tang, which was unusual in advancing a range of other traditions including Daoism and Buddhism. These shifts resulted in a comparatively diminished status for Confucianism during those periods.
That the Tang, despite being considered one of Chinese history's golden ages, was an exception to the rule of Confucian supremacy has been attributed by some scholars to other exceptional qualities of this state. Tang society's approach to gender is seen as different from earlier and particularly later periods. Women were more visible in leadership roles throughout the various strata of elite Tang society, and representations of noble women in Tang literature and art often portray a much more robust and active vision of femininity.Footnote 32 The Tang state is also generally regarded as having been unusually culturally open and diverse. The dynasty's ruling house “had roots in the Turkic peoples.”Footnote 33 It looked west and was more economically, socially, and culturally integrated with Central Asia than previous Chinese states had been. Experts on Daoism have related this cultural characteristic of the Tang state to the high level of Daoist influence within it, which to some extent came at the expense of Confucianism.Footnote 34
The Tang period is important to Japanese history because of not only the influences on Japan that emanated directly from Tang China, but also the level of Tang influence on the Korean peninsula, particularly the Korean state of Silla, the dominant power in southern Korea during the seventh century.Footnote 35 From the sixth century, the Silla state identified itself with Buddhism, as did its northern Korean neighbor Goguryeo. This meant that all the major continental states in touch with Japan in the seventh and eighth centuries, Sui and Tang China, Silla, Goguryeo, and Bohai all identified with a non-Confucian tradition, usually Buddhism.Footnote 36 Herman Ooms has argued that the influence of continental Daoism was equal to if not greater than that of Buddhism in eighth-century Japan. Direct references to Daoism in the Japanese historical texts are, as Ooms puts it, “elusive.” But he argues nonetheless that many elements in Japanese society often attributed to Shinto or Confucian influence are actually Daoist “deposits,” which can be shown to have a deep Daoist history.Footnote 37 He links the strong position of Daoism in Tang dynasty China to the particular religious plurality of early Japan.
Thus early Japanese state structures formed at a moment in Chinese history when Confucianism was in many ways at its institutionally weakest. In Japan at this time, however, the emperor and imperial institutions were even weaker than during the Tang.Footnote 38 Because these institutions were the locus of Confucian activity, this limited the capacity of Confucianism to be developed or utilized critically in society. As Marian Ury has written, “A bureaucratic system carefully modelled after that of the Tang reached its apogee in the eighth century … by the beginning of the … tenth century, although the conception and rhetoric of Confucian government remained, as did its forms and usages, many of its functions were being carried out by other means.”Footnote 39
In what follows, I will examine primary sources from that eighth-century period of “apogee” that Ury mentions, when Japan formalized administrative and legal codes based on those of Tang China. These codes suggest that, even from the beginning of this formalization process, Confucian ritual was sequestered. Although Confucian ritual was relatively weak in the Tang, it was still the basis of imperial legitimation articulated in the Tang state codes, and in the related ritual practices prescribed in those codes and carried out in the inner sanctum of the imperial palace. But this was not the case in Japan.
Comparison of Ritual in Tang and Japanese Administrative Codes
The most influential legal and administrative codes of early Japan, the Yōrō Ritsuryō 養老律令 (757), closely followed continental examples from Tang China. The ritsu (Ch. lü) sections, or criminal codes, of the Yōrō Ritsuryō in most respects match Tang examples, while the ryō (Ch. ling) sections, or administrative codes, sometimes diverge considerably. Experts have judged that the most likely continental model for the administrative codes was the Tang dynasty code Yonghui ling 永徽令 (651).Footnote 40 Though that code is no longer extant, comparison is possible because there do exist versions of significant elements of other, closely related Tang codes. In the early twentieth century, Japanese Sinologists reconstructed a Tang code that they thought most closely matched those used in the construction of the Japanese state regulatory documents.Footnote 41 In the comparison below, that 1930s reconstruction is my primary referent for the Tang codes, but I have also cross-referenced it with another version, the Tang Kaiyuan ling 開元令 (737), which is the code produced closest in time to the mid-seventh-century dates usually associated with the completion of the Yōrō Ritsuryō.
My comparison will concentrate on regulations related to the conduct of central imperial rituals involving the emperor. These regulations appear in parallel articles located in the same position in both codes, although they have different titles. In the Tang code, this section is called the “Regulations of Worship” (Ci Ling 祠令), in the Japanese code the “Regulations of the Heavenly and Earthly Gods” (Jingiryō 神祇令). Both regulations describe rituals performed in the presence of the emperor to sanctify his position, in which he usually played a personal role. These are therefore the key rites of sacred kingship in both traditions.
Tang Regulations
The “Regulations of Worship,” or Ci Ling section in the Tang codes clearly deals with Confucian rituals formulated in line with classic Confucian texts, most conspicuously the Rites of Zhou (Zhou Li 週禮).Footnote 42 The text states that each of these rituals is to be performed at one of two imperial state shrines: the Ancestral Shrine (miao 廟) or the Heavenly Shrine (jiao 郊).Footnote 43 There is an unequivocal separation of the places of worship based upon the different objects of worship. Celestial objects of Heaven and Earth, especially the Lord of Heaven (Tiandi 天帝), were worshipped at the Heavenly Shrine (jiao), and imperial ancestors (usually ancestors of the current lineage) at the Ancestral Shrine (miao). This is important because it separated the rites designed for imperial legitimation through associating the emperor with Heaven from those that legitimated him through association with the ancestors of the lineage or dynasty. This facilitated new lineages preserving the imperial institutions because it allowed them to utilize the Heavenly Shrine to legitimate themselves as children of Heaven regardless of the lineage worship at the Ancestral Shrine. It eased dynastic transition and mitigated any problems associated with lineage slip.
This separation between heavenly and ancestral objects of worship, and therefore between different and alternate sources of state legitimation, was already regularized during the Later Han. The separation is clearly indicated in texts marking the crystallization of imperial ritual in the Later Han, particularly those associated with the editing of Ban Gu (32–92 CE) like the Confucius House Records (Kongzi jiayu 孔子家語) and the so-called “Old text” ritual classic the Rites of Zhou.Footnote 44
The establishment of the parallel shrine system in China was also significant because it included institutionalization of these shrines and their practices in both metropolitan and provincial centers of the empire. The system of Ancestral Shrines and Heavenly Shrines was supposed to function with matching shrine institutions at the regional and central levels, with the Principal Ancestral Shrine (宗廟 zongmiao) in the imperial center, and Provincial Ancestral Shrines (guomiao 國廟) in the regions.Footnote 45 Rites carried out in the center were replicated in the provinces, thereby binding center and periphery, sovereign and vassal. This element in the Tang codes follows the basic model outlined in the Confucian ritual classics, particularly the Rites of Zhou. In this way, the Tang regulation of the ritual schemes of sacred kingship established a sharp differentiation between ancestral and heavenly objects of worship and legitimation, and a parallel relationship between rituals performed in the imperial metropole and rituals in the provincial centers.
Before moving on to look at the parallel section of the Japanese code, I want to touch briefly on how language in the Tang code reinforced the separation of heavenly and earthly objects of worship. The opening line of the “Regulations of Worship” chapter defines worship/sacrifice as follows: ”[Sacrifice/worship] for Heaven is called si 在天称祀, for the Earth we use zhai 在地為祭.”Footnote 46 The words si and zhai both mean sacrifice, but this opening passage of the Tang code indicates that they refer to different kinds of sacrifice, with different sacrificial objects represented by Heaven and Earth. This line resonates with, or can be seen as an indirect lifting from, a section in the Rites of Zhou, which makes this point more specifically, directing what kinds of songs and dances should be used for sacrifices to Heaven and to Earth, respectively.Footnote 47 In other words, the Rites of Zhou, and the Tang codes citing it, not only distinguish different objects of worship, but also differentiate ritual practices to be performed for legitimation of celestial or legendary figures of Heaven from those carried out for the more closely related imperial ancestors.
Japanese Regulations
The parallel section in the Japanese codes, the Jingiryō 神祇令 (Regulations of the Heavenly and Earthly Gods) also regulated the performance of the core state rituals, establishing a system whereby the emperor's authority was legitimated through his or her mediation between the temporal world and the world of the gods. The opening sentence of the Japanese codes resonates with that of the Tang, identifying the Jingi (Ch. Shenqi) (Heavenly and Earthly Gods) of its chapter title as referring to “Gods of Heaven and Earth” by picking out characters also employed in the opening section of the Tang code just discussed.Footnote 48 One immediate difference in the Japanese document, however, is that the distinction between rites to Heaven and Earth is not indicated with separate verbs si (for Heaven) and zhai (for Earth) as it is in the Tang codes. In the form of Japanese glossing seen in the notations on the manuscript texts of the Japanese codes held today in the Library of the Imperial Household Agency (shoryōbu) in Tokyo, both si and zhai are glossed with an identical vernacular Japanese reading—matsuri.Footnote 49 Perhaps due to this, the most common character for sacrifice employed in the Tang codes and in the Rites of Zhou, si, appears very rarely in the Japanese chapter. Instead, zhai (read in Japanese as matsuri) is used almost universally. In this way, the significance of the difference between rituals for Heaven and Earth is underplayed from the first sentence of the Japanese article. In fact, worship to a Heavenly Shrine jiao—that is, worship to celestial objects of political legitimation (like Heaven) separately from the worship of imperial ancestors—was carried out only twice in Japanese history, by Emperor Kanmu in 785 and 787.Footnote 50
This was simply because the actual objects of reverence identified in the Japanese code were not the same as those in the Chinese code. The Tang code articulates, as initial objects of reverence, a hierarchy from the Lord of Heaven shangdi down through the gods to mythic imperial ancestors—all celestial objects.Footnote 51 The Japanese code, instead, ultimately identifies Japanese proto-Shinto shrine deities as its objects of worship. Initially the Japanese code fudges this by using pseudo-Confucian nomenclature from the Rites of Zhou to identify the worship objects as “Heavenly Gods and Earthly Gods,” a term also used in the Tang code article: “The Office of Heavenly and Earthly Gods sacrifices to the Heavenly Gods and Earthly Gods in line with the standard regulations (凡天神地祇者.神祇官皆依常典祭之).”Footnote 52
But these “Heavenly and Earthly Gods” are not the gods referred to in the Chinese text. This is clear firstly from the gloss annotations on the manuscript version of the text in the Library of the Imperial Household Agency. There, the readings of “Heavenly Gods” and “Earthly Gods” are respectively given as amatsu yashiro and kunitsu yashiro.Footnote 53Amatsu yashiro kunitsu yashiro are normally written not with the Chinese characters 天神地祇 as in this text, but rather with the characters, 天社・国社. For a Japanese reader, these words in this rendering clearly refer, not to the Confucian Heaven and Earth Gods, but rather to two different kinds of Japanese kami (later labelled Shinto) shrine.Footnote 54 If the text had meant “Heavenly Gods and Earthly Gods” to refer to the generic gods of the Confucian tradition, or East Asian religion in general, then it would have glossed these characters in different Japanese, as amatsu kami kunitsu kami or tenshin chishin.
Later in the text, the rites relating to these kami shrine objects of reverence are specifically named, which provides a second layer of evidence that the Japanese codes are not referring to Confucian rites. The first rite mentioned, Toshigoi no matsuri 祈年祭, also known as Kinensai, is a specific kami shrine ritual. It is the only one of the rites contained in this section of the Japanese codes that was intended to occur at multiple shrines throughout the country at the same time, twice annually. In this sense it was the only ritual mentioned that had something like the geographic reach across the country of the Confucian rituals referred to earlier. But this rite was already defunct by the late ninth century. The other rites the Japanese regulations name are all designed for specific, individual kami shrines. Different rituals were performed at different shrines: the Hanashizume no matsuri 鎮花祭, also known as Chinkasai, occurred at the Ōkami Shrine in Yamato; the Kamumiso no matsuri 神衣祭 was a ritual of the Ise Shrine; the Ōimi no matsuri 大忌祭 is a ritual specific to the Hirose Shrine.Footnote 55 The other rituals mentioned all similarly focus on a single festival or rite restricted to one or two kami shrines. The Tang chapter mentions many rites by name, most intended to be performed across the country through duplication in the central and regional state system of Ancestral Shrines miao and Heavenly Shrines jiao. None of the rites mentioned in the Japanese regulations, however, match any of those Tang Confucian rituals. We can see that all the particular rites mentioned in this chapter of the Japanese code are from traditions now commonly labeled “Shinto.”
One of the most significant differences in the ritual systems articulated in the Tang and Japanese codes was that the Japanese code did not establish a scheme for parallel performance of imperial ritual practices at both the imperial and regional levels. The exclusive positioning of imperial ritual in individual Shinto shrines, and the restriction of Confucian practice to the central imperial academy, meant that the rituals described in the system never reached across the expansive provincial geography of the political realm. The Japanese ritual scheme lacked the tool of political integration afforded by the Chinese scheme's replication of ritual practice in imperial and regional state centers. Joan Piggot has argued that, as an administrative system, the Ritsuryō statutory state system did penetrate the countryside and was effective in linking center and periphery.Footnote 56 It is interesting to note that, by contrast, the ritual aspects of the statutory state system seem not to have performed that function.
This is crucial for considering both the motivations and consequences of the replacement of Confucian rituals with Shinto Shrine rituals in this code. A long-standing Japanese nationalist argument explaining the lack of penetration of Confucian ritual in early Japan has been that Shinto ritual already occupied the space and carried out the political functions of Confucianism in China. Confucianism was thus not only foreign, but superfluous, and also late on the scene. But we can see here that Shinto shrine practices associated with the imperial state do not seem to have been widely disseminated. In fact, their geographic reach was much more constrained than that of the administrative procedures of the early Japanese state. That is, it seems the ritual space was not already occupied. Furthermore, as Mark Teeuwen has recently argued, the imperial cult—as it is currently generally understood, with the emperor descended from Amaterasu, and the centrality of the Ise Shrine—was itself probably devised as part of the same political changes that crystallized in the statutory state system. “Shinto,” far from being the original underlying tradition, was probably constructed during the same late seventh-century period during which the imperial state institutions and Confucian influenced forms crystallized.Footnote 57 And its geographic reach never in any way mirrored that of the political realm. So the argument about “indigenous Japanese religious culture” resisting “Chinese Confucianism” stands on shaky ground.
Missing completely from the Japanese codes are the core Confucian rites, the shidian, (called sekiten or shakuten in Japanese).Footnote 58 These are mentioned in the first lines of the “Regulations of Worship” chapter of the Tang code, but are nowhere to be found in the Japanese code's parallel “Regulations of the Heavenly and Earthly Gods” article.Footnote 59 So, while in the Chinese codes the shidian ritual is part of the practices in which the emperor participates, and through which the emperor is legitimated, this is not the case in the Japanese code. In the latter, Confucian ritual in the form of the shidian was instead limited to a separate chapter on the state Confucian academy (Ch. Daxue; Jp. Daigaku). This section also exists in the Chinese codes as a separate chapter, but in the Japanese code the Confucian religious practices are only mentioned in the section on the academy: “At the imperial academy and each of the provincial state academies every year on two occasions, in Spring and Autumn, the shidian ritual will be performed for Confucius and the ancient sages.”Footnote 60
This tells us that it was usually only the academy staff who led these rituals.Footnote 61 Confucian ritual practice within the state was thus institutionally limited to spheres of government education in which lower aristocrats who served in lower to middle levels of the imperial bureaucracy were trained. Unlike in China, in Japan the shidian rite was completely disengaged from the emperor or empress and the direct legitimation of their rule.
Even in this limited setting of the academy, however, Confucianism was positioned relatively weakly in Japan. The ideal Confucian state academy system, comprising a central academy in the capital and regional academies throughout the provinces, had never been realized in early Japan. The college in Dazaifu, the imperial outpost in Kyushu in charge of continental relations, was the only regional academy established, and by 1097 it had ceased to function.Footnote 62 The Confucian academy in the capital, lacking sufficient funds to sustain itself, also sold positions. Eventually, the leadership of the Confucian academy became hereditary, thereby diminishing the real-world value of Confucian scholarship in matters of appointment even within the academy itself. The central academy became increasingly irrelevant. Its decline can be discerned in the dearth of teachers and funding. Interestingly, vegetables replaced meat as the sacrificial offering from 1163 onward, although it is unclear if this was due to a shortage of funds, Buddhist sensibilities, or both. The lack of state support for and general malaise of Confucianism were demonstrated when a fire in 1177 destroyed the Confucian academy buildings and no one bothered to reconstruct them.Footnote 63
CONCLUSION
In comparison with both imperial ritual practice in contemporaneous China and the development of other elements of the statutory state administrative system in Japan itself, the early Japanese state seems to have had a remarkably limited array of ritual schemes available to it to legitimate its sacred kingship. Confucian rites were not the prime vehicle for worship of the imperial ancestor gods in the imperial state, and they never permeated the space of the imperial palace's inner sanctum. Confucianism did not occupy a structurally preeminent position in the state rituals that legitimated the imperial line. The so-called Shinto rites inserted in their place were obviously incapable of effectively legitimating imperial rule within the kind of complex administrative imperial state envisioned in the statutory state system. These rites lacked a number of the legitimating aspects of the Confucian rites, specifically the suspension of belief in reality, the linkage of the central and regional state apparatuses, and the division between real ancestors and the legitimating but mythological realm of the legendary emperors and Heaven. The rites also did not possess the reach into provincial centers offered by the Confucian model.
Scholars of Japanese history interested in Confucianism often dismiss its weakness in the early Japanese state as simply one symptom of the overall weakness of the statutory state system.Footnote 64 I suggest that we can turn this around to argue that the removal of Confucian ritual from the imperial structures, from the very beginning of the organized Japanese imperial state, may have contributed to that state's weakness over the long term. Nationalist historiography has traditionally praised the longevity of the Japanese imperial house as a sign of the strength of Japanese “native” (“Shinto”) culture in comparison with China. The present study suggests an alternative: rather than focusing upon the alleged longevity of the imperial house, we can reflect on the imperial system's remarkably short reign on power. In comparison to other elements in the statutory state system researched by scholars like Piggott, state ritual practice, including both the remnant Confucian and dominant pre-Shinto elements, appears to have been particularly weak in terms of social and geographic penetration, even when compared with the rest of the early Japanese imperial state system. Teeuwen's argument that the imperial cult was itself probably devised as part of the same political changes crystallized in the statutory state system, and constructed during the same late seventh-century period, suggests that both the statutory state system and the Japanese imperial kami cult represent, in realpolitik terms, short-lived and ultimately unsuccessful experiments in governance, which were politically effective for a couple of centuries at most.Footnote 65
Chinese imperial rituals articulated in Confucian terms were of course also a political invention, predominantly of the Later Han. They claimed a history much deeper than the actual time of their invention. They were, however, successfully utilized in governance for a very long period after that. Chinese dynasties were governed by imperial institutions for most of the last two thousand years, but states in Japan for less than two hundred years. In China, when a lineage was surpassed, the imperial institutions, systems, and culture carried on through a new lineage. This was the cultural power of the Chinese imperial system: it could adapt to any family or lineage that occupied it, and this is one reason Chinese imperial institutions endured. The ritual system acknowledged that the emperor was not really the son of Heaven, but was a system for acting “as if” he was. Puett has explained this:
The goal [of imperial rituals examined in the Book of Rites] was not to instill a belief that Heaven had in fact given birth to the ruler. The goal was to create a ritually constructed relationship of lineage between Heaven, the ruler, and the populace—a relationship which would be seen along the same “as if” lines: the participants should act as if Heaven was the father of the ruler and the ruler the father and mother of the people, ritually developing the proper dispositions that such a constructed lineage relationship would require.Footnote 66
In the Japanese Yōrō Ritsuryō Regulations, and in particular throughout the “Regulations of the Heavenly and Earthly Gods” (jingiryō) section, Confucian ritual was replaced by the also relatively new imperial kami rituals that later came to be associated with “Shinto.” The “Shinto” manifestation of imperial rituals changed the identity and nature of the “Heaven” being worshipped in the Confucian system, and also its relationship to the imperial ancestors. In the Japanese version, Heaven was no longer an abstracted entity linked to a similarly abstract ancient world of the founding emperors, whose mythological status was confirmed in practice through ritual positioning with Heaven in the Heavenly Shrine jiao, separated from and thereby separable from the veneration of the closer imperial ancestors in the Ancestral Shrine miao. Instead, Heaven was held as the organic ancestor of the emperor, and was worshipped with no disjunction between the Heavenly and the recent imperial line. There was less discontinuity between the ancient legitimation of the celestial mythological and the modern reality of the imperial family. This was what made dynastic change relatively difficult. This, in turn, meant new rulers in Japan seldom enjoyed what many new Chinese dynasties did: continuity in political culture and governing practice, and the state's personnel, knowledge, and security apparatuses. This may be one reason why physical force was resorted to so often in premodern Japanese history, and why sociological functions of stable governance, like broad expansion of markets and popular literacy, only began to emerge in Japan in the seventeenth century, nearly half a millennium after their emergence in China.
This case is remarkable when compared with other instances of ritual and political bifurcation in sacred kingships globally, in that the Japanese emperor system was constructed in such a way as to encourage this bifurcation despite the availability of a more politically robust Chinese Confucian model. Indeed, the Confucian ritual scheme for legitimatizing the sovereign was one of the few elements of Chinese statecraft rejected in the process of Japanese cultural reproduction through which the early imperial state structures were constructed. This Japanese example of what Alan Strathern calls “the ritualization trap”—the loss (or giving-up) of real political power of a sacred sovereign—was thus born not simply of the complex local cultural, political, and religious factors at work in that space.Footnote 67 Rather, the formulation of ritualistic schemes later identified as “indigenous” or “Shinto,” such as those discussed by Teeuwen in relation to Amaterasu, and those Confucian rituals I have discussed here (later in history labeled “foreign” or “Chinese”), can both be understood through the same long durée, yet historically grounded and limited (and thereby researchable), processes of cultural reformulation and transmission. Study of these processes of cultural reformulation and transmission, rather than only the cultural history of each player in these interactions, provides a potentially powerful tool for analyzing various formulations of sacred kingship both in the particular polities of East Asia and in the comparative realm offered up by the region's shared and textually interrelated political and religious traditions.
