Printer Friendly

The "Religion of Images"? Buddhist Image Worship in the Early Medieval Chinese Imagination.


Religious icons and their associated beliefs and practices have been central to Chinese Buddhist life at all points in its history. Whether we turn to contemporary ethnography, ancient literary sources, or the material record itself, Buddhist sacred images are there and Chinese Buddhists are worshiping them, manufacturing them, and attributing them with power.

Images with Buddhist iconography were first produced in China no later than the second century CE, as early as, or perhaps even earlier than, the first Chinese translations of Indian Buddhist texts. Surviving examples from this period are for the most part small, Buddha-like figures incorporated into the design of Chinese funerary wares. (1) Though the actual purpose of these objects is unclear, and their immediate ritual context was not necessarily "Buddhist" in any meaningful sense (Wu 1986), textual evidence records Buddhist forms of image worship taking place among Chinese patrons of Buddhism as early as the late second century. (2) Actual examples of freestanding Buddhist icons designed for ritual worship survive, in increasing numbers, from the late third century. (3)

Chinese Buddhists interacted with their sacred images in diverse ways. They made offerings to them, composed poems in praise of them, (4) told stories about their divine power and the misfortune befalling those who despoiled them (Gjertson 1981: 297), carried out city-wide rituals of image-procession (Tsukamoto 1985: 272-77), and gained religious merit from their manufacture, (5) which by the early fifth century began to take on the massive proportions still seen today in the cave shrines of Binglingsi, Maijishan, Dunhuang, Longmen, and Yun-gang, among many other sites. Though episodes of rhetorical and actual hostility towards Buddhist images did occasionally break out among opponents of Buddhism in China and elsewhere in East Asia (Rambelli and Reinders 2012), the Chinese Buddhist tradition itself rarely if ever called the worship of images into question. (6)

Many modern scholars have also proposed that not only were sacred images crucially important to Chinese Buddhists, but also that Buddhism itself introduced to China fundamentally new ideas about such objects and their powers. As John Kieschnick (2003: 58-59) observes:
Evidence for sacred icons in China before the entrance of Buddhism is
as sketchy as it is for sacred icons in pre-Buddhist India... although
images existed in ancient China, they were rarely attributed with
divine power. And the peculiar idea that a powerful deity could be
induced to inhabit a man-made likeness was not common, if it existed
at all... just a few centuries after the arrival of Buddhism, all of
this changed dramatically, as the countryside was quickly populated
with images that not only represented deities but were also deities
themselves, capable of profoundly affecting the lives of those around

As Kieschnick goes on to note, both physical and textual evidence suggests that in pre-Buddhist China painted or sculpted images used for religious purposes were not focal points of worship, but were placed in tombs. Above-ground ancestral sacrifice, the model for many other forms of religious worship, used only a tablet bearing the written name of the deceased. (7) Though the general lack of above-ground archaeological remains from pre-Buddhist China is worth remembering, Kieschnick's point is still well taken: sacred icons appear to rise to prominence in China precisely when Buddhism was first gaining a foothold. Some scholars have seen evidence for Buddhist influence on Chinese image practices beginning from as early as the second century, when we first find Chinese gods and ancestors depicted in an "iconic" (front-facing) pose that some art historians have interpreted as presuming a context of worship. (8)

But in addition to positing Buddhism as the source of a broader cultural shift towards religious image worship in China, modern scholars have also frequently proposed the stronger thesis that this influence was apparent to the Chinese themselves. According to Erik Zurcher, Buddhist approaches to sacred images were "so characteristic and so radically different from the pre-Buddhist Chinese heritage, that they retained their specific Buddhist identity." (9) In this view, the key practices and ideologies associated with Buddhist image worship--Zurcher here mentions specifically the veneration of sacred beings in painted or sculpted form and the belief that the construction of such images would generate merit--were in China marked as distinctly Buddhist, at least for a time, just as were other initially novel Buddhist beliefs and practices such as celibate monasticism or the concept of rebirth.

Yet we should pause to consider whether Buddhist image practices actually were radically different from "the pre-Buddhist Chinese heritage" and whether, assuming, for the moment, that this was so, this would necessarily imbue them with a distinct identity. Identity, after all, is not simply given in raw difference, but must be actively constructed by those who claim it or attribute it to others, and China was not so insular, as scholars perhaps once assumed, that anything novel or foreign always and immediately stood out. (10) To assess whether particular Buddhist practices and ideologies were or were not the bearers or markers of a distinct identity therefore requires looking beyond their putative or actual novelty within the Chinese world to the particular ways that they were, or were not, actively conceived as Buddhist by the medieval Chinese. It is this question of the supposedly distinct identity of Buddhist image practices in China that I will take up below, both as a topic of intrinsic historical interest and also as a way to consider certain broader questions about how religious and other identities are constructed and the way that cultural influence, real and imagined, is experienced and represented.

My argument, which will be based primarily on a close investigation of how the worship of sacred images is discussed in literary sources, (11) can be stated easily in advance--that the worship of sacred icons was in China first singled out as a distinctly Buddhist practice only in the late fifth century. This idea thus gains currency only long after the worship of sacred icons had been fully accepted as part of many if not most forms of Chinese religion, and is moreover presented within a polemical context that should make us question just how widespread it ever truly became. The history of the Chinese conception of image worship does not, I will thus suggest, necessarily map smoothly onto the history of Chinese image worship itself.


Names, to paraphrase Confucius, are important. For good reason, then, modern scholars discussing Chinese Buddhist images have often noted that Buddhism itself was known in China as the "religion of images" (xiangjiao), (12) a designation that can only have been coined (it is suggested) because Buddhism's emphasis on such objects and their associated practices was seen as distinctive, or that Buddhism was believed to have introduced sacred icons to China, or at the very least because image worship was somehow marked as a characteristically Buddhist activity in the same way as were practices such as celibate monasticism that had no counterpart in pre-Buddhist China. (13)

Yet the history of this name xiangjiao is somewhat more complicated than has been appreciated. This history is worth examining both for its own sake and also because, as we shall see, "the religion of images" turns out not to have been simply a generic and widely used designation for Buddhism, indicative of how it was ever and always perceived in China, but a polemical term of abuse first deployed by anti-Buddhist Chinese authors in the late fifth century. Tracing this history will serve as a first step towards charting what I will suggest was indeed a broader change that took place in the late fifth century, one in which image worship was for the first time identified as not just an important Buddhist practice but as distinctively Buddhist.

Most of the major modern dictionaries of literary Chinese include an entry for xiangjiao, where it is defined as "the religion of images" and said to have been a generic name for "Buddhism." (14) Yet as Paul Pelliot (1928) pointed out many years ago (in a short article overlooked by most later scholars), the examples cited by these dictionaries do not actually support this interpretation. The earliest passage mentioned is invariably the Dhuta Temple Inscription (Touduosi bei ming [phrase omitted]) of Wang Jin [phrase omitted] (d. 505), (15) where commenting on the fate of Buddhism in India after the death of the Buddha he writes that: "the true teaching having already perished, the xiangjiao too then began to decline" [phrase omitted], [phrase omitted] (Wen xuan [phrase omitted] 59.812).

As Pelliot observed, xiangjiao here means neither "the religion of images" nor even "Buddhism" tout court. Relevant here is, rather, the technical meaning of xiangjiao as a Chinese translation of saddharma-pratirupaka ("the semblance of the true teaching"), an eschatological term from Indian Buddhism denoting a historical period beginning roughly 500 years after the Buddha's demise in which, it was argued, the "true teaching" (saddharma) of Buddhism will have disappeared with only its "semblance" (pratirupaka) remaining. (16)

The concept of the "semblance teachings" appears in Chinese Buddhist texts, beginning from the late third century, (17) in a variety of similar but slightly different expressions. (18) East Asian Buddhist dictionaries duly note these many terms and include the concept of the "semblance teachings" under their definitions of xiangjiao. Yet these Buddhist dictionaries also further give, as an additional meaning for xiangjiao in particular, "the religion of images," (19) sometimes explaining (without giving a clear source for this interpretation) that this term meant "Buddhism" by way of contrast with other Chinese traditions such as Confucianism in which images were not worshiped. (20)

A certain amount of confusion thus reigns about the meaning of the term xiangjiao and, equally importantly, how and when it may have acquired its different senses (if any there actually be). Pelliot, after noting the problems with the dictionary entries available to him, eventually suggested that interpreting xiangjiao as "the religion of images" must have first occurred in the late Tang or early Song, when secular Chinese authors had forgotten its technical meaning (Pelliot 1929). The full story, however, proves to be somewhat different--and more interesting--than Pelliot suspected.

The term xiangjiao is first attested in the early fifth century. (21) Interestingly, it initially appears not in texts translated directly from Indian languages, but in the compositions of Chinese authors. (22) In the earliest examples, the meaning is clearly no more than the age of "the semblance teachings," the later epoch of Buddhist history in which the original teachings had begun to decline. A preface by the monk Tanying [phrase omitted] (d. 405/419) may be the earliest attested example:

When the stream [of Buddhism] reached into later eras, during the time of the "semblance teachings," people's capacities became weak, and their comprehension of the path was poor. They therefore discarded the fish but kept the fish-trap, preserved the finger but forgot the moon [to which the finger points], and they considered the teachings of emptiness to mean that neither sin nor merit exists.

[phrase omitted] (23)

For Tanying, the age of the "semblance teachings" is thus one in which living beings, owing to their inferior capacities, cling to the Buddha's words at the expense of his true meaning and thereby wrongly believe that emptiness implies the negation of karma.

This notion that the passage of time had eroded human abilities was clearly attractive to medieval Chinese Buddhists, no doubt in part because it dovetailed with Chinese ideas about the decline of civilization since the golden age of the ancient sage-kings. It was also a useful way to counter criticism of Buddhism. When a certain Li Miao [phrase omitted] noted that if the Buddha really existed he would surely appear and be seen by people of the present day, the monk Daogao [phrase omitted], writing in the late fifth century, (24) could thus respond that for beings living in the age of the "semblance teachings" it would be unrealistic to expect the Buddha to manifest himself. (25)

Daogao's reply here is revealing because he links the age of the "semblance teachings" to the problem of human beings not seeing the Buddha. Had Daogao held there to be--or expected readers to hold there to be--an inherent connection between the "semblance" (xiang) age and the use of Buddhist "icons" (xiang), we might expect him to mention that the Buddha, despite his absence, is indeed still visible in this era through his images. But Daogao does not make this move, and indeed likely could not have because he takes the "true teachings" to have endured many centuries after the Buddha's death (as was the normative understanding), and yet like other learned Buddhists he would also certainly have believed that the worship of images began while the Buddha was still alive (as is told in the story of the Udayana image, which 1 will discuss below).

Chinese Buddhist texts of a less scholastic nature did sometimes conflate the "semblance" era with the mere absence of the Buddha. The opening passages of the eponymous Scripture on the Resolution of Doubts concerning the Age of the Semblance Law (Xiangfa jueyi jing [phrase omitted]), an influential late fifth- or early sixth-century apocryphal scripture, thus presents the "semblance" period as, among other things, the time when living beings can no longer behold the physical body of the Buddha. (26) Yet while image worship is indeed then mentioned at several places in this text, it is not so as to associate it with the "semblance" age, but as so to argue that it is precisely in this period that correct forms of image worship will become lost. (27) Image worship, in its ideal form, is here associated not with the "semblance" age, but with the true teachings from which the semblance age deviates.

Though initially referring to a specific period of Buddhist history, Chinese authors eventually began to deploy the trope of the "age of the semblance teachings" in certain conventionalized contexts, most particularly when discussing the transmission of Buddhism to China. Sengyou [phrase omitted] (445-518), in the preface to his famous catalogue of Buddhist texts, thus invokes the "semblance teachings" when linking the history of Buddhism to the history of Buddhism in China: "It was in the Han dynasty, during the age of the semblance teachings, that the wondrous scriptures first made their way [to us in China]" [phrase omitted] (28)

On the grounds that Buddhism arrived in China during this age, the "semblance teachings" eventually came to refer to that which was transmitted from India to China, that is to say, Buddhism itself. This usage was most common when presenting the history of Buddhism in China in sweeping terms. (29) Often, an explicit contrast is made between the disappearance of the "true teaching" in India and the spread of the "semblance teaching" to China. The Tang emperor Gaozong [phrase omitted] (r. 650-684) thus declared:
I have heard that after the true teaching perished in India, the
semblance teaching spread to China. From ancient times down to the
present, many years have passed, yet through the likes of Ananda,
Kas'yapa, As'vaghosa, and Nagarjuna, it has been like pouring water
from one vessel to another, like a flame passing [from one lamp to
another]. For this reason when we get hold of [the Buddha's] marvelous
instruction and subtle words, poring over the texts we are able to see
the meaning.

[phrase omitted]. (30)

Here, then, "the semblance teaching" does denote "Buddhism," as modern dictionaries suggest. But its connotation is more specific: not just Buddhism, but "Buddhism as we in China have come to know it." In these cases, xiangjiao means "Buddhism" from the perspective of its transmission, development, or degeneration over long periods of time, not simply Buddhist ideas, practices, or institutions writ large.

Gaozong's edict also specifically connects the spread of the "semblance teaching" to Buddhist texts and their interpretation. A similar connection is drawn in many of the other examples cited above, such as Tanying's preface, in which the age of the semblance teaching is characterized by the problematic interpretive tendencies of those living in this time, who are prone to mistake the words of the Buddha's teaching for its true meaning, to "discard the fish but keep the fish-trap."

Tanying's invocation of this famous parable of the fish-trap (originally from the Zhuangzi [phrase omitted]) is significant, for this comparison stood in many ways at the center of the distinctive hermeneutic theories of the intellectual current known to modern scholars as "dark learning" (xuan xue [phrase omitted]), which dominated early post-Han Chinese philosophy, and in which the concept of xiang [phrase omitted]., "phenomenal image," played a significant role.

These theories took as their starting point the widely held principle that the "words [of the sages] do not exhaust their meaning" (yan bujin yi [phrase omitted]), which in the Han dynasty was read as implying that the mind of the sages could only be fully plumbed by analyzing the "images" (xiang [phrase omitted].) of the Book of Changes (the hexagrams composed of broken and solid lines). In the post-Han world of "dark learning," it became common to suggest that the ultimate meaning of the sages' teachings was not to be identified with any representation whatsoever (Tang 2001a: 33). The earliest and most influential exponent of these ideas was Wang Bi [phrase omitted] (226-249). "Images," according to Wang Bi, allow one to capture the meaning, but just like the fish-trap, which must be discarded after catching the fish, "when one grasps the meaning one should forget the image" [phrase omitted]. (31)

Wang Bi here addresses the problem of how one can reliably interpret signs that are distant from that which motivated their production, either in time or in form. (It is here worth observing that, as a verb, xiang [phrase omitted]. could also denote translation.) (32) This problem of the gap between signifier and signified was in turn linked to the ontological question of the relationship between the manifest forms of the world ("being," you [phrase omitted]) and the inexpressible root from which they come ("non-being," wu [phrase omitted]). For Wang Bi, this root was "the Way" (dao [phrase omitted]), which, though the source of all things, could not be directly represented:
"Way" designates "non-being." There is nothing it does not connect,
and nothing it is not the source of. Using a metaphor for it, we call
it a "way." It is still, without substance. It cannot be given a
phenomenal image.

[phrase omitted]

However, that the Dao is not representable in "phenomenal images" (xiang) does not mean we can do without representation altogether:
Phenomenal images are that which bring out the meaning [of the sages].
Words are that which clarify the phenomenal images. For fully
elucidating the meaning [of the sages] there is nothing as good as
phenomenal images... Therefore [the sages] set up phenomenal images so
as to fully elucidate their meaning, and yet the phenomenal images
[are ultimately] to be forgotten.

[phrase omitted]

Thus for Wang Bi, that the meaning of the sages (and the Way itself) lies beyond all words and all "phenomenal images" implies not that we should dispense with such things, but that we should interpret them "gesturally," as Robert Ashmore (2004: 461) puts it, as devices for pointing to a meaning that lies beyond them. Not without reason, for medieval Chinese Buddhists these issues, first raised by Wang Bi and his epigones, were seen as identical to key Buddhist philosophical conundrums, notably how and whether emptiness (sunyata; kong [phrase omitted]) can be expressed through the words of the scriptures even though it is, supposedly, beyond all conceptual distinctions (Tang 2001a: 48-52).

We can, therefore, appreciate why xiangjiao was a productive concept for medieval Buddhist thinkers. Through it, Buddhist and Chinese hermeneutical concerns about the relationship between language (or other forms of representation) and meaning could be brought together with a Buddhist historical mythology according to which human understanding of the Buddha's teachings grows progressively more difficult.

It was in light of these connections that xiangjiao eventually became a term that could refer to all the concrete forms in which Buddhism necessarily appears while simultaneously alluding to the more expansive emptiness from which these forms had come and to which they could give access. These ideas are given eloquent form in the opening lines of a preface (to the new translation of the Huayan jing [phrase omitted]) attributed to the Empress Wu Zetian (r. 684-704) at the end of the seventh century:
True emptiness is without any phenomenal image, yet without the
phenomenal teachings there is no way to convey its truth. In ultimate
truth there are no words, yet without words and explanations there is
no way to make clear its truth... in becoming a student of the wisdom
that has no teacher, one must rely on the sutras; in learning the
principle that cannot be studied, in the end one must have recourse to
the gevas [versified Buddhist teachings].

[phrase omitted]. (35)

Here, the "phenomenal teachings" (as I would now propose we translate xiangjiao) still seem to point most directly to Buddhist texts (it is thus made parallel with a genre of Buddhist writing). But this same "dark learning" logic could easily be applied more broadly, and it is clear that some medieval Chinese authors recognized the philosophical and poetical possibilities of invoking it when discussing Buddhist icons, which were after all also called "images" (xiang), and which just as much as texts could be seen as the outward forms paradoxically necessary to convey the ultimately formless truth of Buddhism.

This move is occasionally made in dedicatory inscriptions carved on Buddhist images (or on the hybrid image-steles common in the fifth and sixth centuries), particularly within their opening passages where we often find an abstract or poetical statement concerning the ultimate ineffability of the Buddha's teaching and the Buddha himself. (36) That carved on the base of a statue of Amitabha Buddha (housed in the Museum fur Asiatische Kunst in Berlin) dated to 746 is typical of these opening formulas:
The ultimate truth is without any image (xiang), yet in the absence
of images there is no way to manifest this truth. The ultimate
principle is without words, yet without words how could one set out
this principle?

[phrase omitted]. (37)

"Image" here naturally evokes the very icon upon which the words are inscribed. So, too, the word zhen [phrase omitted], which I have translated as "truth," was also a common way of referring to sacred images or portraits--to "manifest this truth" thus also reads as "to make visible this portrait [of the Buddha]." (38) The image of the Buddha and the words carved upon it are here declared the necessary means of expressing that which ultimately lies beyond material or verbal representation.

Similar formulaic declarations, simultaneously celebrating and disclaiming the value of the object itself, are found in dedicatory Buddhist image inscriptions as far back as the late fifth century. (39) And in at least some of these examples the inexpressibility of the ultimate truth is equated with, or at least analogized to, the temporal and geographic distance of Chinese Buddhists from the original source of Buddhism in India during the time of the Buddha. An image inscription dated to the year 562 thus opens:
When the true visage [of the Buddha] hid its splendor, the true
doctrine ceased its reverberation. Vast emptiness is distant and
expansive--without "images" there is no way to manifest its "truth."

[phrase omitted]. (40)

The historical and ontological inaccessibility of the highest truths of Buddhism is here conflated. The problems occasioned by the death of the Buddha ("when the true visage hid its splendor") and the ending of the era in which the truths of Buddhism were preached or directly accessible ("the true doctrine ceased its reverberation") are joined together and likened to the ungraspability of the ultimate truth of "vast emptiness." Both problems are to be overcome through "images" (xiang), a word that here points both to the present icon of the Buddha (which allows us to see the Buddha's form) and also to all the manifest forms of Buddhism that remain in the later age of "semblance teachings" and that are simultaneously degradations of and yet also gateways to the original and ultimate.

Inscriptions such as these give us some sense that for Buddhist-inspired authors working between the late fifth and early eighth centuries, the worship of images could easily have been part of what was invoked under the label xiangjiao. But even if so--and I have yet to find the term xiangjiao itself used within these inscriptions (41)--any such connotation would have been articulated against the background of a more ancient and pervasive discourse concerning "phenomenal images" in which the issue was representation in general and that pertaining to words and written texts in particular.

Buddhism in medieval China was thus indeed the "phenomenal teachings" (xiangjiao), because as something accessible to the eyes and minds of ordinary mortals living in China through all its various forms, it was necessarily removed, in time and ontology, from its true but inaccessible origins, in the time of the Buddha and the era of the "true teachings" on the one hand, and in the ultimately un-representable emptiness of its highest meaning on the other. What it was not--at least not usually--was the "religion of images" in the sense of the particular tradition in which sacred images and their worship were emphasized.


If xiangjiao did not mean "the religion of images," why do modern dictionaries give this interpretation? It is possible to trace some of the route whereby this understanding entered modern reference works, a route that wends its way not through any particular examples of the usage of xiangjiao (which do not, as we have seen, support this reading) but through previous dictionaries and lexicons all the way back to an influential eighth-century gloss to the very line from the Dhuta Temple Inscription that modern dictionaries cite as the locus classicus for this term in this meaning. (42)

Within its context this early gloss (and the dutiful copying of it by later reference works) might well be seen as simply a mistake, as evidence that, as Pelliot suggested, later "Confucian" authors had forgotten the technical Buddhist meaning of this term. Yet as I will now suggest, an earlier, more calculated source for these ideas can be found, within a strain of anti-Buddhist polemical writing in which Buddhist image worship was singled out for censure. In medieval China Buddhism was not, as a rule, "the religion of images"; but at a certain moment in the late fifth century, someone did try to argue that it was. And it is at this moment, I will suggest, that first emerges the notion that image worship per se was distinctively Buddhist.

Chinese anti-Buddhist writings first appeared no later than the early fourth century. (43) In addition to demonstrating that Buddhism's acceptance in China met with resistance from those who viewed its key practices and doctrines as incompatible with traditionally held Chinese customs and beliefs, anti-Buddhist writings can also be used as a window onto broad features of how Buddhism was represented in China. The elements of Buddhism that this literature focuses on show us which parts of Buddhism were seen as most distinctive relative to what was, or what was imagined to be, the non-Buddhist Chinese world. (And of course, this literature itself participates in the active construction of those elements as distinctive.)

Even the very earliest anti-Buddhist polemics already contain most of the key tropes that would be repeated down through the centuries--that Buddhist celibacy contradicts filial piety, that Buddhists aim for a dangerous independence from the state, that Buddhism destabilized the economy by consuming precious resources, that Buddhism is not as ancient (and thus not as good) as the teachings of the Chinese sages. Given the endurance of these arguments, it is notable that while criticism of Buddhist image worship does eventually appear among them, it first does so only in the late fifth century, nearly two hundred years (if not more) after the first known examples of this genre of writing. It is here, moreover, that we find the first attempt to interpret the word xiangjiao as the "religion of images."

The crucial document is the Treatise on the Three Destructions [Wrought by Buddhism] (Sanpo lun [phrase omitted]), an essay of unknown authorship which survives through citations in pro-Buddhist essays written in south China during the early sixth century. (44) The Treatise on the Three Destructions appears to have been an extended presentation of the ways that Buddhism harms Chinese society. In many of its passages, the author invokes the famous "conversion of the barbarians" (huahu [phrase omitted]) (45) story according to which Buddhism was actually based on teachings that the Chinese sage Laozi delivered to the Indians. (46) Although this story may have originally been intended to promote Buddhism by linking it to a Chinese sage, it was quickly adopted for overtly polemical ends. The practices of Buddhism are, in these tellings, not so much alternative versions of Laozi's true teachings, but harsh and cruel punishments that Laozi imposed on the barbarians of India in order to break them of their evil, animalistic natures. (47)

Such stories thus argue that the distinctive practices of Buddhism, being intended to tame and subdue the barbarians, are thus inherently unsuited for the naturally superior Chinese. It is in this context that the Treatise on the Three Destructions brings up the worship of images:
[Buddhism] was not spread [by Laozi] here in China, but was,
originally, used to bring order to the Western Regions [of India], The
barbarians [of India] were inhumane. Fierce, obdurate, and lacking
ritual decorum, they were no different than birds or beasts. They had
no faith in the principle of vacuous non-being [as taught in Daoism].
For these reasons, when Laozi went through the passes [into India] he
created the "teaching of the material images" to instruct them. And it
is further said that the barbarians were coarse and wild. Desiring to
eliminate their evil seed, [Laozi] prohibited the men and women from
marrying, so that when the entire country had been converted to the
teachings, the barbarians would naturally become extinct.

[phrase omitted]. (48)

Although this passage has been discussed by previous scholars, its significance has not, I think, been fully appreciated. (49) Wu Hung, for example, has interpreted this story that Laozi created for the Indian barbarians the "teaching of the material images" as an attempt by fifth-century Daoists to justify their own recent adoption of sacred images (whose iconography was indeed patently Buddhist) by claiming that images of the Buddha are, in fact, images of Laozi. (50) Yet while this interpretation is plausible for certain later versions of the "conversion of the barbarians" story, (51) it works less well for the Treatise on the Three Destructions, whose main point is not that Buddhist images were really images of Laozi, but that Buddhist image worship, like other distinctively Buddhist practices such as celibacy, began as a punishment imposed by Laozi on the uncivilized Indians. Image worship itself is, in this account, a distinctly Buddhist activity, something suitable only for the inferior Indians, who were unable to believe in the "vacuous non-being" of Daoism.

The Treatise on the Three Destructions contrasts this "vacuous non-being" to Buddhism as the "teaching of material images." This label is surely an intentional twisting of the term xiangjiao, which as we have seen was indeed a common term for "Buddhism." Similar pseudo-etymologies, in which the true significance of a Buddhist technical term is revealed to be something pernicious, are found in many texts from the "conversion of the barbarians" genre and throughout the remainder of the Treatise on the Three Destructions. The early transliteration of "Buddha" as futu [phrase omitted] is thus explained, by taking the literal meaning of each character in Chinese, to mean "superficial executioner," on the grounds that Laozi merely shaved the Indians' heads as punishment for their crimes instead of beheading them. (52) A similar hidden meaning is found in the old transliteration sangmen [phrase omitted] (sramana, a Buddhist monk), read to mean "gate of mourning" in reference to Buddhism's being, supposedly, a religion of death. (53)

The Treatise on the Three Destructions reveals, it thus seems, the true origins of the interpretation of xiangjiao as "religion of images," an interpretation that, as discussed above, would later enter Chinese dictionaries and lexicographical commentaries but that, as we can now see, derives not from the ordinary usage of this term in Buddhist or even non-Buddhist texts, but from a blatantly polemical attempt to denigrate Buddhism as a debased teaching suitable only for non-Chinese barbarians.

The Treatise on the Three Destructions is, moreover, the earliest known example of a "conversion of the barbarians" story to discuss image worship, a point of considerable significance for our understanding the history of how such practices were thought about in China. These narratives, as we have seen, present Buddhism as originating in punishments imposed by Laozi on the Indian barbarians to curb their unruly temperaments. While room is thereby left for Buddhist philosophical insights to be valued (as originally Chinese ideas), its practices and institutions are dismissed as distinctly un-Chinese. The features of Buddhism here rejected--typically celibacy, begging for food, and abstinence from alcohol--must be precisely those the authors of such stories considered, and assumed their audience (whom they were trying to persuade) would also consider, most distinctively Buddhist.

Though the surviving versions of the "conversion of the barbarians" story, fragmentary and otherwise, are often difficult to date precisely, many examples can be reliably placed in the Liu Song dynasty (420-479) or earlier. In none of these is image worship among the practices imposed on the Indians by Laozi. (54) Within this body of literature, at least, image worship was first represented as characteristically Buddhist, and hence characteristically un-Chinese, only toward the end of the fifth century. (55)

Although the fragmentary nature of the "conversion of the barbarians" stories makes it difficult to establish any conclusions based merely on those examples that do survive, that image worship was indeed not routinely cited as a distinguishing feature of Buddhism prior to the late fifth century is confirmed through a number of other kinds of sources.

First of these is the broader body of anti- and pro-Buddhist writings authored between the third and fifth centuries. Image worship is thus not a topic of contention in any of the major anti-Buddhist (and at times, patently xenophobic) essays composed in south China during the first half of the fifth century--such as the Treatise on the Black and the White (Bai hei lun [phrase omitted]) of Huilin [phrase omitted] (composed in 436) or the Treatise on the Barbarians and the Chinese (Yi xia lun [phrase omitted]) of Gu Huan [phrase omitted] (composed in 467)--nor in the many replies to these essays by pro-Buddhist authors (and the counter-replies by further critics). (56) Perhaps the most famous and extensive early Buddhist apologia, Master Mou's Treatise on Removing Doubts (Mouzi lihuo lun [phrase omitted]), defends the compatibility of Buddhism with Chinese culture from many angles and pleads for the acceptability of doctrines such as rebirth and practices such as celibacy and head-shaving that critics, as we have seen, frequently contended were antithetical to Chinese culture. (57) Yet with the exception of passing reference to the story of the Han emperor Ming having put an image of the Buddha on his tomb (58) (I will return to this story below), image worship is neither mentioned nor explicitly defended or justified. Sun Chuo's [phrase omitted] (ca. 314-ca. 371) Essentials of the Practice of Buddhism (Fengfa yao [phrase omitted]), our most important summary of what a mid fourth-century southern Chinese layman thought were the key features of Buddhist doctrine and practice needing explanation for a Chinese audience, similarly declines to raise the question of the worship of Buddhist images as a distinct topic. (59) In the very few passages from this literature in which Buddhist images are even mentioned at all--such as in the probably early fourth-century (60) Treatise on the Rectification of Criticisms (Zhengwu lun)--the context is not the criticism of sacred icons as such, but the material ostentation of Buddhist temples in general and the economic argument that such expenditures are useless. (61) A similarly economic argument informs the earliest governmental attempts to restrict the production of Buddhism images (in 435 CE, under the Liu Song dynasty), which prohibited not Buddhist images in general, but the casting of metal icons in particular, something that, as critics frequently noted, took a valuable material resource out of circulation. (62)

Prior to the end of the fifth century image worship as such was thus not something that critics of Buddhism singled out for attack, nor that defenders of Buddhism felt the need to justify or explain. (63) Once again, then, we see that the Treatise on the Three Destructions appears to be doing something rather novel when it links the worship of images to the many other practices that in polemical literature had long been singled out as characteristic of the foreign elements of Buddhism. The Treatise on the Three Destructions, written most likely at the tail end of the fifth century, seems to mark--if it did not itself inaugurate--a new attitude towards Buddhist image worship. Critics of Buddhism could and did now begin to include this practice among the standard repertoire of blameworthy elements of Buddhism.


The many later Buddhist responses to the criticisms of image worship whose earliest clear example is the Treatise on the Three Destructions lie beyond the immediate scope of this study. What is most significant given my present purposes is simply that these criticisms first appear at the same time as do our first traces of a corresponding Buddhist anxiety about the status and origin of Chinese Buddhist traditions of image worship. (64) Such anxiety emerges, naturally enough, in the Buddhist replies to the Treatise on the Three Destructions itself. The monk Sengshun [phrase omitted] (fl. late fifth to early sixth century), replying directly to the claim that Laozi imposed image worship on the Indians, thus writes:
The first establishment of material images was not part of the
fundamental teachings [of the Buddha]. Rather [images were made]
because after [the Buddha] had died [his disciples] would otherwise
continue to long for him without cease. This is clearly explained in
the story of the sandalwood image [created by King Udayana].

Sengshun here mentions the legend of the first image of the Buddha, supposedly made while the Buddha was still alive at the request of King Udayana, a tale that was well known in China by this time. (66) Curiously, however, Sengshun does not invoke this tale merely to argue that the Treaties on the Three Destructions has gotten the history of image worship wrong, or to claim that image worship was sanctioned by the Buddha, but to insist that the making of Buddhist images, having originated with King Udayana, is not part of the Buddha's fundamental teaching.

Although Sengshun's interpretation of the Udayana story was perhaps idiosyncratic, his invocation of it points us to another set of sources in which we can also see, in a somewhat different manner, a rising concern toward the end of the fifth century for justifying Chinese traditions of Buddhist image worship. These sources are the various legends pertaining to the dream of Emperor Ming, a tale that in its developed versions describes the first transmission of Buddhism to China as including the importation of one of the Udayana images.

As the Chinese Buddhist tradition's self-understanding of its origins (or one of them), this legend has been taken by modern scholars to show that the transmission of Buddhism to China was represented as, in part, the transmission of Indian Buddhist icons and their associated ritual practices. (67) Just as does the notion that Buddhism was known in China as the "religion of the images," this story thus seems to confirm that Buddhist icons and their worship were conceptualized as among the distinguishing or distinctive features of Buddhism relative to the broader Chinese world.

However, the many versions of the legend of Emperor Ming differ in the role they assign to images. (68) What is possibly the earliest surviving exemplar--that found in the preface to the famous Scripture in Forty-Two Sections (Sishi'er zhang jing [phrase omitted])--does not mention images at all, and posits the initial transmission of Buddhism to China as entirely a matter of translating texts and establishing a Buddhist temple. (69) A similar version appears in the Daoist Scripture of the Inner Explanations of the Three Heavens (Santian neijie jing), composed during the Liu-Song dynasty. (70)

Images are, however, mentioned in versions of the story firmly dateable to the mid fourth century. Yet even here the images in question are not brought from India. According to the Record of the Later Han (Hou Han ji [phrase omitted], ca. 375 CE): "[Emperor Ming] thereupon dispatched envoys to India to inquire about their religious techniques. They then returned to China, where they made a likeness of [the Buddha's] form" [phrase omitted], [phrase omitted]. [phrase omitted], [phrase omitted]. (71) This rather cursory account was probably based on a more detailed version that is attested both in Master Mou's Treatise on Removing Doubts and in the Declarations of the Perfected (Zhen gao [phrase omitted]), a collection of Daoist texts composed between 364 and 370. In these versions, after the envoys return from India with Buddhist scriptures (only), Emperor Ming establishes a Buddhist temple and then has images of the Buddha made on the walls of his palace and atop his future tomb. (72) The eventually most recognizable version of the story--in which the envoys return from India with a/the Udayana image that serves as the model for additional images made by Emperor Ming--is first found only in a late fifth-century collection of miracle tales, the Signs from the Unseen Realm (Mingxiang ji [phrase omitted]), (73 ) also the earliest source for various other legends about famous Chinese Buddhist icons with individual histories traceable to India. (74) A vaguely similar evolution may also be evident in a different, less well-known story about the origins of Buddhism in China, the legend of Han Emperor Wu's (r. 140-87 BCE) worship of the "golden man" captured from the Xiongnu. Discussed in earlier sources simply as the worship of a nameless foreign god, in the telling of Wei Shou [phrase omitted] (506-572) in his History of the Wei (Wei shu ft [phrase omitted]) this story marks the first entry of Buddhism into China, accompanied, as in the later tales of Emperor Ming, by the importation of a new and distinctive kind of icon. (75)

The legend of Emperor Ming (or comparable tales such as the story of Emperor Wu) is, of course, not a set of historical facts, but evidence for how the arrival of Buddhism in China was narratively constructed (Campany 2012: 70-71). But these narratives themselves have a history. Though we are limited by the vagaries of the transmission and survival of our sources, the evidence we do have suggests that the earliest versions of Emperor Ming's tale either did not mention Buddhist images or else claimed only that it was the emperor who first ordered such images made in China. These stories did not claim that Buddhist icons were a new kind of object, only that the Buddha was a new deity. (76) In the late fifth century, however, an additional element was added. Buddhist icons themselves now make the journey from India to China and establish a lineage of such objects physically linked to the Buddhist homeland and parallel to the lineage of Buddhist texts. (77) Though the motivations for positing such a lineage were no doubt many, at least part of what this seems to imply is the existence of a new rhetorical climate in which Buddhist imagery and Buddhist image practices required a previously unnecessary level of justification. Whether or not this was a direct response to the kinds of criticisms seen in sources such as the Treatise on the Three Destructions is unclear. But in a similar if inverse fashion, here too image worship is newly made a central and distinctive element of Buddhism and Buddhism's transmission to China.


In the late fifth century a change thus seems to have occurred. Only from this time did the worship of images per se come to be seen as something that opponents of Buddhism might criticize or that apologists need to defend. That image worship was a distinctively Buddhist activity thus does not seem to have become a widespread assumption until this time, and seems furthermore to have been an idea that originated within the specific context of anti-Buddhist polemics.

For this conclusion I have so far presented largely negative evidence--an absence of debate or contention about Buddhist image worship prior to the late fifth century. Yet the literary record also preserves abundant positive evidence that, during roughly the first four hundred years of major Buddhist activity in China, the worship of sacred icons was seen as a generic religious practice, not something associated specifically with Buddhism.

It will be convenient to examine these references in reverse chronological order, working backwards from the late fifth century. As mentioned above, the first polemic to highlight image worship as a distinctive (and reprehensible) feature of Buddhism is the late fifth-century Treatise on the Three Destructions. We need only look roughly one generation earlier to find anti-Buddhist authors expressing a very different understanding. The Treatise on the Barbarians and the Chinese (Yi xia lun) of Gu Huan [phrase omitted] (d. 483), composed in 467, was the most important anti-Buddhist essay of its day, and inspired many replies and counter-replies among the educated elite of south China. (78) Gu Huan argued that while the fundamental truths of Buddhism and Daoism are the same, they are tailored to the specific qualities of the Indians and Chinese respectively. Despite this nominal profession of equality, Gu Huan's essay was intended, and understood, as an assertion of the cultural superiority of "Chinese" ways relative to Indian ones.

The Treatise on the Barbarians and the Chinese itself does not criticize Buddhist image worship or even mention it. Sacred images are mentioned, however, in Gu Huan's later reply to Yuan Can's [phrase omitted] (420-477) criticism of the original treatise. In his criticism, Yuan Can had disputed Gu Huan's claim that Daoism was the equal of Buddhism (Nan Qi shu 35.933). In reply, Gu Huan repeats his central thesis: that anything profound in Buddhism is also found, in a more refined form, in Daoism. It is at this point that Gu Huan mentions images:
Daoism grasps what is fundamental so as to govern what is secondary,
while Buddhism brings salvation to what is secondary so as to preserve
what is essential. So tell me, then, in what way do these two
differ?... If you were to say that they differ in that [Buddhism] sets
up images [while Daoism does not], then [I would say that] even common
spirit mediums set up images! [Images] are not the ultimate point of
Buddhism. The ultimate point is, rather, the eternally abiding, and
what difference is there between the Dao and the "eternally abiding

It would thus seem, first of all, that in the circles frequented by Gu Huan Daoist ritual practice did not involve images, and other sources confirm that this was an important difference, in theory if not in practice, between Buddhism and some (though not necessarily all) elite forms of Daoism in south China at this time. (80) But second and more importantly, Gu Huan here suggests that image worship is not fundamental to Buddhism because it is also found among the "common spirit mediums" (suwu [phrase omitted]) of Chinese popular religion.

The contrast between how Buddhist image worship is discussed by Gu Huan and Seng-shun (in his reply to the Treatise on the Three Destructions), authors separated in time by only several decades, is telling: both downplay the significance of sacred images for Buddhism, but to opposite ends, the former to denigrate Buddhism, the later to praise it. We can, once again, glimpse in this contrast the emergence, in the late fifth century, of a new polemical and rhetorical climate in which the default value of image worship had changed, from something whose absence needed to be explained (Gu Huan), to something whose presence needed to be explained (Sengshun).

Indeed, as Gu Huan himself suggests, in his day the worship of images was evidently something commonly associated not just with Buddhists, but with the ritual practices of "popular" Chinese religion. A variety of other sources confirm this impression. (81) The Daoist master Lu Xiujing [phrase omitted] (406-477), writing in the south around the same time as Gu Huan, thus instructed his followers not to install images in their home shrines, something done, he says, only by "common and vulgar families"[phrase omitted]. (82) Although the target of Lu's criticism is not clear, setting up images for worship is here seemingly criticized not for resembling Buddhist practice, but for its perceived affinity with "popular" forms of worship. (83)

An association between popular religion and sacred images of deities is also assumed in Buddhist-inspired sources from this period. The late fifth-century Signs from the Unseen Realm records the story of Chen Anju [phrase omitted], a pious Buddhist who lived with his uncle, a follower of "the customs of spirit mediums" (wusu [phrase omitted]). The central plot of the tale is a classic return-from-death narrative in which Chen is mistakenly taken to the underworld in place of his uncle. But this story is revealing for its incidental (and therefore revelatory as concerns the assumptions of author and audience) description of the religious practices of Chen's uncle's household, which was filled with "drumming, dancing, sacrifices, images of gods, (84) and shrine spaces" [phrase omitted], [phrase omitted]. (85) A similar link between popular religion and sacred images of gods is made in the story of Zhang Ying [phrase omitted], set in the early fourth century, whose conversion to Buddhism away from popular cults and their bloody sacrifices is marked when he "destroyed and cast away his images of [popular] gods and set up a grand merit-offering [in the Buddhist manner]" [phrase omitted], [phrase omitted]. (86)

By the time the Signs from the Unseen Realm was compiled in the second half of the fifth century, Buddhism had permeated all levels of Chinese society. We might therefore suppose that it was under Buddhist influence--inspired, perhaps, by the unprecedented splendor and scale of Buddhist image-making during this time--that the worship of sacred icons had by then become relatively common within popular Chinese religion.

Yet such a conclusion is frustrated by continuing references to non-Buddhist sacred icons in ever earlier textual sources. Miraculous images of popular Chinese gods--images that move on their own, shed tears, and so forth--are mentioned in several tales among the surviving fragments of Liu Yiqing's [phrase omitted] (403-444) Record of the Hidden and Visible Worlds (Youming lu [phrase omitted]). (87) A century earlier, Gan Bao's [phrase omitted] (d. 336) (88) Records of an Inquest into the Spirit-Realm (Soushen ji [phrase omitted]) also includes several tales of miraculous images of popular gods housed in local shrines (ci [phrase omitted] or miao [phrase omitted]). (89) In one such story, the daughter of a certain Zhang Pu visits a local shrine on Mt. Lu. Her servant jokingly points at one of the "statues" (xiangren [phrase omitted]) and says "let's marry you to him!" That night, Zhang Pu is visited in his dreams by the god of Mt. Lu, who thanks him for choosing his son as a match for the daughter. (90) In another tale, a sick woman out looking for medicinal herbs stumbles on a small stone shaped like a human. She makes reverence to it, and promises to serve it and build a shrine for it if it heals her illness, which, after a prophetic dream, it does. (91)

Stories of non-Buddhist icons that perform miracles can in fact be dated as far back as the Han dynasty, to the famous tale of Ding Lan [phrase omitted], the exemplar of filial piety who carved a wooden statue of his deceased father (or, in some versions, his mother) and reverenced it as if it were alive. A version of this story is found in the Records of an Inquest into the Spirit-Realm, (92) but paintings and carvings depicting the tale, along with inscriptions narrating the events, are found in tomb murals from the Eastern Han (20-220). Textual sources that record the story, some of which date to this same period of time, similarly note that the wooden statue exhibited various miracles, including showing pleasure or displeasure in its facial expressions and bleeding when cut by an angry neighbor. (93) The trope of the "animate icon," though eventually much enriched by Buddhism, clearly had its own history in China. (94) Stories of specifically Buddhist icons that perform such miracles are actually first attested in Chinese sources only centuries after the earliest non-Buddhist ones discussed above (though, of course, the surviving Buddhist tales may well have circulated long before). (95)

The tale of Ding Lan also points to the broader notion that images of the deceased could be ritually worshiped as part of ancestral sacrifice. This idea is not mentioned in the earliest Chinese ritual treatises, and scholars remain uncertain when it became common and to what extent it reflects the influence of Buddhism. But leaving aside questions of origins and influence, it is significant that this practice is alluded to, with no sense that it is non-Chinese or otherwise problematic, in relatively early sources. (96) The tale of Ding Lan is one such example, but there are others. (97)

The mid fourth-century Records of Xiangyang (Xiangyang ji [phrase omitted]), (98) cited in Pei Songzhi's commentary to the Record of the Three Kingdoms, gives what it claims is a memorial written by officials of the state of Shu after the death of the famous general Zhuge Liang [phrase omitted] [phrase omitted] (181-234) urging that a "shrine-image" (miaoxiang [phrase omitted]) be erected for him, something that, the Record of the Three Kingdoms itself records, was done in 263. Having first noted that this is what had been done in the past for many ancient worthies as far back as Zhou times, the authors continue:
[And so too] the king of Yue, holding in mind the merit of [his able
minister] Fan Li [536-448 BCE], cast metal so as to preserve his
likeness. Ever since the Han dynasty, there have been many of minor
excellence and merit for whom images have been made and temples
erected. How much more so, then, [should this be done for] [Zhuge]
Liang, whose model of virtue extends far and wide and whose influence
covers this benighted age! That the house [of you, my] king has not
perished is indeed in dependence on no other than this man. Yet the
performances of the spring and fall sacrifices [to him] do not take
place outside of his own family, and [an official] "shrine image"
[for him] has not been set up, leading the common people to make
sacrifices to him in the alleyways, and the non-Chinese tribespeople
to perform rituals to him in the wilds.

Whether or not this passage accurately records a memorial from 263, it at the least shows that by the middle of the fourth century, when the Records of Xiangyang was composed, the use of metal images within sacrificial shrines was attributed by educated Chinese authors to antiquity and was promoted in the present as a suitable way of worshiping deceased worthies such as Zhuge Liang. (100)

Other records, from the fourth century and earlier, similarly reveal that using sculpted images in the context of ancestral sacrifice, though not necessarily common, was not unknown. The Biographies of the Lang Family (Lang jia zhuan [phrase omitted]), another lost work known only through citations in the commentary to the Record of the Three Kingdoms, (101) states that when Wang Lang [phrase omitted] (d. 228) served as prefect of Kuaiji [phrase omitted] he "made sacrifices to the First Emperor of Qin. He carved a wooden statue of the emperor [and placed it] in the same temple [where were honored the ancient sage kings] Xia and Yu." [phrase omitted]

[phrase omitted]. (102) The Record of the Three Kingdoms itself, completed in the late third century, similarly mentions that when the general Cao Cao's [phrase omitted] (155-220) commander Bao Xin [phrase omitted] was killed in battle against the Yellow Turban rebels, Cao "made a wooden carving in the likeness of [Bao] Xin, and, in his grief, made sacrificial offerings to him" [phrase omitted], [phrase omitted]. (103) Another story of an ancient sage king venerated as a deity through his sacred image comes from the Treatise on Curiosities (Bowu zhi [phrase omitted]) of Zhang Hua [phrase omitted] (232-300): "When the Yellow Emperor departed as an immortal [i.e., died], his ministers pined for him without end. Some carved his image out of wood and made reverence to it." [phrase omitted]. (104)

Although it is not my intention to speculate on the ultimate point of origin of such practices, let us note, in closing, that sacrificing to images of gods or spirits is mentioned in literary sources even in the Eastern Han. The Comprehensive Meaning of Customs and Habits (Fengsu tongyi [phrase omitted]), completed in ca. 200, thus mentions the custom of sacrificing to the god known as Controller of Destiny (siming [phrase omitted]) using a small wooden statue in human form. (105) A century earlier, Wang Chong [phrase omitted] (d. ca. 100) describes another kind of popular sacrificial ritual that used images:

When constructing a building, after the work of digging the earth and excavating the soil has been completed, it is customary to appease the earth god. This is what is called "appeasing the earth." One makes an idol out of clay that reassembles the form of this god and has a shaman make prayers of invitation to it so as to appease the earth god. Having made sacrificial offerings, their minds happy and at ease, [people] say that the spirits and gods have been appeased and thanked, and that future misfortune has been eliminated.

[phrase omitted]. (106)

Wang Chong's subsequent critique makes clear that the "idols" (ouren [phrase omitted]) in question were images in the form of the earth god:

Proper ritual stipulates that upon entering an ancestral temple, having nothing upon which to focus the mind, one cuts a piece of wood to the size of ten inches [upon which is written the name of the deceased]. This is called the "ancestral tablet" (zhu [phrase omitted]), as one "focuses "(zhu [phrase omitted]) one's mind on it and makes reverence to it. One does not make a human image for this. Yet today, when performing the sacrifice for the appeasement of the earth, one makes an idol out of clay that reassembles the form of that spirit. How could this possibly allow one to appease [the earth spirit]?

[phrase omitted]. (107)

Though Wang Chong here disapproves of the practice, his comments reveal that he and his contemporaries were aware of forms of popular sacrifice and worship that involved images of the gods to whom one was sacrificing. We also see here that in Wang Chong's understanding these rituals used images as the focal point of worship in the same manner that official ancestral sacrifice used tablets bearing the name of the deceased, a ritual in which images were apparently not normally present in Wang Chong's days.

The various examples discussed above, drawn from written sources dating between the first and late fifth centuries, do not necessarily allow us to draw any conclusions about the origin, antiquity, or pervasiveness of the use of sacred icons of gods, deceased ancestors, or dead kings within Chinese ritual practice. They also do not necessarily contradict the notion that Buddhist influence should be seen as somehow responsible for the later widespread use of sacred icons in all forms of Chinese religion. But leaving aside such questions, whose resolution has in any case not been my objective, these sources show us something of equal significance--that between 100 and 500 CE using sacred icons as objects of worship, or considering them as the abode of divinities or the loci or even performers of miracles, was not something that Chinese authors assumed their readers would consider to have any necessary connection to Buddhism or to foreign religious practices more generally.


It is clear that image worship had a history in China. As I hope to have shown, it also had a historiography--there is a history to how the Chinese themselves understood the history of image worship in China. I have sketched certain details of this historiography, some explicit and some implicit, stemming from the first several hundred years of Chinese Buddhism, between roughly 100 and 500 CE. Once again, in doing this my primary aim has not been to shed light on the spread and popularization in China of Indian Buddhist images and forms of image worship (themselves neither fixed nor unitary), (108) a question for whose answer the material evidence itself would necessarily take center stage. Although the literary sources I have consulted are not irrelevant to that question, my focus has been, rather, on the differing ways that the worship of sacred icons as such was imagined and represented. While the literary sources are, like the material evidence itself, often fragmentary, they still suggest with reasonable clarity that only towards the very end of the fifth century did anyone in China begin to suggest that image worship per se was something distinctively Buddhist. It was at this time that its opponents began mocking Buddhism as "the religion of images," and that defenders of Buddhism began to justify explicitly their use of such images.

The argument that the worship of images is a distinctively Buddhist and hence distinctively foreign and un-Chinese practice was thus first made only long after sacred icons had already become common throughout other Chinese religions (if indeed they had ever not been common). This argument was, furthermore, not made in the spirit of a disinterested observation about cultural difference or the spread of new religious practices, but in a distinctly polemical context, and may therefore never have had much currency outside of a small circle of anti-Buddhist Chinese intellectuals.

Stated most broadly, the conclusion we may thus draw is that in early medieval China the status of image worship as Buddhist, foreign, or otherwise non-Chinese was not simply given: it was argued, or not, at particular moments in history, and for this reason it was subject to change and fluctuation independent of the degree to which such practices were actually distinctive within Chinese society.

And this may well have been true not only of the very earliest periods of Chinese Buddhist history, which I have surveyed above, but also in later times. The Confucian scholar Qiu Jun [phrase omitted] (1421-1495), for example, is known to have argued that the then widespread use of images when sacrificing to Confucius should be discontinued because image worship had been borrowed from Buddhism. (109) Qiu Jun thus believed, or found it convenient to argue, that Buddhism had introduced the worship of images to China, and for this reason he saw it as the bearer of a distinct Buddhist identity. (110) Yet interestingly, some earlier critics of Confucian image worship, such as Song Na [phrase omitted] (1310-1390) and Song Lian [phrase omitted] [phrase omitted] (1310-1381), had articulated their desire to return to the aniconic rituals of the Confucian classics without linking deviations from such practices to Buddhist influence (Sommer 2002: 118). This was true even for otherwise fierce critics of Buddhism such as Zhu Xi [phrase omitted] (1130-1200), who knew that sacred icons were not discussed in the Confucian ritual classics, but who attributed their widespread use to a precedent established in the second century BCE by a pious Confucian official. (111) Zhu Xi did identify what he took to be pernicious Buddhist influence on the image-based practices of his contemporaries. But he located this influence not in the mere use of sacred icons, but in details of style, such as the tendency to portray the sacred figures as seated in chairs (a way of sitting, he rightly pointed out, not used in China in ancient times). (112)

The contrasting examples of Qiu Jun and Zhu Xi show a dynamic similar to that at work in the earliest periods of Chinese Buddhism. There is here not at all a smooth evolution in which the initially foreign practices of Buddhism gradually come to be accepted as Chinese, as the story of the Buddhism's Chinese acculturation is still often told. The "foreign" or "Buddhist" nature of image worship was, rather, a site of contestation. Cultural difference--if we allow ourselves to use this potentially anachronistic label--was here not coterminous with what was actually different. What had long since ceased to be different, or even what never had been, could easily be framed as such, again or for the first time.


References to the standard Chinese histories (Hou Han shu, Sanguo zhi, Song shu, Nan Qi shu, and Nan shi) are to the 1963 Zhonghua shuju editions.

HY: The Daoist Canon, cited by page (of each text) and column (a or b) from the 1988 Wenwu chubanshe 36-vol. rpt. of the Ming canon, using the numbering system given in Weng Dujian [phrase omitted], ed., Daozang zimu yinde [phrase omitted] (Beijing: Hafo Yanjing xueshe, 1935).

T: CBETA electronic edition (version 5.2, 5/28/2014) of (with corrections) Taisho shinshu daizokyo [phrase omitted], ed. Takakusu Junjiro [phrase omitted] and Watanabe Kaigyoku [phrase omitted] [phrase omitted] (Tokyo: Taisho issaikyo kankokai, 1924-1932). Passages are cited by text number, followed by volume, page, register (a, b, or c), and line number(s).


Abe, Stanley K. 2002. Ordinary Images. Chicago: Univ. of Chicago Press.

Abramson, Marc Samuel. 2008. Ethnic Identity in Tang China. Philadelphia: Univ. of Pennsylvania Press.

Ashmore, Robert. 2004. Word and Gesture: On Xuan-School Hermeneutics of the Analects. Philosophy East and West 54: 458-88.

Beal, Samuel. 1871. A Catena of Buddhist Scriptures from the Chinese. London: Trtibner.

_________1885. Buddhist Records of the Western World. Boston: J. R. Osgood.

Behr, Wolfgang. 2004. "To Translate" Is "to Exchange" [phrase omitted]: Linguistic Diversity and the Terms for Translation in Ancient China. In Mapping Meanings: The Field of New Learning in Late Qing China, ed. M. Lackner and N. Vittinghoff. Pp. 173-209. Leiden: Brill.

Bokenkamp, Stephen R. 1996. The Yao Boduo Stele as Evidence for the "Dao-Buddhism" of the Early Lingbao Scriptures. Cahiers d'Extreme-Asie 9: 55-67.

_________1997. Early Daoist Scriptures. Berkeley and Los Angeles: Univ. of California Press.

Boucher, Daniel. 2000. On Hu and Fan Again: The Transmission of "Barbarian" Manuscripts to China. Journal of the International Association of Buddhist Studies 23: 7-28.

Bowu zhi jiaozheng [phrase omitted]. Ed. Fan Ning [phrase omitted]. Beijing: Zhonghua shuju, 1980.

Brinker, Helmut. 2002. Early Buddhist Art in China. In Return of the Buddha: The Qingzhou Discoveries, ed. Lukas Nickel. Pp. 20-33. London: Royal Academy of Arts.

Bumbacher, Stephan. 2006. A Buddhist Sutra's Transformation into a Daoist Text. Asiatische Studien 60:799-831.

Buswell, Robert E., Jr., and Donald S. Lopez, Jr. 2014. The Princeton Dictionary of Buddhism. Princeton: Princeton Univ. Press.

Campany, Robert Ford. 1996. Strange Writing: Anomaly Accounts in Early Medieval China. Albany: State Univ. of New York Press.

_________2007. Two Religious Thinkers of the Early Eastern Jin: Gan Bao and Ge Hong in Multiple Contexts. Asia Major (3rd Series) 18.1: 175-224.

________, tr. 2012. Signs from the Unseen Realm: Buddhist Miracle Tales from Early Medieval China. Honolulu: Univ. of Hawai'i Press.

Cao Zhi ji [phrase omitted]. Ed. Zhao Youwen [phrase omitted]. Beijing: Renmin wenxue chubanshe, 1984.

Carter, Martha L. 1990. The Mystery of the Udayana Buddha. Naples: Istituto Universitario Orientale.

Ch'en, Kenneth. 1952. Anti-Buddhist Propaganda during the Nan-Ch'ao. Harvard Journal of Asiatic Studies 15: 166-92.

Chappell, David. 1980. Early Forebodings of the Death of Buddhism. Numen 21 1: 122-54.

Chen Pan [phrase omitted]. 1975. Han Jin yijian shi xiao qi zhong [phrase omitted]. Taipei: Zhongyang yanjiuyuan lishi yuyan yanjiusuo.

Choi, Sun-ah. 2012. Quest for the True Visage: Sacred Images in Medieval Chinese Buddhist Art and the Concept of Zhen. PhD dissertation, Univ. of Chicago.

Ci yuan [phrase omitted]. Beijing: Shangwu yinshuguan, 1989.

Csikszentmihalyi, Mark. 2006. Readings in Han Chinese Thought. Indianapolis: Hackett.

Dai Kanwajiten [phrase omitted]. 13 vols. Ed. Morohashi Tetsuji [phrase omitted]. Tokyo: Taishukan shoten, 1955-60.

DeCaroli, Robert. 2015. Image Problems: The Origin and Development of the Buddha's Image in Early-South Asia. Seattle: Univ. of Washington Press.

Delahaye, Hubert. 1983. Les antece'dents magiques des statues chinoises. Revue d'esthetique 5: 45-54.

Demieville, Paul. 1930. Butsuzo. Hobogirin 2: 210-15.

DeWoskin, Kenneth J., and J. I. Crump. 1996. In Search of the Supernatural: The Written Record. Stanford: Stanford Univ. Press.

Digital Dictionary of Buddhism, Ed. Charles A. Muller.

Ding Fubao [phrase omitted]. 1991. Foxue da cidian [phrase omitted]. 1922. Rpt. Shanghai: Shanghai shudian chubanshe.

Dongguan Han ji jiaozhu [phrase omitted]. Ed. Wu Shuping [phrase omitted]. Zhengzhou: Zhengzhou guji chubanshe, 1987.

Dudnick, Ad. 2000. The Poem Laojun bianhua wuji jing. In Linked Faiths: Essays on Chinese Religions and Traditional Culture in Honour of Kristofer Schipper, ed. Jan A. M. De Meyer and Peter M. Engelfriet. Pp. 53-147. Leiden: Brill.

Endo Yusuke [phrase omitted]. 2014. Rikuchoki ni okeru bukkyo juyo no kenkyu [phrase omitted]. Tokyo: Hakuteisha.

Fengsu tongyi [phrase omitted]. SBCK edn.

Foguang da cidian [phrase omitted]. Gaoxiong: Foguang chubanshe, 1988.

Forke, Alfred, tr. 1911. Lun-heng. Leipzig.

Foulk, T. Griffith, and Robert H. Sharf. 1993-94. On the Ritual Use of Ch'an Portraiture in Medieval China. Cahiers d'Extreme-Asie 1: 149-219.

Fukui Kojun [phrase omitted]. 1965. Dokyo no kisoteki kenkyu [phrase omitted]. Tokyo: Shoseki bunbutsu ryutsukai.

Funayama Toru [phrase omitted] 2007. Kyoten no Gisaku to Henshu: Yuikyozanmaikyo to Sharihomonkyo [phrase omitted] . In Chugoku shukyo bunken kenkyu [phrase omitted], ed. Kyoto daigaku jinbun kagaku kenkyujo [phrase omitted]. Pp. 83-108. Kyoto: Rinsen shoten.

Giles, Herbert. 1912. A Chinese English Dictionary. Shanghai: Kelly & Walsh.

Gimello, Robert M. 2004. Icon and Incantation: The Goddess Zhunti and the Role of Images in the Occult Buddhism of China. In Images in Asian Religions: Texts and Contexts, ed. Phyllis Granoff and Koichi Shinohara. Pp. 225-56. Vancouver: Univ. of British Columbia Press.

Gjertson, Donald E. 1981. The Early Chinese Buddhist Miracle Tale: A Preliminary Survey. JAOS 101:287-301.

Granoff, Phyllis. 2001. Portraits, Likenesses and Looking Glasses: Some Literary and Philosophical Reflections on Representations and Art in Medieval India. In Representation in Religion: Studies in Honor of Moshe Barasch, ed. Jan Assmann and Albert I. Baumgarten. Pp. 63-105. Leiden: Brill.

________. 2006. Reading between the Lines: Colliding Attitudes towards Image Worship in Indian Religious Texts. In Rites Hindous, ed. Gerard Colas and Gilles Tarabout. Pp. 389-420. Paris: Editions de l'Ecoles des hautes etudes en sciences sociales.

Halperin, Mark. 2006. Out of the Cloister: Literati Perspectives on Buddhism in Sung China, 960-1279. Cambridge, Mass.: Harvard Univ. Press.

Hansen, Valerie. 2000. The Open Empire: A History of China to 1600. New York: Norton.

Hanyu da cidian [phrase omitted]. Ed. Luo Zhufeng [phrase omitted] Shanghai: Shanghai cishu chubanshe, 1986.

Hou Han ji jiaozhu [phrase omitted]. Ed. Zhou Tianyou [phrase omitted]. Tianjin: Tianjin guji chubanshe, 1987.

Hou Xudong [phrase omitted]. 1998. Wu, liu shiji beifang minzhong fojiao xinyang: Yi zaoxiang ji wei zhongxin de kaocha [phrase omitted]. Beijing: Zhongguo shehui kexue chubanshe.

Huang, Shih-shan Susan. 2012. Picturing the True Form: Daoist Visual Culture in Traditional China. Cambridge, Mass.: Harvard Univ. Press.

Hurvitz, Leon. 1956. Treatise on Buddhism and Taoism: An English Translation of the Original Chinese Text of Wei-shu CXIV and the Japanese Annotation of Tsukamoto Zenryu. Kyoto: Jimbunkagaku kenkyusho, Kyoto Univ.

_________1961. Additional Observations on the "Defense of the Faith." In Bukkyo shigaku ronshu:

Tsukamoto hakushi shoju kinen [phrase omitted]. Pp. 28-41. Kyoto: Tsukamoto hakushi shoju kinenkai.

Jiang Peijun [phrase omitted]. 2010. Dunhuang Laozi huahu jing yanjiu [phrase omitted]. Taipei: Hua Mulan wenhua chubanshe.

Julch, Thomas. 2014. Bodhisattva der Apologetik: Die Mission des buddhistischen Tang-Monchs Falin. 3 vols. Munich: Utz.

_________2016. In Defense of the Samgha: The Buddhist Apologetic Mission of the Early Tang Monk Falin. In The Middle Kingdom and the Dharma Wheel, ed. Thomas Julch. Pp. 18-93. Leiden: Brill.

Kalinowski, Marc, tr. 2011. Balance des discours: Destin, providence et divination. Paris: Les Belles Lettres.

Kamituska Yoshiko [phrase omitted]. 1999. Rikucho dokyo shiso no kenkyu [phrase omitted]. Tokyo: Sobunsha.

Karashima. Seishi. 1998. A Glossary of Dharmaraksa's Translation of the Lotus Sutra. Tokyo: The International Research Institute for Advanced Buddhology, Soka Univ.

Keenan, John P. 1994. How Master Mou Removes Our Doubts: A Reader-Response Study and Translation of the Mou-tzu Li-huo lun. Albany: State Univ. of New York Press.

Kieschnick, John. 2003. The Impact of Buddhism on Chinese Material Culture. Princeton: Princeton Univ. Press.

Kim, Minku. 2011. The Genesis of Image Worship: Epigraphic Evidence for Early Buddhist Art in China. Ph.D. dissertation, UCLA.

Kleeman, Terry F. 1994. Licentious Cults and Bloody Victuals: Sacrifice, Reciprocity and Violence in Traditional China. Asia Major (3rd Series) 7.1:185-211.

Knapp, Keith N. 2005. Selfless Offspring: Filial Children and Social Order in Medieval China. Honolulu: Univ. of Hawai'i Press.

Knechtges, David R., and Taiping Chang, eds. 2010-14. Ancient and Early Medieval Chinese Literature: A Reference Guide. 4 vols. Leiden: Brill.

Kohn, Livia. 1995. Laughing at the Tao: Debates among Buddhists and Taoists in Medieval China. Princeton: Princeton Univ. Press.

________. 1998. God of the Dao: Lord Lao in History and Myth. Ann Arbor: Center for Chinese Studies.

Kusuyama Haruki [phrase omitted]. 1979. Roshi densetsu no kenkyu [phrase omitted]. Tokyo: Sobunsha.

Lancaster, Lewis R. 1974. An Early Mahayana Sermon about the Body of the Buddha and the Making of Images. Artibus Asiae 36.4: 287-91.

Le Grand Ricci. Leiden: Brill, 2013 (online edition).

Li Jianguo [phrase omitted]. 2007. Xinji Soushen ji [phrase omitted]. Beijing: Zhonghua shuju.

Li Xiaorong [phrase omitted]. 2005. Hongming ji Guang Hongming ji shulun gao [phrase omitted]. Chengdu: Ba Shu shushe.

Liebenthal, Walter. 1955. Chinese Buddhism during the 4th and 5th Centuries. Monumenta Nipponica 11.1:44-83.

Link, Arthur E. 1961. Cheng-wu lun: The Rectification of Unjustified Criticism. Oriens Extremus 8.2: 136-65.

Liu chen zhu Wen xuan [phrase omitted]. Beijing: Zhonghua shuju, 1987.

Liu Yi [phrase omitted]. 2010. Jingdian yu lishi: Dunhuang daojing yanjiu lunji [phrase omitted]: [phrase omitted] [phrase omitted]. Beijing: Renmin chubanshe.

Lu Xun [phrase omitted]. 1973. Lu Xun quanji [phrase omitted]. 20 vols. Beijing: Renmin wenxue chubanshe.

Lunheng jiaoshi lift [phrase omitted]. Ed. Huang Hui [phrase omitted]. Beijing: Zhonghua shuju, 1990.

Maspero, Henri. 1910. Le songe de l'ambassade de l'empereur Ming: Etude critique des sources. Bulletin de l'Ecole francaise d'Extreme-Orient 10.1: 95-130.

Makita Tairyo [phrase omitted]. 1976. Gikyo kenkyu [phrase omitted]. Kyoto: Jinbun kagaku kenkyujo.

Mather, Richard B. 1963. Wang Chin's "Dhuta Temple Stele Inscription" as an Example of Buddhist Parallel Prose. JAOS 83: 338-59.

_________1992. Chinese and Indian Perceptions of Each Other between the First and Seventh Centuries. JAOS 112: 1-8.

Mathews, R. H. 1996. Mathews' Chinese-English Dictionary. Revised American Edition of 1943. Cambridge, Mass.: Harvard Univ. Press.

McNair, Amy. 2007. Donors of Longmen: Faith, Politics, and Patronage in Medieval Chinese Buddhist Sculpture. Honolulu: Univ. of Hawai'i Press.

Murray, Julia K. 2009. Idols in the Temple: Icons and the Cult of Confucius. Journal of Asian Studies 68.2:371-411.

Nakajima Ryuzo [phrase omitted]. 1985. Rikucho shiso no kenkyu: Shidaifu to bukkyo shiso [phrase omitted] %[phrase omitted]. Kyoto: Heirakuji shoten.

Nakamura Hajime [phrase omitted]. 2001. Kosetsu bukkyogo daijiten [phrase omitted]. Tokyo: Tokyo shoseki.

Nattier, Jan. 1991. Once Upon a Future Time: Studies in a Buddhist Prophecy of Decline. Berkeley: Asian Humanities Press.

Oda Tokuno [phrase omitted]. 1928. Bukkyo daijiten [phrase omitted]. Tokyo: Okura shoten.

Ofuchi Ninji [phrase omitted]. 1991. Shoki no dokyo [phrase omitted]. Tokyo: Sobunsha.

Omura Seigai [phrase omitted]. 1915. Shina bijutsushi choso hen [phrase omitted]. Tokyo: Bussho kankokai zurobu.

Orzech, Charles. 2011. On the Subject of Abhiseka. Pacific World 13: 113-28.

Palumbo, Antonello. 2001. La Scrittura di Laozi che convene i barbari: Sincretismo e conflitto ideologico in un ciclo di letteratura religiosa della Cina medievale. PhD dissertation, Istituto Universitario Orientale, Napoli.

_________2003. Dharmaraksa and Kanthaka: White Horse Monasteries in Early Medieval China. In Buddhist Asia: Papers from the First Conference of Buddhist Studies Held in Naples in May 2001, ed. Giovanni Veradi and Silvio Vita. Pp. 167-216. Kyoto: Italian School of East Asian Studies.

________. 2007. Literary Evidence on Han Buddhism: A Brief Reassessment. Paper presented at the conference Cultural Encounters between Han and Tang: New Research on Early Imperial China (SOAS, February 23, 2007).

Pelliot, Paul. 1920. Meou-tseu ou les doutes leves. T'oung Pao 19: 255-433.

_________1928. Le terme de [phrase omitted] siang-kiao comme designation du bouddhisme. T'oung Pao 25: 92-94.

________. 1929. Encore un mot a propos du "Sutra des causes et des effets" et de l'expression "siangkiao." T'oung Pao 26: 51-52.

Quan Tang wen [phrase omitted]. 11 vols. Beijing: Zhonghua shuju, 1983.

Rambelli, Fabio, and Eric Robert Reinders. 2012. Buddhism and Iconoclasm in East Asia: A History. New York: Continuum International Publishing Group.

Rawson, Jessica. 2002. The Power of Images: The Model Universe of the First Emperor and Its Legacy. Historical Research 75.188: 123-54.

Raz, Gil. 2014. Conversion of the Barbarians [huahu [phrase omitted]] Discourse as Proto Han Nationalism. The Medieval History Journal 17: 255-94.

Ren Jiyu [phrase omitted]. 1985. Zhongguo fojiao shi [phrase omitted]. 3 vols. Beijing: Zhongguo shehui kexue chubanshe.

Rhie, Marylin M. 1999. Early Buddhist Art of China and Central Asia, vol. 1. Leiden: Brill.

_________2002. Early Buddhist Art of China and Central Asia, vol. 2. Leiden: Brill.

Robson, James. 2012. Sin, Sinification, Sinology: On the Notion of Sin in Buddhism and Chinese Religions. In Sin and Sinners: Perspectives from Asian Religions, ed. Phyllis Granoff and Koichi Shinohara. Pp. 73-92. Leiden: Brill.

Sato Chisui [phrase omitted]. 1977. Hokucho zozomei ko [phrase omitted]. Shigaku zasshi [phrase omitted] 86.10: 1421-67.

Satsuho shogen koji taizen [phrase omitted]. 12 vols. Tokyo: Yoshikawa Hanshichi, 1887.

Schipper, Kristofer. 1994. Purity and Strangers: Shifting Boundaries in Medieval Taoism. T'oung Pao 80:61-81.

Seidel, Anna. 1984. Le Sutra merveilleux du Ling-pao supreme, traitant de Lao-tseu qui convertit les barbares (manuscrit S. 2081). In Contributions aux etudes de Touen-houang, vol. 3, ed. M. Soymie. Pp. 305-52. Paris: EFEO.

Shantung sikao [phrase omitted]. SKQS edn.

Sharf, Robert H. 1996. The Scripture on the Production of Buddha Images. In Religions of China in Practice, ed. Donald S. Lopez Jr. Pp. 261-67. Princeton: Princeton Univ. Press.

_________2001. Prolegomenon to the Study of Japanese Buddhist Icons. In Living Images: Japanese Buddhist Icons in Context, ed. Robert H. Sharf and Elizabeth Horton Sharf. Pp. 1-19. Stanford: Stanford Univ. Press.

Shen Jiaben [phrase omitted]. 1964. Gu shumu sanzhong [phrase omitted]. Rpt. Beijing: Zhonghua shuju.

Shishuo xinyu jiaojian [phrase omitted]. Yang Yong [phrase omitted], ed. Beijing: Zhonghua shuju, 2006.

Shinohara, Koichi. 1992. Guanding's Biography of Zhiyi. In Speaking of Monks: Religious Biography in India and China, ed. Phyllis Granoff and Koichi Shinohara. Pp. 97-232. Oakville: Mosaic Press.

_________1998. Changing Roles for Miraculous Images in Medieval Chinese Buddhism: A Study of the Miracle Image Section in Daoxuan's "Collected Records." In Images, Miracles, and Authority in Asian Religious Traditions, ed. Richard H. Davis. Pp. 141-88. Boulder: Westview Press.

Shuijing zhu jiaozheng [phrase omitted]. Ed. Chen Qiaoyi [phrase omitted]. Beijing: Zhonghua shuju, 2007.

Silk, Jonathan A., ed. 2013. Buddhism in China: Collected Papers of Erik Zurcher. Leiden: Brill.

Smith, Thomas E. 1994. Where Chinese Administrative Practices and Tales of the Strange Converge: The Meaning of gushi in the Han Wudi gushi. Early Medieval China 1: 1-33.

Sommer, Deborah. 2002. Destroying Confucius: Iconoclasm in the Confucian Temple. In On Sacred Grounds: Culture, Society, Politics, and the Formation of the Cult of Confucius, ed. Thomas A. Wilson. Pp. 95-133. Cambridge, Mass.: Harvard Univ. Asia Center.

Soothill, William Edward, and Lewis Hodous. 1937. A Dictionary of Chinese Buddhist Terms. London: Kegan Paul.

Soper, Alexander Coburn. 1958. Northern Liang and Northern Wei in Kansu. Artibus Asiae 21.2: 131-64.

_________1959. Literary Evidence for Early Buddhist Art in China. Ascona: Artibus Asiae.

Strickmann, Michel. 1979. On the Alchemy of T'ao Hung-ching. In Facets of Taoism, ed. Holmes Welch and Anna Seidel. Pp. 123-92. New Haven: Yale Univ. Press.

________. 1996. Mantras et mandarins: Le bouddhisme tantrique en Chine. Paris: Gallimard. Taiping yulan [phrase omitted]. Beijing: Zhonghua shuju, 1995.

Taishang lingbao Laozi huahu miaojing [phrase omitted], Stein no. 2081 (Images accessed from the International Dunhuang Project,

Takata Osamu [phrase omitted]. 1967. Butsuzo no kigen [phrase omitted]. Tokyo: Iwanami shoten.

Tang Yongtong [phrase omitted]. [1957] 2001a. Wei Jin xuanxue lungao [phrase omitted]. Taipei: Foguan shuju.

________. [1938] 2001b. Han Wei Liang Jin Nanbeichao fojiao shi [phrase omitted]. 2 vols. Taibei: Foguang shuju.

Teiser, Stephen F. 1994. The Scripture on the Ten Kings and the Making of Purgatory in Medieval Chinese Buddhism. Honolulu: Univ. of Hawai'i Press.

Tian, Xiaofei. 2005. Seeing with the Mind's Eye: The Eastern Jin Discourse of Visualization and Imagination. Asia Major (3rd Series) 18 (2): 67-102.

Tokuno, Kyoko. 1983. A Case Study of Chinese Buddhist Apocrypha: The Hsiang-fa chueh-i ching. M.A. Thesis, Univ. of California, Berkeley.

Tsiang, Katherine R. 2016. Visualizing the Divine in Medieval China. In A Companion to Chinese Art, ed. Martin J. Powers and Katherine R. Tsiang. Pp. 158-76. Hoboken: John Wiley and Sons.

Tsukamoto, Zenryu. 1985. A History of Early Chinese Buddhism. Tr. Leon Hurvitz. Tokyo: Kodansha International.

Tsukamoto Zenryu [phrase omitted]. [1964] 1974. Gisho Shaku Roshi no[phrase omitted] Tokyo: Daito shuppansha.

von Glahn, Richard. 2004. The Sinister Way: The Divine and the Demonic in Chinese Religious Culture. Berkeley and Los Angeles: Univ. of California Press.

Wang Bijijiaoshi [phrase omitted]. Ed. You Yulie [phrase omitted]. Beijing: Zhonghua shuju, 1980.

Wang, Eugene Y. 2014. The Shadow Image in the Cave: Discourse on Icons. In Early Medieval China: A Sourcebook, ed. Wendy Swartz et al. Pp. 405-27. New York: Columbia Univ. Press.

Wang, Michelle. 2016. Early Chinese Buddhist Sculptures as Animate Bodies and Living Presences. Ars Orientalis 46: 13-38.

Wen xuan [phrase omitted]. Beijing: Zhonghua shuju, 1977.

Wenzel, Claudia. 2011. The Image of the Buddha: Buddha Icons and Aniconic Traditions in India and China. Transcultural Studies 1: 263-305.

Whitfield, Roderick. 2005. Early Buddha Images from Hebei. Artibus Asiae 65.1: 87-98.

Wong, Dorothy C. 2004. Chinese Steles: Pre-Buddhist and Buddhist Use of a Symbolic Form. Honolulu: Univ. of Hawai'i Press.

Wu, Hung. 1986. Buddhist Elements in Early Chinese Art (2nd and 3rd Centuries A.D.). Artibus Asiae 47.3/4: 263-303.

________. 1989. The Wu Liang Shrine: The Ideology of Early Chinese Pictorial Art. Stanford: Stanford Univ. Press.

________. 2000. Mapping Early Taoist Art: The Visual Culture of Wudoumi Dao. In Taoism and the Arts of China, ed. Stephen Little and Shawn Eichman. Pp. 77-93. Chicago: Art Institute of Chicago.

_________2002. A Deity without Form: The Earliest Representation of Laozi and the Concept of Wei in Chinese Ritual Art. Orientations 33.5: 38-45.

_________2006. From the Neolithic to the Han. In Chinese Sculpture, ed. Li Song et al. Pp. 17-103. New Haven: Yale Univ. Press.

Xiong, Victor Cunrui. 2006. Emperor Yang of the Sui Dynasty. Albany: State Univ. of New York Press. Yan Juanying [phrase omitted] ed. 2008. Beichao fojiao shike tapian baipin [phrase omitted]. Taipei: Zhongyang yanjiuyuan Iishi yuyan yanjiusuo.

Yang, Liu. 1997. Manifestation of the Dao: A Study in Daoist Art from the Northern Dynasty to the Tang (5th--9th Centuries). Ph.D. dissertation, School of Oriental and African Studies, London. _________2001. Origins of Daoist Iconography. Ars Orientalis 31: 31-64.

Yang, Shao-yun. 2014. Reinventing the Barbarian: Rhetorical and Philosophical Uses of the Yi-Di in Mid-Imperial China, 600-1300. Ph.D. dissertation, Univ. of California, Berkeley.

Yoshikawa Tadao [phrase omitted]. 1998. Chugokujin no shukyo ishiki [phrase omitted]. Kyoto: Sobunsha.

_________1984. Rikucho seishinshi [phrase omitted]. Kyoto: Dohosha.

Yoshioka Yoshitoyo [phrase omitted]. 1976-80. Dokyo to bukkyo [phrase omitted] 3 vols. Tokyo: Kokusho kankokai.

Zhang Mantao [phrase omitted], ed. 1978. Sishi'er zhang jing yu Mouzi Lihuolun kaobian [phrase omitted] [phrase omitted]. Taipei: Dasheng wenhua chubanshe.

Zhang, Xunliao. 2010. Daoist Stelae of the Northern Dynasties. In Early Chinese Religion: Part Two, The Period of Division (220-589 AD), ed. John Lagerwey and Lu Pengzhi. Pp. 437-543. Leiden: Brill.

Zhang Zexun [phrase omitted]. 2009. Bei Wei Guanzhong daojiao zaoxiang ji yanjiu [phrase omitted]. Macau: Univ. of Macau.

Zhang, Zhenjun. 2014. Buddhism and Tales of the Supernatural in Early Medieval China: A Study of Liu Yiqing's (403-444) Youming lu. Leiden: Brill.

Zhongwen da cidian [phrase omitted]. 10 vols. Taipei: Zhongguo wenhua daxue chubanbu, 1985.

Ziegler, Harumi Hirano, tr. 2015. The Collection for the Propagation and Clarification of Buddhism, vol. 1. Moraga: Bukkyo dendo kyokai America.

Zurcher, Erik. [1959] 1972. The Buddhist Conquest of China: The Spread and Adaptation of Buddhism in Early Medieval China. Leiden: Brill.

_________1980. Buddhist Influence on Early Taoism: A Survey of Scriptural Evidence. T'oung Pao 66: 84-147. (Rpt. in Silk 2013: 105-64)

_________1990. Han Buddhism and the Western Region. In Thought and Law in Qin and Han China. Ed. W. L. Idema and E. Zurcher. Pp. 158-82. Leiden: Brill. (Rpt. in Silk 2013: 353-76)

_________1995. Buddhist Art in Medieval China: The Ecclesiastical View. In Function and Meaning in Buddhist Art: Proceedings of a Seminar Held at Leiden University 21-24 October 1991, ed. K. R. van Kooij and H. van der Veere. Pp. 1-20. Groningen: Egbert Forsten. (Rpt. in Silk 2013: 477-512)


Earlier versions of this paper have been much improved thanks to the comments, suggestions, criticisms, and corrections of Shao-yun Yang, Phyllis Granoff, Koichi Shinohara, Robert Sharf. and James Benn. Three anonymous reviewers for JAOS were also particularly generous with their careful attention to many remaining issues and problems, some of which may remain, but all of which have been improved thanks to their feedback.

(1.) Wu 1986: Rhie 1999: 47-66, 112-32; Abe 2002: 11-102. The exceptions to this are the cliff-carvings from Kongwangshan [phrase omitted] whose usual dating to the late Han is not undisputed (Zurcher 1990: 165-66 [Silk 2013: 359-611; Rhie 1999:27-47).

(2.) Sanguo zhi [phrase omitted] 49.1185; Zurcher 1972: 27-28. This story, of the warlord Ze Rong [phrase omitted] (d. 195) carrying out the ritual "bathing" of a metal icon of the Buddha, survives in a late third-century text, and is the earliest clear reference to Buddhist image worship in China. The two other famous early accounts of Buddhist ritual in China (by King Ying of Chu in 65 CE and by Emperor Huan in 166 CE) do not mention images (Zurcher 1972: 27, 37). Interestingly while neither the Hou Han shu [phrase omitted] (320) nor the Dongguan Han ji [phrase omitted] (Taiping yulan [phrase omitted] It 526.2387b; Dongguan Han ji jiaozhu 126) mentions an image in connection with Emperor Huan's joint sacrifice to the Buddha and Laozi (Wu 2002: 43; Kamituska 1999: 525; cf. Brinker 2002: 24), later Buddhist sources often rewrite the account to include images of both deities (Lidai sanbao ji [phrase omitted], T.2034:49.49bl7-18).

(3.) The earliest example dated by inscription (to 338 CE) is the small bronze Buddha held in the Asian Art Museum in San Francisco. Some of the many similar, undated images are believed to date from the late or even early third century (Rhie 1999: 133-53; Rhie 2002: 243-68; Whitfield 2005).

(4.) See n. 63 below.

(5.) The merit of image construction was promoted in some of the earliest Indian Buddhist scriptures translated into Chinese (Sharf 1996). A few potentially very early inscriptions (that survive in literary sources) attest to this idea operating on the ground as early as the early third century (Kim 2011: 395-403).

(6.) Demieville (1930: 210) notes the antinomian rhetoric of Chan (which had little impact on the actual veneration of icons by Chan masters or their followers) as one of the few counterexamples.

(7.) Tablets inscribed with the names of ancestors have been found in Shang dynasty tombs, and those with the names of other deities in tombs from the fourth century BCE (Lai 2016: 102-4).

(8.) Wu 1989: 130-34; Tsiang 2016: 161 (cf. Yang Liu 2001).

(9.) Zurcher 1995: 11 (Silk 2013: 504), emphasis mine.

(10.) Premodern China was considerably more open to outside influence than was once believed (see Hansen 2000). Recent research on early Chinese religions has also revealed many parallels with aspects of Buddhism once presumed to be antithetical to early Chinese culture--such as the concept of "sin" (Robson 2012) or belief in a punitive afterlife (von Glahn 2004: 45-77).

(11.) I rely on written sources (including inscriptions) both because of my own areas of expertise, but also because the question I ask concerns whether and how the ideologies and practices of image worship as such implied a distinct Buddhist identity, not whether particular images (or kinds of images) were seen or understood as "Buddhist." This later question has been skilfully and influentially addressed from the material evidence itself, in the context of the earliest surviving Chinese Buddhist images, by Wu Hung (1986).

(12.) Also written [phrase omitted]. The two graphs [phrase omitted]. and [phrase omitted], in the meaning used here, are interchangeable in medieval texts. The later form, originally meaning "elephant," is older, but the former is found already in the Chu silk manuscript (ca. 300 BCE), where it is used as a verb meaning "to depict" (Behr 2004: 217).

(13.) Beal 1871: 141; Beal 1885: 106 n. 30; Soper 1958: 141; Soper 1959: 93; Tsukamoto 1985: 483 n. 10t (this note is by Leon Hurvitz); Teiser 1994: 77; Sharf 2001: 3; Kieschnick 2003: 53; Gimello 2004: 246; Tian 2005: 90; Halperin 2006: 5; Xiong 2006: 153; Zhang Xunliao 2010: 457; Orzech 2011: 114; Huang 2012: 347 n. 10. Scholars writing in English give various slightly different but ultimately equivalent translations of xiangjiao, such as "teaching by images" (Soper), "religion of the images" (Sharf), "teaching of icons" (Kieschnick), "Doctrine of the Images" (Tian), "teaching of images" (Halperin, Orzech), or "teaching by means of icons" (Huang).

(14.) Dai Kanwa jiten 1.920, 10.659; Ci yuan 139; Hanyu da cidian 10.19a; Giles 1912: 535; Mathews 1996: 380. The Zhongwen da cidian explains the word slightly differently, as "Buddhist icons and Buddhist scriptures" (p. 1216).

(15.) On the Dhuta Temple Inscription, see Mather 1963. The Zhugong yishi [phrase omitted], the first example cited in Morohashi's Dai Kanwa jiten, is a Tang text.

(16.) On these ideas and terminology, see Chappell 1980; Nattier 1991: 66-89. There were various dating systems as to when the age of the "semblance teachings" began and how long it would last.

(17.) The earliest examples, using the term xiangfa [phrase omitted], occur in the translations of Dharmaraksa (Karashima 1998: 495; Nattier 1991: 71).

(18.) Semantically similar terms include xianghua [phrase omitted], xiangfa, xiangyun [phrase omitted], and xiangji [phrase omitted]. Of these, modern dictionaries typically connect only xiangjiao to image worship (but cf. Dai Kanwa jiten 10.659 and Foguang da cidian 5757).

(19.) Oda 1928: 675; Ding: 1991: 2224; Soothill and Hodous 1937: 420; Foguang da cidian 5756; Digital Dictionary of Buddhism (, entry for [phrase omitted], accessed January 10, 2016 ([phrase omitted] directs one to [phrase omitted], under which are given the two definitions of "religion of the image" and the "semblance period"). Nakamura's Kosetsu bukkyogo daijiten does not include xiangjiao, but does include xianghua, for which it gives both definitions (Nakamura 2001: 1062). The recent Princeton Dictionary of Buddhism closely follows the definitions given in Chinese and Japanese dictionaries (Buswell and Lopez 2014: 1008).

(20.) See in particular Oda Tokuno's early and influential Bukkyo daijiten (Oda 1928: 675), which explains that by this title Buddhism was distinguished from Confucianism, the "teaching of names" (mingjiao [phrase omitted]).

(21.) Earlier we do find the unrelated use of xiangjiao in the phrase ru shi xiang [phrase omitted], meaning "such as this" (evamrupa-; see Karashima 1998: 357-58; Nattier 1991: 73; see similarly Puyaojing [phrase omitted], T.186:3.517al).

(22.) A digital search of the Chinese Buddhist canon reveals a single early example in a text purporting to be a translation of an Indian scripture: the Shelifu wenjing [phrase omitted] (T. 1465:24.902a26). This text, however, was at least partially composed and compiled in China, no earlier than the early fifth century (Funayama 2007).

(23.) Chu sanzangjiji [phrase omitted], T.2145:55.77a25-27.

(24.) Another reply to Li Miao was authored by a certain Faming [phrase omitted], listed in the Gaoseng zhuan [phrase omitted] as the contemporary of someone who died in 492 (T.2059:50.408b22-23). We may presume that Daogao and Li Miao were active in the late fifth century.

(25.) Hongmingji [phrase omitted] T.2102:52.70b6-9.

(26.) T.2870:85.1335c27-1336a2. On this text, see Makita 1976: 304-19; Tokuno 1983.

(27.) The reader is thus told, for example, that in this later, benighted age images of the Buddha will be crafted that are incomplete (missing hands or other body parts), or that living beings will in this age fail to repair old and broken images, or even that they will set up temples without images altogether (Xiangfa jueyi jing, T.2870:85.1337c8-16).

(28.) Chu sanzang ji ji, T.2145:55.1al7-18. See also the very similarly worded comments of Shen Yue [phrase omitted] (441-513) at Guang Hongmingji [phrase omitted], T.2103:52.201a2.

(29.) See, for example, the preface by Emperor Wu (r. 502-50) of Liang (Guang Hongmingji, T.2103:52.244cl5) and the edict by the Wei emperor Xiaowen #3t (r. 471-500) (ibid., T.2103:52.272b28-29).

(30.) Xu Gaoseng zhuan [phrase omitted], T.2060:50.445a8-ll; see also Quan Tang wen 11.135. I tentatively emend "Nanda" [phrase omitted] to "Ananda" [phrase omitted].

(31.) Wang Bi ji jiaoshi 609.

(32.) Behr 2004: 215-18.

(33.) Wang Bi ji jiaoshi 624. This is Wang Bi's commentary to Lunyu [phrase omitted] 7.6. Translation based on Ashmore 2004: 460, with slight modifications.

(34.) Wang Bi ji jiaoshi 609.

(35.) Quart Tang wen 97.1001. Geya (qiye [phrase omitted]) is the second of the nine or twelve divisions of the Buddha's teaching. Always associated with verses, there are varying explanations of what texts in particular this category denotes (see, e.g., Lamotte 1944-81: 5.2284-87).

(36.) On the different formats of inscriptions found on images and steles during the fifth and sixth centuries, see Sato 1977. See also Wong 2004: 63-70.

(37.) Brinker 2002: 20; Wenzel 2011: 266.

(38.) On zhen meaning icon or portrait, see Foulk and Sharf 1993-94: 161-62, and especially Choi 2012: 29-89.

(39.) Amy McNair refers to these opening formulas as "apologies for image-making" (McNair 2007: 10-11). See also Hou 1998: 231-46; Choi 2012: 34-48.

(40.) Hou 1998: 236.

(41.) I have, to be sure, not examined all of the thousands of extant Buddhist image inscriptions dating between the fifth and seventh centuries (for a relatively comprehensive list, see Hou 1998: 312-78). Even if, however, some such examples were found, it is clear that this word xiangjiao (or its close variants) was not common in this literature, something that we would certainly expect if they had been generic terms for Buddhism meaning "the religion of images." The closest example I have found, using the related term xianghua, actually comes from an inscription (dated to 520) for an image whose iconography is Daoist rather than Buddhist: "Though the true visage has passed away, the image teaching has spread far and wide" [phrase omitted] (Yan 2008: 52; Zhang Zexun 2009: 167; on the iconography of this image, see Yang Liu 1997: 110).

(42.) The Dai Kanwa jiten, for example, draws its definition of xiangjiao from the Satsuho shogen koji taizen (4.8), a nineteenth-century Japanese version of a thirteenth-century Chinese primer. We see similar definitions in other Chinese dictionaries dating from the Song and Ming (see, e.g., the late sixteenth-century Shamang sikao [phrase omitted], 145.13). Though I have not traced every instance of it, this interpretation seems almost certain to go back to one text in particular: the Wu chen jizhu [phrase omitted] commentary to the Wen xuan (completed in 718), where according to the glosses of Li Zhouhan [phrase omitted] (n.d.) the word xiangjiao from the Dhuta Temple Inscription means "making material icons in order to teach people" [phrase omitted] (Liu chen zhu Wen xuan 59.1089).

(43.) The most useful introduction to the early anti-Buddhist polemical literature remains Zurcher 1972: 254-87. Other important studies include Ch'en 1952; Mather 1992; Hurvitz 1961; Liebenthal 1955: 78-83; Kohn 1995: 159-86; and Julch 2014 and 2016. Among the many Japanese and Chinese studies, see Tang 2001b: 2.2-8; 41-45; Ren 1985: 3.93-127; Yoshikawa 1984: 490-546; Nakajima 1985: 116-272; Endo 2014: 238-437.

(44.) Citations from the Treatise on the Three Destructions followed by refutations thereof are found in the Miehuo lun [phrase omitted] of Liu Xie [phrase omitted] (d. ca. 520) and the Xi Sanpo lun [phrase omitted] of Sengshun [phrase omitted] (n.d.), both preserved in the Hongming ji (T.2102:52.49c3-51cl0; 51cll-53cl9). Given Liu Xie's dates, it is usually inferred that the Treatise on the Three Destructions was composed in the late fifth century (Ch'en 1952: 172 n. 28).

(45.) Hu [phrase omitted] was not always a term of abuse (Boucher 2000; Yang Shao-yun 2014: xix-xxii). But within anti-Buddhist literature it did carry a clearly negative connotation (Abramson 2008: 61), making "barbarian" an appropriate translation.

(46.) On the many texts, passages, and fragments associated with the huahu genre, see Zurcher 1972: 288-320; Zurcher 1980: 93-96 (Silk 2013: 113-16); Schipper 1994; Seidel 1984; Raz 2014; Kohn 1998: 24-28, 275-90; Ofuchi 1991: 470-84; Yoshioka 1976-1980: 3.39-73; Liu 2010: 1-116; Fukui 1965: 256-320; Kusuyama 1979; Li Xiaorong 2005: 94-240; Jiang 2010. I have not had access to Palumbo 2001, which also contains a thorough study of the genre.

(47.) See, for example, Dudnick 2000: 110.

(48.) T.2102:52.50.c 19-23.

(49.) For slightly different translations of this passage, see Ch'en 1952: 172-73; Mather 1992: 4.

(50.) Wu 2000: 91

(51.) See in particular Stein no. 2081, the Marvellous Scripture of the Supreme Lingbao on the Conversion of the Barbarians by Laozi (Taishang lingbao Laozi huahu miaojing [phrase omitted]), dated by Anna Seidel to the sixth century (Seidel 1984: 330-36). In this version Laozi's instructions to make images are not connected with the punishments such as celibacy and head shaving, and Laozi himself chastises those who believe that he has no visible form. The (likely) Tang-era Dunhuang manuscripts titled Laozi huahu jing [phrase omitted]. tell a very similar version of this story (Jiang 2010).

(52.) Hongmingji, T.2102:52.50c4-6. Head-shaving was a common judicial punishment in ancient China. (There are also Indian Buddhist stories in which head-shaving is a substitute for execution; see Zhu Weimojie jing [phrase omitted] [phrase omitted], T.1775:38.399b23-25.)

(53.) T.2102:52.50c7-9. Similarly fanciful interpretations of Buddhist terminology are discussed in the Huahu jing cited in the Xiaodao lun [phrase omitted] (T.2103:52.147bl 6-23). and this rhetoric was clearly a staple of the genre.

(54.) Apart from the early brief allusions to the huahu trope in Eastern Han (25-220) and Three Kingdoms (220-280) texts (Zurcher 1972: 290-93; Ofuchi 1991: 470-84), I here include the version contained in the probably mid third-century Dadao jia lingjie [phrase omitted] (HY 788, 12a-19b; see Bokenkamp 1997: 170), the anti-Buddhist polemic cited at the beginning of the Zhengwu lun [phrase omitted] (T.2102:52.7a24-29; Link 1961: 139), the Laojun bianhua wuji jing [phrase omitted] (HY 1186; Dudnick 2000: 110), the Taishang miaoshi jing [phrase omitted] [phrase omitted] (HY 658, 6a2-6b8; Raz 2014, 272-73). the Liu-Song era Santian neijie jing [phrase omitted] (HY 1196; Bokenkamp 1997: 211-12), and the Wenshi neizhuan [phrase omitted] (cited in the Xiaodao lun, T.2103:52.145c 11-21; Kohn 1995: 70). We may also provisionally include the Mingwei huahu jing [phrase omitted], cited in the Bianzheng lun [phrase omitted] [phrase omitted] (T.2110:52.535a10-25), which some scholars think may be the original huahu scripture from the early fourth century (Zurcher 1972: 298-99). Apart from the early Han- and Three-Kingdoms-era allusions, all of these versions mention at least some concrete elements of the teaching Laozi delivered to the Indians, but none include image-making or worship among them.

(55.) Several examples dating to later than the Treatise on the Three Destructions also contain the episode of image making. These include the Taishang lingbao Laozi huahu miaojing (Seidel 1984), the ten-fascicle Dunhuang Huahu jing (Jiang 2010), as well as the Zaoli tiandi chuji [phrase omitted], cited in the Xiaodao lun (T.2103:52.144b 14-26; Kohn 1995: 52-55), and Stein no.1857. the Xisheng huahu jing [phrase omitted]. (Kusuyama 1979: 448-52; Fukui 1965: 267-84). Many of these versions date to periods when image worship was well established among Daoists, and no doubt for this reason the image-worship narrative is often contextualized differently than in the Treatise on the Three Destructions. Thus in the ten-fascicle Huahu jing and the Xisheng huahu jing, image worship is not mentioned alongside Laozi's imposing of the harsh punishments of celibacy and head shaving (as we find it in the Treatise on the Three Destructions as well as in the Zaoli tiandi chuji), but as part of a separate episode.

(56.) These and other pro- and anti-Buddhist writings of the fourth and fifth centuries are collected in fascicles 3-7 of the Hongming ji (Ziegler 2015: 91-288).

(57.) Hongming ji, T.2102:52.1bl-7a22 (Pelliot 1920; Keenan 1994; Ziegler 2015: 5-48). The preface attributes the text to a second-century author, but few modern scholars accept this. Dates ranging from between the third and fifth centuries have been proposed (Zhang Mantao 1978; Li Xiaorong 2005: 20-29).

(58.) Hongming ji, T.2102:52.5a3-8.

(59.) Hongming ji, T2102:52.86a23-89b2 (Zurcher 1972: 164-75). Sun Chuo makes one comment, when discussing the highest principles of Buddhism, to the effect that even though Buddhism does make use of material objects ([phrase omitted]) in the course of worship, its true aim lies beyond these (89a22-24). These comments seem quite general (in the vein of typical xuanxue analysis of the relationship between the formless Dao and the visible world of its "traces" discussed above) and are certainly not a defense of image worship specifically.

(60.) On internal evidence Zurcher (1972: 15) suggests an early fourth-century dating.

(61.) Hongming ji, T.2102:52.8al8-b6. The interlocutor raises the following objection: "The priests [of Buddhism] assemble the masses and construct large temples with lavish ornamentation. This is a useless waste." [phrase omitted] [phrase omitted], [phrase omitted], [phrase omitted], [phrase omitted]. One passage within the lengthy reply mentions Buddhist images: "Therefore those who serve the Buddha revere the traces he has left behind, and inwardly attempt to ponder his likeness. For this reason they inscribe and set up images and likeness so as to further their reverent devotion, and they separate themselves from their jewels and treasures [as offerings] so as to abundantly pay homage in sacred temples." [phrase omitted], [phrase omitted], [phrase omitted], [phrase omitted], [phrase omitted], [phrase omitted]. The context here is thus a broader defense of lavish displays of material devotion (a standard anti-Buddhist criticism), not an attempt to justify the representation and worship of deities in iconic form.

(62.) Song shu 97.2386; Hongming ji, T.2102:52.69al5-22.

(63.) Note should be made, however, of the hymns of praise (zan [phrase omitted]) and dedicatory "inscriptions" (ming [phrase omitted]) written for particular images or paintings, examples of which survive from the fourth century and beyond (Guang Hongming ji, T.2013:52.195.cl2-199cl2). These documents are of interest as a potentially new form of Chinese literature. But still, the early examples, such as the famous inscriptions to the Buddha's "shadow image" by Huiyuan [phrase omitted] (334-416/17) and Xie Lingyun [phrase omitted] (385-433) (Eugene Wang 2014) or Zhi Dun's [phrase omitted] (314-66) inscriptions for images of various Buddhist figures (Guang Hongming ji, T2013:52.195cl2-197c6), do not seem to make any effort to defend or justify Buddhist images or their worship, and thus do not suggest a context in which Buddhist image worship was being subjected to criticism or attack.

(64.) The later debates about image worship, between Buddhists and their critics, especially the rich records of such debates from the Tang dynasty, are the subject of the forthcoming dissertation by Kwi Jeong Lee (Princeton).

(65.) Hongming ji, T.2102:52.52b19-21. Some versions read [phrase omitted] instead of [phrase omitted].

(66.) On the legend of the Udayana Buddha, in China and elsewhere, see Takata 1967: 10-17; Carter 1990: 1-16; Sharf 1996: 262-64. Some scholars have even argued that the "conversion of the barbarians" stories in which the origins of Buddhist image worship are attributed to Laozi were at least in part drawing from versions of the Udayana story (Seidel 1984: 331). Remarkably, the Udayana story is sometimes cited in reference to Daoist images, such as in a 565 inscription for an image of "Lord Lao" (Lao jun [phrase omitted]; Kamituska 1999: 511-12). This inscription seems to show that at least some people in north China believed that the Buddha and Laozi were one and the same.

(67.) Sharf 2001: 2; Kieschnick 2003: 53; Brinker 2002: 24; Wu 2006: 102; Rambelli and Reinders 2012: 6.

(68.) For overviews of the sources for the legend, see Maspero 1910; Tang 2001b: 1.21-37; Li Xiaorong 2005: 46-93.

(69.) Chu sanzang ji ji, T.2145:55.42c 19-28. The dating of this preface is not certain. Maspero, after comparing it to other versions of the story, considered it to be the earliest and places it in the late second century (Maspero 1910: 98). This is probably far too early. Other scholars suggest the early third century (Zurcher 1990: 159 [Silk 2013: 354]), and yet others no earlier than the mid fourth century (Li Xiaorong 2005: 61-63).

(70.) ZUrcher 1980: 93-96 (Silk 2013: 113-16); Bokenkamp 1997: 214-15.

(71.) Hou Han ji 277. The editors here follow the citations of the Hou Han ji preserved in the Taiping yulan. As Antonello Palumbo has noted, however, the received version of the Hou Hanji, compiled by Huang Jishui [phrase omitted] in 1548 within the Liang Han ji [phrase omitted], does not mention envoys sent to India. Palumbo believes that this may in fact be the original form of the text, and that therefore, as he puts it, "a 4th-century historian was prepared, among other things, to describe the making of Buddha-images under the Han as an entirely Chinese business" (Palumbo 2007: 9). This conclusion, if correct, would add significant weight to my own arguments here.

(72.) Hongming ji, T.2102:52.5a5-7. For the version from the Zhen gao, which has a few small but interesting differences, see Yoshioka 1976-80: 3.6; Bumbacher 2006: 810.

(73.) Mingxiang ji [phrase omitted], cited in Fayuan zhulin [phrase omitted] (T.2122:53.383b4-13). The Mingxiang ji was compiled sometime during the Southern Qi [phrase omitted] (479-502) dynasty (Palumbo 2003: 176 n. 26). For a translation and analysis of this story, see Campany 2012: 68-71. Several other late fifth- or early sixth-century sources--including Tao Hongjing's [phrase omitted] (456-536) commentary to the Zhen gao (to be distinguished from the Zhen gao passage itself; see previous note), the Shuijing zhu [phrase omitted], and the Luoyang qielan ji [phrase omitted]--provide an intermediate version of the story, in which the envoys, while in India, make copies of scriptures and an unspecific image(s), which are then brought back to China (Maspero 1910: 1 17-18; for the Zhen gao commentary passage, see Yoshioka 1976-80: 3.7).

(74.) Notably, stories of images made and distributed by the famous Indian Buddhist ruler Asoka. Tales in which Asoka images are miraculously discovered in China appear in many later sources, but the Mingxiang ji is the earliest collection from which such a story is known (Shinohara 1992: 203-4; tr. Campany 2012: 151). Interestingly, as Shinohara notes, the story of this particular image's miraculous discovery is in fact mentioned in an earlier source, but is there not linked to Asoka or to India at all (Hongming ji, T.2102:52.71c20-24).

(75.) Wei shu 8.3025. Sorting out the different versions of this story, and what they mean for our understanding of how Buddhist image worship was conceptualized, is a somewhat thorny problem (for the relevant sources, which I cite below, see Hurvitz 1956: 27 nn. 6-8). In Wei Shou's telling. Emperor Wu acquires the image, puts it in his palace, and offers incense to it. This, Wei Shou says, was the initial germ of Buddhism in China. The story of the "golden man" is mentioned already in the Shi ji and Han shu, where it is said only that a "golden man for sacrificing to Heaven" [phrase omitted] A was acquired as booty from a conquered Xiongnu tribe, not that the Emperor then worshiped it and not that it had any connection to Buddhism. A tantalizing but brief reference to Buddhism is, however, noted already in Zhang Yan's [phrase omitted] third-century commentary to the Han shu, where he adds simply that "Buddhists sacrifice to a golden man" [phrase omitted]. Given the early references to the use of wooden icons in China, which I discuss below we might take Zhang Yan as intending to emphasize the anomalousness of metal as the material for constructing sacred images. More details appear in the Han Wu gushi [phrase omitted] a collection of tales surviving only through citations and whose original date is contested (Smith 1994 argues for a third-century date; it has also been linked to the late fifth-century author Wang lian [phrase omitted]; see Knechtges and Chang 2010: 1.349). The earliest preserved citation of the "golden man" story from this collection is given by Liu Xiaobiao [phrase omitted] (d. 521) in his commentary to the Shishuo xinyu [phrase omitted] (Shishuo xinyu jiaojian 1.192). In this version, Emperor Wu himself sacrifices (or orders sacrifices) to the image. It is also noted specifically that the emperor followed the customs of the country of the god's origin by not sacrificing animals and only offering incense and prostrations. Although the influence of Buddhism here thus seems clear. Buddhism itself is not mentioned, and this story is not framed as being about Buddhist image worship at all, merely about a strange god from foreign lands. Whatever the history of the formation of this legend, only in Wei Shou's version is it an account of the origins of Chinese Buddhism or Chinese Buddhist image worship.

(76.) That Emperor Ming places Buddhist images on (or perhaps in?) his tomb is an interesting detail that may reflect either the context, or the historical memory, of the Chinese practice of placing such images in tombs, attested beginning from the late Han in the archaeological record (see above).

(77.) We also see in these later versions a potentially new claim that the emperor did not just make such images but also inaugurated distinctive Buddhist practices relative to them: the Signs from the Unseen Realm version has the Emperor Ming not only crafting images of the Buddha and placing them in ritually important sites, but as "making [Buddhist-style] offerings to them" [phrase omitted] (Fayuan zhulin, T.2122:53.383bl2).

(78.) Nan Qi shu 35.931-32 (see also Nan shi 75.1875-80). Replies by Buddhist authors are included in the Hongming ji (T.2012:52.41b29-48a2). On the Yi xia lun, see Yoshikawa 1984: 492-99; Nakajima 1985: 244-51; Ch'en 1952: 168-72; Raz 2014: 282-86.

(79.) Nan Qi shu 35.933-34.

(80.) Kamituska 1999: 465-67. See also Lu Xiujing's comments cited below.

(81.) In what follows below, I have made much use of the material collected in Chen 1975: 56-58, but also and especially Omura 1915: 112-38.

(82.) Lu xiansheng daomen keliie [phrase omitted] (HY 1119), 4b5-8.

(83.) Bokenkamp 1996: 64. See also Yoshikawa 1998: 135-36. Daoism frequently defined itself in opposition to popular Chinese cults (Strickmann 1979; Kleeman 1994).

(84.) The term shen ying [phrase omitted], "spirit reflections," in the meaning "portrait of a deceased person," is attested from at least the Han dynasty (Xin lun [phrase omitted], cited at Taiping yulan 699.3210a). Images of Indian deities, Buddhist and otherwise, are also commonly called "reflections" (pratibimba; Granoff 2001: 95).

(85.) Fayuan zhulin, T.2122:53.756c21-22. Tr. Campany2012: 174.

(86.) Fayuan zhulin, T.2122:53.756c3-4. Tr. Campany 2012: 114-16.

(87.) Lu 1973: 8:378; 392; and 396 (see also Delahaye 1983: 50). The Youming lu contains a few stories with Buddhist characters or themes (Zhang Zhenjun 2014). Liu Yiqing also compiled a collection of explicitly Buddhist miracle tales, the Xuanyanji [phrase omitted] (Gjertson 1981).

(88.) On the date of Gan Bao's death, see Campany 2007: 175 n. 2; Li Jianguo 2007: 26.

(89.) The Soushen ji was lost during the Song. In the Ming, several reconstructions were made, the earliest and most commonly used being a twenty-fascicle version published in 1603 (Campany 1996: 55-62). The recent work of Li Jianguo [phrase omitted] (2007) has now found other preserved citations and has also excluded some stories in the twenty-fascicle version. Though it is difficult to be completely certain that any given passage attributed to the Soushen ji was truly included in Gan Bao's original text, Li's new edition allows us to be reasonably confident.

(90.) Li Jianguo 2007: 104 (no. 69); tr. DeWoskin and Crump 1996: 47-48.

(91.) Li Jianguo 2007: 109 (no. 72); tr. DeWoskin and Crump 1996: 52. See also DeWoskin and Crump 1996: 54-55, but this story has been excluded in Li Jianguo's new edition.

(92.) Li Jianguo 2007: 135 (no. 85), drawing the citation from the Taiping yulan.

(93.) On the tale of Ding Lan, see Knapp 2005: 191-94; Kim 2011: 27-75. Visual depictions of the story are found among the reliefs from the Wuliang shrine (Wu 1989: 284). and also from the Eastern Han tombs of Helinge'er and Dawenkou (Knapp 2005: 260 n. 2). Early textual references to the story include the late second-century Fengsu tongyi (3.1b) and a poem by Cao Zhi [phrase omitted] (d. 232), which mentions the miracles of the image weeping and shedding blood (Cao Zhi ji 321). A fuller version of the story is found in the Soushen ji, as well as in Sun Sheng's [phrase omitted] (302-373) Yiren zhuan [phrase omitted] (cited at Taiping yulan 414.2038).

(94.) Although both Rawson (2002) and Delahaye (1983: 51) have noted that this idea clearly predates Buddhism in China, it has remained common for ideas about animate statues to be considered Chinese solely in reference to Buddhism or Buddhist influence (see, e.g., Strickmann 1996: 165-211; Wang 2016).

(95.) As far as I am aware, the earliest Chinese text containing a story of a miracle-performing Buddhist image is Faxian's [phrase omitted] (d. 422) diary of his travels to India (Gaoseng Faxian zhuan [phrase omitted], T2085:51.858a3-4; 860bl9-29). Slightly later or contemporaneously, we find some tales of Buddhist statues with miraculous powers in Liu Yiqing's Xuanyanji (Lu 1973: 8:551; 555; 558; 559; Gjertson 1981: 297; Delahaye 1983: 48-49). These kinds of stories are very common in later collections (Shinohara 1998).

(96.) For an interpretation of the tale of Ding Lan that draws quite different conclusions about these points than my own, see Kim 2011: 27-104.

(97.) Some of these sources have been discussed by Yang Liu (2001: 41).

(98.) On the Xiangyang ji, compiled by Xi Zuochi [phrase omitted] (fl. mid fourth century), see Shen 1964: 2.9.

(99.) Sanguo zhi 35.928.

(100.) The memorial appears to aim to establish an official shrine to Zhuge Liang so as to co-opt popular sacrificial rituals to him. Zhuge Liang, it seems, was already becoming a popular god (as Fan Li may well have been, the ancient precedent the memorial cites, who eventually becomes identified as cai shen [phrase omitted], the god of wealth).

(101.) On the Lang jia zhuan, see Shen 1964: 2.12.

(102.) Sanguozhi 13.407.

(103.) Sanguo zhi 1.9.

(104.) Taiping yulan 79.370; see also 396.1832. The received version of the Bowu zhi is worded slightly differently (Bowu zhi jiaozheng 93 no. 266).

(105.) Fengsu tongyi 8.9b-10a; tr. Csikszentmihalyi 2006: 180.

(106.) Lunheng jiaoshi, 25.1044; tr. Forke 1911: 1535, Kalinowski 2011: 268-69.

(107.) Lunheng jiaoshi, 25.1045; tr. Forke 1911: 1536, Kalinowski 2011: 270.

(108.) The adoption of image worship within Buddhism and other Indian religions was also a gradual process, accompanied by much debate (Lancaster 1974; Granoff 2006; DeCaroli 2015).

(109.) Sommer 2002: 118-19. The government eventually followed Qiu Jun's proposals and tried to eliminate images of Confucius, except at the temple in Confucius' home town of Qufu.

(110.) Qiu Jun was not the only Ming author to argue that image worship came to China via Buddhism (Murray 2009: 384-385), and it is interesting to consider this in light of the definitions of xiangjiao as "the religion of images" that, as discussed above (see n. 42), were found in Ming-dynasty dictionaries.

(111.) Sommer 2002: 103. The earliest clear reference to images of Confucius as objects of worship occurs in a sixth-century text (Shuijing zhu jiaozheng 25.594).

(112.) The chair was indeed introduced to East Asia via Buddhism (Kieschnick 2003: 222-48).

[Please note: Some non-Latin characters were omitted from this article]
COPYRIGHT 2018 American Oriental Society
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2018 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Greene, Eric M.
Publication:The Journal of the American Oriental Society
Article Type:Essay
Geographic Code:9CHIN
Date:Jul 1, 2018
Previous Article:Sudarabien im Altertum: Kommentierte Bibliographie der Jahre 1997 bis 2011.
Next Article:The Assyrian Sculptures from the Nergal Gate Museum at Nineveh before the Islamic State's Attack.

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |