The Decline and Fall of Chinese Buddhist Literary Historical Consciousness: The Compilation of the Lidai sanbao ji [phrase omitted], in Light of the Dunhuang Fragments of the Zhongjing bielu [phrase omitted].

• Author: Greene, Eric M.

Pretnodern Chinese Buddhists transmitted detailed records concerning the production of their canonical scriptures--that is, information about when and by whom these texts were translated into Chinese from their Indic originals. These records allow modern scholars to date much canonical Chinese Buddhist literature with reasonable precision. This, in turn, has made the Chinese Buddhist canon a foundational source for the study of not only Chinese Buddhism, but also Indian Buddhism and Chinese Daoism, whose texts are notoriously difficult to date, not to mention Chinese historical linguistics, Chinese manuscript culture, and many other domains. For good reason, then, modem scholars have expended considerable labor on the historical bibliography of the Chinese Buddhist canon, a topic that even when not making for thrilling reading is nevertheless an indispensable prerequisite for so many other domains of inquiry.

Within the annals of Chinese Buddhist historical bibliography no source is quite so notorious as the Lidai sanbao ji [phrase omitted] (Records of the Three Jewels by dynasty), compiled by Fei Changfang (1) [phrase omitted] (fl. 562-598) in the early Sui dynasty (58I-618). (2) As is well known, relative to earlier records the Lidai sanbao Ji (LDSBJ hereafter) introduced many questionable translator attributions for the earliest Chinese Buddhist canonical scriptures. Most of these attributions were eventually accepted by later catalogers and, as a result, enshrined in the later tradition down even to modern editions of the canon. That Fei Changfang was an unreliable bibliographer is accepted by most modern scholars. But did he err because he relied on earlier, faulty sources? Because of carelessness? Or was his a deliberate deception?

In a recent paper, Michael Radich (2019) has put forth a forceful argument for the latter of these conclusions. In addition to bringing welcome attention to the general unreliability of LDSBJ, Radich proposes that we can see evidence of how Fei Changfang worked and, moreover, his bad faith--his having deliberately fabricated information--by examining patterns in the LDSBJ's treatment of those texts classified only as "translator unknown" in our earliest completely surviving catalog of Chinese Buddhist texts, Sengyou's [phrase omitted] (445-518) Chu sanzang ji ji [phrase omitted] (Collected records on the production of the Tripitaka; CSZJJ). These patterns show, Radich believes, that Fei Changfang arrived at his many new translator attributions by manipulating Sengyou's list of translator-unknown texts and assigning them to new translators in batches, the traces of which manipulation we can detect. The LDSBJ's dramatic inflation of the oeuvres of many early translators was, accordingly, not just sloppy history, let alone the accurate transmission of information contained in other, now-lost catalogs, but knowingly groundless propaganda. Fei Changfang transformed a hitherto conservative bibliographic tradition, in which the presence of many texts of unknown origins was tolerated and in which those attributions that did exist were ones that even modern scholars find plausible, into something much less accurate.

I will argue in this essay, however, that this assessment of Fei Changfang's working methods need revisiting in light of an often overlooked primary source for the study of early Chinese Buddhist historical bibliography: the Dunhuang fragments of the oldest surviving catalog of Chinese Buddhist texts, one predating even CSZJJ. These fragments have hitherto remained little known among American and European scholars. And to be clear from the start, they do not call into question the basic notion that LDSBJ is untrustworthy. But they do offer evidence that the errors of LDSBJ did not result from forgery or knowing manipulation, but from a particular reading strategy that Fei Changfang applied to his sources, of which these Dunhuang fragments are our only direct window apart from CSZJJ.

If our primary aim is to date texts, the causes of LDSBJ's errors are perhaps less important than the methodological point that we should be skeptical of its claims and, hence, of the received attributions of much Chinese Buddhist canonical literature. Yet how these errors arose is also significant because it implicates the nature and evolution of what we might call the literary historical consciousness that medieval Chinese Buddhists had vis-a-vis their canonical texts. Chinese Buddhists of the seventh century and beyond somehow felt comfortable attributing to Han-dynasty translators (for example) texts that, when judged in terms of language, vocabulary, and style, are clearly products of a much later era. There was, then, between the early 500s and the mid-600s a dramatic change in Chinese Buddhist literary historical consciousness as it shaped, and was shaped by, the hermeneutic structures that made Chinese Buddhist literature available--in this case, the canons, libraries, and catalogs that dated texts and ascribed them to differing translators or presented them only as "translator unknown." Whereas many elite Buddhist readers and catalogers prior to Fei Changfang accurately grasped, if only implicitly, the historical evolution of the idioms of canonical Chinese Buddhist literature, those subsequent to him as a rule did not. (3) For readers of Buddhist scriptural literature in China over the course of this time, it was as if a new generation were suddenly to be taught, and book publishers began to agree, that Shakespeare had written both The Tempest and Waiting for Godot and that, contrary to what had previously been assumed, English playwrights had been speaking in a relatively similar voice across the centuries.

How did this dramatic change occur, and what accounts for it? A full answer to such questions goes far beyond the small matter of Fei Changfang's good or bad faith. Yet as a preliminary step, if we can show that Fei was not simply making things up, we will have the beginnings of an argument that this change was part of a broader development within the history of Chinese Buddhism, not the side effect of the work of a single rotten historian.

SENGYOU'S LIST OF ANONYMOUS SCRIPTURES AND THE FRAGMENTS OF THE ZHONGJING BIELU

Radich is not the first to argue that Fei Changfang was a deliberate forger. (4) But his is now the most compelling presentation of this thesis and his evidence requires careful consideration. The problems of LDSBJ have traditionally been identified by comparing its main section--the "historical catalog" (dailu '[phrase omitted]) that arranges the history of Chinese Buddhist translation activity chronologically by translator and dynasty (5)--to Sengyou's CSZJJ, our oldest fully extant catalog of Chinese Buddhist texts. (6) Such a comparison quickly shows that LDSBJ dramatically expanded the oeuvres of many early translators by assigning to them hundreds of texts that CSZJJ lists only as "translator unknown" [phrase omitted] (7) Fei Changfang often supports these new attributions by citing what he claims are other (lost) catalogs, many of which purport to predate CSZJJ. That this information is in general not reliable, despite its veneer of historicity, was established convincingly by early twentieth-century scholars. (8)

Radich argues that we can see how Fei Changfang arrived at these new attributions by examining certain "troubling patterns" in his treatment of the CSZJJ list of translatorunknown texts, which, following Radich, I will call "Sengyou's list." For, as it turns out, LDSBJ assigns translators to the texts on Sengyou's list not randomly, as we might expect if Fei Changfang truly had access to new sources of information about them, but in short "batches." For example, LDSBJ assigns fifty-four texts to the translator Nie Daozhen [phrase omitted] [phrase omitted], of which forty-one can be found within in stretch of only sixty-five texts on Sengyou's list that includes 1,306 titles in total (Radich 2019: 822). It is implausible, Radich argues, that Fei Changfang would just happen to find reasons to assign to the same translator a batch of texts so closely grouped on Sengyou's long list, which was itself not ordered by translator (being, precisely, a list of translator-unknown texts). Finding many such examples, Radich concludes that Fei must have arrived at his hundreds of new attributions by gathering up batches from Sengyou's list and randomly assigning them to specific translators (ibid.: 825). Although Radich leaves the door open for the alternative possibility that Fei Changfang was merely dependent on faulty sources, he believes that "the most economical explanation... is that Fei himself is deliberately falsifying information" (ibid.: 832). (9)

Yet in drawing these conclusions Radich fails to note that we do, in fact, have direct access to part of one of the earlier catalogs that Fei Changfang claims to draw from: the Zhongjing bielu [phrase omitted] (Analytical catalog of the canon), of which there survive two Dunhuang manuscript fragments, Pelliot 3747 (Fig. 1) and Stein 2872, listing ninety texts in total. The Zhongjing bielu is also cited in CSZJJ, and on this and other grounds it is dated by scholars to no later than the early sixth century, making it the oldest Chinese Buddhist catalog to survive, even partially, as an independent work. (10) (It is moreover the second oldest independently surviving Chinese catalog of any kind, after the Han-dynasty Yiwenzhi [phrase omitted] [phrase omitted] (11) Discussed occasionally by Chinese and Japanese scholars over the years, the existence of the Zhongjing bielu fragments has been noted only in passing in Western scholarship. (12)

LDSBJ claims to draw from the Zhongjing bielu and cites it over one hundred times, including sometimes when making its novel translator attributions, such as those of Nie Daozhen mentioned above. (13) The fragments of the Zhongjing bielu therefore constitute an...