28 July 2023

The Lost Translations of the Heart Sutra

If there is anything eternal, it may well be Buddhist anxieties about the authenticity, legitimacy, and authority. These anxieties seem to be present in the earliest strata of Buddhist writing and continue down to the present. One of the principle methods of making a text seem more authentic (etc) is to claim that it is old. There is a Buddhist heuristic that the older a text is, the more authentic it is. This is one reason that, for some people, the Pāli texts are seen as more authentic and thus more legitimate and more authoritative than other texts.

In the arena of Heart Sutra studies there is an old argument for the antiquity of the text, which is to cite the so-called "lost translations", and one in particular. This essay draws heavily on Watanabe (1990) an article, published in Japanese, but of which I have recently made an English translation, using ChatGPT and some other online translation apps. Watanabe was the first to make this argument and it was made in 1990, two years before Nattier stumbled on the fact that the Sanskrit text is a backtranslation. 

We can see this trope of lost translations invoked, for example, in recent Zen Buddhist commentaries on the Heart Sutra by Red Pine (2004) and Tanahashi Kazuaki (2014). Both men cite a lost translation attributed to Zhi Qian 支謙 (fl. 222–254 AD) that enables them to date the Heart Sutra very early (first or second century CE). Tanahashi Kazuaki (2014: 62) says:

Among the vanished texts, the most noteworthy is the rendition by Zhiqian [sic] of the third century. Traditionally regarded as the oldest Chinese translation of the Heart Sutra, this text was reportedly included in [Sengyou’s Catalogue].

"Sengyou's Catalogue" refers to the Chūsānzàng jìjí «出三蔵記集» (T 2145), compiled by Sēngyòu 僧祐 (445-518 CE). Amongst the resources employed by Sengyou were older catalogues, notably one by Dao-an 道安 (312–385) compiled in 374 CE (itself now lost). In the Dao-an section of Sengyou’s catalogue we find two texts listed:

A. Móhē bānrě bōluómì shénzhòu yī juàn 摩訶般若波羅蜜神呪一卷 “Dhāraṇī of the Great Prajñāpāramitā”; one scroll. (T 2145; 55.31b9)

B. Bānrě bōluómì shénzhòu yī juàn (yìběn) 般若波羅蜜神呪一卷(異本) “Dhāraṇī of the Prajñāpāramitā”; one scroll (different version). (T 2145; 55.31b10).

The astute reader will note that neither text is called a Heart Sutra; or a sutra, for that matter. It is less obvious, perhaps, that neither text is attributed to Zhi Qian. The term shénzhòu 神呪 probably translates dhāraṇī or vidyā, but we don't know. Not only are there no Indic sources for these titles, the texts themselves were lost by the Tang dynasty. So these catalogue entries are almost everything we know about these two shénzhòu texts.

One may compare these with the two entries in the Tuóluóní jí jīng «陀羅尼集經» (T 901), translated by Atikūṭa 654 CE. I translate and comment on these entries in a blog post: Svāhā in The Heart Sutra Dhāraṇī (5 July 2019)

I say almost everything we know, but there is a little more because the texts crop up in some later catalogues with the notation: "produced from the Large Sutra" (Chū dà pǐn jīng 出大品經), which is used to indicate the text is an extract from the larger work. Furthermore, in the Zhòngjīng mùlù (衆經目録) (also known as Yàncóng Lù 録), compiled and written under the guidance of Yàncóng 彦琮 (602 CE), both the A and B shénzhòu texts are classified as “separately produced” (biéshēng 別生). This is a term used for locally produced Chinese Buddhist texts, and has also been applied to chāo jīng 抄經 or digest texts. 

And all this evidence from the catalogues is consistent with the comments of Kuījī (T 1710) and Woncheuk (T 1711) who both composed commentaries on the Heart Sutra in the late-seventh century. They clearly state that they don’t believe the Heart Sutra to be an authentic Buddhist sutra; rather, they both see it as a compilation of passages from other Prajñāpāramitā texts. Moreover, physical and literary evidence stops entirely in the mid-seventh century: earliest artefact is from 661CE, earliest literary mention is from 656 CE.

So there are two processes to try to understand. How did the two shénzhòu come to be associated with Zhi Qian? And how did the shénzhòu texts come to be considered versions of the Heart Sutra?

Zhi Qian and Fèi Chángfáng

Following Sengyou, a series of three catalogues named Zhòngjīng mùlù «衆經目錄», by Fǎjīng 法經 (594 CE), Yancong 彥悰 (602 CE), and Jìngtài 靜泰 (663-665 CE), all list the two shénzhòu texts as "translator lost" (shī yì 失譯). However, in the midst of these we also have the Lìdài sānbǎo jì «歷代三寶紀» (T 2034) compiled by Fèi Chángfáng 費長房 (597 CE). The Lìdài sānbǎo jì is infamous amongst scholars for adding attributions to texts that were previously listed as "translator lost". Many of these attributions are false and the text is widely considered unreliable in matters of attribution. Fei's entry for the A text reads:

摩詞般若波羅蜜呪經 見宝唱録或直云般若波羅蜜呪經〔支謙訳〕
Móhē bānrě bōluómì zhòu jīng. See the Bǎochànglù; in some cases it is just called Bānrě bōluómì zhòu jīng. Translated by Zhi Qian.

Note the subtle change in the title. The character shén 神 "divine" has been dropped and the character jīng 經 "text, sutra" has been added. Still, everyone involved thinks this is the same text as found in Sengyou's Catalogue. Note that the Bǎochànglù is a reference to another catalogue that no longer exists: the Liángshì zhòng jīng mùlù «梁世衆經目錄» compiled by Bǎochàng 寶唱 ca. 520-521. It's possible that Bǎochàng was responsible for this attribution, but Fèi Chángfáng made up so many attributions that the finger points squarely at him. Also note that, contra the Zhòngjīng mùlù catalogues, Fèi Chángfáng considers the version without móhē 摩詞 in the title to be a variant of the A text rather than a distinct B text.

As far as we can tell, then, Chángfáng simply made up this attribution. And there is no reason to suppose that Zhi Qian translated the Móhē bānrě bōluómì shénzhòu or the Móhē bānrě bōluómì zhòu jīng. Rather, such texts were likely just extracts from the Large Prajñāpāramitā text that circulated independently. Note that it is quite definite that the Xīn jīng (T 251) copied multiple passages from the Móhē bānrě bōluómì jīng «摩訶般若波羅蜜經» (T 223), translated by Kumārajīva in 404 CE, as does the Dàmíngzhòu jīng (see below). Assuming that all the catalogue entries relating to the shénzhòu texts are references to the same text, the appearance in Dao-an's catalogue dated 374 definitely rules out it being a Heart Sutra. The passages copied did not even exist until thirty years after this date.

That said, the attribution to Zhi Qian is cited in influential catalogues such as the Neidian Catalogue (Dà Táng nèidiǎn lù «大唐内典録» T 2149), compiled by Dàoxuān 道宣 (664) and the Kaiyuan Catalogue (Dà Táng kāiyuán shìjiào lù «大唐開元釋教錄» T 2154), compiled by Zhìshēng 智昇 (730). The latter was especially influential as it was used to reconstruct the Buddhist canon after the purges of 849 and eventually provided the organisational scheme followed by the Tasihō canon.

At this point, then, the móhē shénzhòu text has been identified as a translation by Zhi Qian, while the B shénzhòu (sans móhē) is either noted as "translator lost" or is said to be the same text with a different title, despite Sengyou's clear note that they were different. What we do not have anywhere in the picture is a Heart Sutra text. We turn to this mystery next.

Zhi Qian and the Heart Sutra

The key moment here is the appearance, already mentioned above, of the Kaiyuan Catalogue by Zhìshēng, in 730 CE. Something new happens in this catalogue, which is the first mention of a text that we know to be a Heart Sutra:

A 摩詞般若波羅蜜呪經 或無摩詞学 見宝唱録〔支謙訳〕 Móhē bānrě bōluómì zhòu jīng. Some texts lack the Móhē characters; see the Bǎochànglù; (translator Zhi Qian).

B 欠 Missing.

C 摩詞般若波羅蜜大明呪經 亦云摩訶大明呪經 初出与唐 訳般若心経等 同本見経題上〔羅什訳〕 C. Móhē bānrě bōluómì dàmíngzhòu jīng. Also called Móhē dàmíng zhòu jīng, first produced in the Tang. A translation of the Heart Sutra, See the same Sutra title above. (Translated by Kumārajīva).

Like Fèi Chángfáng and unlike the earlier catalogues, Zhìshēng considers the texts without Móhē to be a variant title rather than a separate text. 

Text C, the Dàmíngzhòu jīng, is extant and included in the Taishō as T 250. This entry in the Kaiyuan Catalogue is the first mention of the text in history. The Dàmíngzhòu jīng is not included amongst the translations of Kumārajīva in any older catalogue. And this means that it was almost certainly not by Kumārajīva. Indeed, this has long been the consensus. Back in 1932, when listing all the Prajñāpāramitā texts, Matsumoto Tokumyo (1932: 9) noted Er hat aber dieses Sūtra nicht übersetzt “But he has not translated this sutra”. Conze adds the detail that it was translated by one of Kumārajīva's "disciples" a theme recently taken up by Charles Willemen in a series of rather silly articles. Willemen asserts, on the flimsiest evidence imaginable, that Dàmíngzhòu jīng was translated by Zhu Daosheng. But he presents no plausible evidence for this assertion. Indeed, we know that the Heart Sutra per se is not a translation. It was composed in Chinese, in the middle seventh century (actually between 654 and 656 CE).

There is no doubt that this entry in the Kaiyuan Catalogue, dated 730 CE, is also the source of the conflation of the shénzhòutexts with the Dàmíngzhòu jīng, and combined with the idea that Zhi Qian translated the Móhē shénzhòu text, it explains why some people believe in a lost translation of the Heart Sutra by Zhi Qian. To be clear, no such thing ever existed and the evidence for it was always weak.

From the absence of the Dàmíngzhòu jīng in earlier catalogues we can also infer it was composed after the composition of the Xīn jīng. And Watanabe adds that it was not translated from Sanskrit, but composed in Chinese. Thus not only is the Dàmíngzhòu jīng not a translation, it is not (and could not be) a translation by Kumārajīva.


Watanabe (1990) concludes from this that the idea of a lost translation of the Heart Sutra by Zhi Qian was simply made up. The text in question was not a Heart Sutra and was not associated with Zhi Qian. Moveover the Dàmíngzhòu jīng attributed to Kumārajīva was not associated with him, was not even a translation, and was produced after the Xīn jīng

There is no reliable evidence of the Heart Sutra prior to the 650s CE. Moreover, Jan Nattier (1992) showed that the Sanskrit text was a back-translation from Chinese. The first mention of a Sanskrit text is in Woncheuk's commentary, but it is vague and could be a reference to the Sanskrit Large Sutra, since Woncheuk knew that to be the source of most of the copied passages.

All attempts at pushing back the existence to dates earlier the seventh century fail for lack of evidence. The oldest physical evidence of the Heart Sutra from anywhere in the world, is the inscription from Fangshan (see Attwood 2019) dated 13 March 661. The oldest literary mention occurs in letter dated 26 Dec 656, reproduced in Yancong's hagiography of Xuanzang (T 2053), but also preserved independently (See Kotyk 2020). This gives us the terminus ante quem. The earliest commentaries are Chinese texts from the late seventh century by Kuījī (T 1710), Woncheuk (T 1711), and Jìngmài 靖邁 (X 522). Note that the latter has received almost no scholarly attention.

We find evidence of the Heart Sutra in Tibet from roughly the eighth century, though this date is dependent on the attribution of Tibetan commentaries to Indian authors, some of whom are otherwise completely unknown, and some of whom are the most famous Buddhists who ever lived. And from India? There is no evidence of the Heart Sutra from India. No manuscripts, no inscriptions, no mentions in other texts. This is consistent with what we expect given that the Sanskrit text is a back-translation made in China.
All the evidence points to the same conclusion: The Heart Sutra was composed in Chinese ca 654–656 CE, using copied passages from Móhē bānrě bōluómì jīng «摩訶般若波羅蜜經» (T 223) and a dhāraṇī from the Tuóluóní jí jīng «陀罗尼集經» (T 901) translated in 654 CE (giving us the terminus post quem).

It's interesting that translators like Red Pine and Tanahashi have drawn on Japanese scholarship where it suits their purposes, but have entirely ignored this very important work by Watanabe. The false idea of the lost translation by Zhi Qian plays into their anxieties about the authenticity of this sutra that is not a sutra. And they employ the idea uncritically despite a long standing consensus around Watanabe's solid debunking of it. It turns out that, despite being very popular, both Red Pine and Tanahashi belong with D. T. Suzuki and Edward Conze as unreliable guides to this text. 

I have produced a draft English translation of Watanabe (1990) and uploaded it for comment on academia.edu. I will soon submit an article to an academic journal that discusses this material.



Attwood, J. (2019). "Xuanzang’s Relationship to the Heart Sūtra in Light of the Fangshan Stele." Journal of Chinese Buddhist Studies, 32, 1–30.

Kotyk, Jeffrey. (2019). “Chinese State and Buddhist Historical Sources on Xuanzang: Historicity and the Daci’en si sanzang fashi zhuan 大慈恩寺三藏法師傳”. T’oung Pao 105(5-6): 513.

Watanabe, Shōgo. 1990. “Móhē bānrě bōluómì shénzhòu jīng and Móhē bānrě bōluómì dàmíngzhòu jīng, As Seen in the Sutra Catalogues.” Indogaku Bukkyōgaku Kenkyū 39-1: 54–58.

Extant Chinese Bibliographies

  1. Chūsānzàng jìjí «出三蔵記集» (T 2145), compiled by Sēngyòu 僧祐 (445–518 CE)
  2. Zhòngjīng mùlù «衆經目錄». (T 2146), compiled by Fǎjīng 法經 (594 CE)
  3. Lìdài sānbǎo jì «歷代三寶紀» (T 2034), compiled by Fèi Chángfáng 費長房 (597 CE).
  4. Zhòngjīng mùlù «衆經目錄» (T 2147), compiled by Yancong 彥悰 (602 CE)
  5. Zhòngjīng mùlù «衆經目錄» (T 2148), compiled by Jìngtài 靜泰 (663-665 CE)
  6. Dà Táng nèidiǎn lù «大唐内典録» (T 2149), compiled by Dàoxuān 道宣 (664).
  7. Gǔ jīn yìjīng tújì «古今譯經圖紀» (T 2151), compiled by Jingmai 靖邁 (7th century).
  8. Dàzhōu kāndìng zhòngjīng mùlù «大周刊定衆經目錄» (T 2153), compiled by Míngquán 明佺 et al. (695).
  9. Dà Táng kāiyuán shìjiào lù «大唐開元釋教錄» (T 2154), compiled by Zhìshēng 智昇 (730)
  10. Zhēnyuán xīn dìng shìjiào mùlù «貞元新定釋教目錄» (T 2157) compiled by Upāsaka Yuán Zhàozhuàn 照撰, (800)

14 July 2023

Meier's Historicity Criteria

Historians of Buddhism face a difficult task, since the further we go back in time the sketchier the evidence of Buddhism is. Before about 500 CE there are few certainties and this is ostensibly a millennium after the life of the Buddha. Let me give an apposite example. It is widely asserted that the Pāli suttas are the oldest Buddhist texts we possess. This is largely based on a tradition found in the Dīpavaṃsa (a religious text) that the suttas were written down in Sri Lanka around the beginning of the Common Era. That is to say, the idea that Pāli texts were being written down at that time comes from a Pāli text; it comes from within a religious tradition and has not been corroborated.

When we turn to archaeology, however, we get a very different story. The oldest extant Pāli document is from the sixth century CE, while the oldest complete sutta is from the ninth century CE. By contrast, we have a partial and damaged manuscript of the Aṣṭasāhasrikā Prajñāpāramitā from the first century CE. The oldest extant Prajñāpāramitā text is some 500 years older than the oldest Pāli text. Moreover, Prajñāpāramitā was also likely an oral tradition at first and, in my opinion, its origins lie in the pre-Buddhist forms of meditation leading to the cessation of sensory experience.

So which literature is really older? Which evidence is more reliable when it comes to history, religious tradition or archaeology? These are the kinds of questions that historians of religion seek answers to. The efforts broadly fall into two main processes: hermeneutics and criticism. Hermeneutics seeks to establish the authority of the text, based on its history, sources, and relations with tradition. Criticism, sometimes "higher criticism", seeks to establish the authenticity of the text based on the language and interpretation of the text.

Unfortunately, when historians of Buddhism write about this, each of them takes an ad hoc approach, eschewing formalised methods of evaluation and interpretation. This can rise to the level of a rejection of the concept of "methodology". My mentor Richard Gombrich, for example, likes to rail against methodology and says that his method is simply "conjecture and refutation". (The young Gombrich is credited as a proof-reader in Karl Popper's 1963 book Conjectures and Refutations). There's nothing wrong with this, per se, but it is too general to be useful in defining methods. In dealing with, say, the historicity of a given person we are often commenting on individual phrases and even words.

In the typical Buddhism Studies history, each piece of evidence, each relevant phrase, is presented and evaluated on an ad hoc basis. In some cases the reader faces a veritable avalanche of "facts" that are supposed to persuade the reader of the authenticity and authority of the words in question. Since each data point is assessed on a unique basis, the critical reader has the extremely laborious task of addressing each point individually. The idea here is to overwhelm the reader with evidence that is difficult to assess.

The ad hoc nature of hermeneutics in Buddhism Studies can be juxtaposed with the highly formalised and structured approach found in Christian Studies. This is partly due to the Protestant doctrine that everyone should read and interpret the Bible for themselves, and partly to the very strong feeling that everyone should come to more or less the same conclusions about the Bible. One way to accomplish both is to specify what methods are valid.

There are some works on Buddhist hermeneutics, but, as far as I can see, these relate to how ancient Buddhists practiced hermeneutics, not to how we as Buddhists and/or scholars should evaluate the historicity of Buddhist texts in the present. Moreover, it has become fashionable for Theravādin monks to write apologia for the authenticity of the Pāli texts. These efforts are quasi-scholarly. The authors have massive, unacknowledged bias in favour of traditional Theravāda beliefs, but beyond this proceed in a scholarly fashion to produce quasi-scholarly works that confirm their most cherished beliefs.

We see the same lack of insight in Nāgārjuna studies where apologist scholars uncritically adopt Nāgārjuna's worldview and then argue for the modern relevance of his conclusions. We almost never see a Nāgārjuna scholar doing the basic philosophical work of identifying the underlying assumptions of Nāgārjuna and putting them to the test. Nāgārjuna's definition of "real", for example, insists that real objects are permanent and unchanging. Since no object of any kind is permanent or unchanging, no objects are real. What we think of as "the real world" is, in fact, merely an illusion or a delusion, or perhaps both but, anyway, nothing really exists, since everything is contingent on something else. As a definition of "real" or "reality" this is completely incoherent, since it assumes from the outset that "real" is a meaningless category. Those who claim that Nāgārjuna "has no view" are simply blind to the axioms of Nāgārjunian thought because they have internalised the dogma and no one ever talks about it. No one ever talks about the way that Nāgārjuna distorts Buddhist thought away from established norms of the time or what might have prompted this.

The point is that there are no formal hermeneutics for Buddhists or scholars to apply. The only attempt to articulate something like this can be found in Jan Nattier's (2003) book A Few Good Men. However, Nattier provides no references for her "hermeneutic principles" meaning that the task of finding out about them is left to the reader. While I have found it very useful to have Nattier's articulation of these principles, I now want to use them in my work. Citing Nattier would probably satisfy most Buddhism Studies scholars, but I wanted to know more about where they came from.

To cut a long story short, the ideas like "the principle of embarrassment" come from Christian theology, especially from Protestant New Testament scholars concerned with the historicity of Jesus. Once I gave up searching for this ideas in a Buddhism studies context, things got a lot easier. One source in particular was repeatedly cited by authors who write on this topic:

Meier, John P. (1991). A Marginal Jew: Rethinking the Historical Jesus. 4 Vol. New York: Doubleday.

In Meier we find a clear exposition and evaluation of the most common and/or useful criteria by which theologians have applied to the idea of an historical Jesus. While the obvious comparison might be the historicity of the Buddha, at present I'm more interested in Xuanzang as an historical character. The history of Xuanzang is typically—in a book like Sally Wriggins' (2004) The Silk Road Journey with Xuanzang—based on naïve readings of a narrow range of sources. I have been reading and making notes on a whole series of articles by Max Deeg critiquing the naïve use of historical sources for the life of Xuanzang. I'm also thinking about Jeffrey Kotyk's (2019) article that compares Buddhist and state histories and argues that, at the very least, we cannot understand Xuanzang as an historical character without considering all of the evidence. And into this mix has dropped the fascinating book chapter by Liu Shufen which argues for a very different picture of Xuanzang's later years under Emperor Gaozong.

Briefly, Liu (2022) argues that Xuanzang was held under house arrest by Gaozong for a number of years, and that during this time he was denied qualified assistants to help with translations. Indeed, his translation output dropped precipitously after the death of Emperor Taizong in 649. Moreover, Gaozong appointed a board of censors empowered to change Xuanzang's translations as they saw fit. These observations are based mainly on passages in the hagiography of Xuanzang attributed to Yancong (T 2053).

What I eventually want to do is write an article in which I analyse the historicity of Xuanzang according to criteria that I will state at the outset and apply even-handedly. To this end, I will here articulate Meier's hermeneutic criteria and give examples from my own published works or closely related publications. Meier articulates five primary and five secondary criteria which are tuned to recovering reliable historical information about Jesus from the Bible (and thus will require some tweaking for use in a Buddhist context).

Meier's Criteria

Primary criteria

  1. the criterion of embarrassment
  2. the criterion of discontinuity
  3. the criterion of multiple attestation [aka corroboration]
  4. the criterion of coherence
  5. the criterion of rejection and execution [specifically related to reports of Jesus violent death]

Secondary criteria

  1. the criterion of traces of Aramaic
  2. the criterion of Palestinian environment
  3. the criterion of vividness of narration
  4. the criterion of tendencies of developing synoptic tradition
  5. the criterion of historical presumption.

Primary Criteria

1. Embarrassment

When religieux write about their religion, they quite naturally portray it in the best possible light. Accounts of historical figures become hagiographies (from the Greek hagios "sacred, saintly"), that is, they are idealised accounts in which the characters become idealised vehicles for the values of the author. Thus in many Buddhist texts the Buddha is unfailingly good. He is not simply good, but his behaviour in the stories defines for Buddhists what good behaviour is. Scholars call this aspect of texts normative; it defines the norms of behaviour for Buddhists. This does not mean that Buddhists ever behaved the way the Buddha is portrayed as behaving. Indeed, there is ample evidence in the Vinaya to suggest that Buddhist monks were every bit as problematic as other human beings (I'll come back to this).

So, for example, when we read (per my 2004 article on the attitudes to suicide in Pāli suttas) that the Buddha taught a group of monks to meditate on death before going on retreat. On returning from retreat the Buddha finds that the monks have all committed suicide. Whoops. This reflects badly on the Buddha. It is an embarrassment. As a commentator, Buddhaghosa was obviously embarrassed, for example, and concocts a fanciful story in which it is secretly a good thing that the Buddha taught those monks the practice that drove them to suicide.

Another example of this is the story from the Ambaṭṭa Sutta in which the Sakya tribe are revealed to have practiced sibling incest marriages in the past (see Attwood 2012). Across Indian culture there are very strong social prohibitions on sibling incest (though in some parts of India first cousin marriages were and are common). This story doesn't simply reflect badly on the Buddha's tribe. It is one of the most taboo of all social conventions. "Sister-fucker" (Hindi bhenchod) is a very heinous insult for a man in modern India. So no one is going to simply make up a story about their ancestors practising sibling incest.

The criterion of embarrassment applied here makes the idea that Sakya's practiced sibling-incest marriages seem plausibly historical. In my article, I contrasted the strength of the incest taboo in India with the practice of sibling-incest marriages amongst Persian royalty as a way of consolidating power. Persian kings marrying their sisters may well have been following the example of Egyptian royalty who also practiced this.

However, this criterion has limitations. Clear cut cases of embarrassment are rare in religious literature for obvious reasons. Even if the embarrassing fact is true, there is no need to blurt it out. There are things that we simply don't talk about. So while we think this criterion is useful, there are not that many times we can apply it.

Moreover, what seems embarrassing for us may not have been embarrassing for the author. It is common to see people who encounter the myths of Buddhism to see the going forth of the Buddha in the standard account of his life (according to say the Lalitavistara or the Mahāvastu) in a negative way; the Buddha is seen to abandon his wife and baby to go off and become a wandering mendicant. This is sometimes seen as an indictment of his moral character. Even if we grant historicity to the myth of the Buddha we have to see the leaving home in context. In that version of the story, the Buddha lives with his family in a palace. The wife naturally, according to millennia old practice, goes to live with her husband's family. In that palace, Yasodharā lacks for nothing. She lives in luxury. Later, in this version of the story, the Buddha returns home and teaches Yasodharā and Rahula how to be liberated themselves. So it all turns out well in the end and there is no "embarrassment". We might go further and point out that other sources of biography, notably the Ariyapariyesanā Sutta, make no mention of the wife and baby, the Buddha appears to be an unmarried youth and, what's more, his mother lives to see him leave home rather than dying in the first week of his life." It is unlikely that the authors of either account thought of these details as embarrassing. It is only modern sensibilities that make it so.

2. Discontinuity

The idea here is to identify words or deeds attributed to an historical character that cannot be derived from their milieu or from later development of their religion. If we could identify such speech or actions in Buddhist texts, we might feel justified in supposing these to be attributable to the authors of the texts.

The problem for Buddhism is that its not a linear development from an existing culture. The authors of the texts attribute authorship to the Buddha. We only know the Buddha from normative Buddhist texts, since there are no contemporary records from other traditions that mention Buddhism or Buddhists.

The early Buddhist texts are the "Old Testament" of Buddhism, so, unlike New Testament scholars arguing about Jesus, there is no clear backdrop from which the Buddha emerged. That said, it is obvious that early Buddhists did draw on existing culture in formulating Buddhism. The presence of deva and asura figures is clear evidence of Brahmanical influence. The Buddhist doctrine of karma is likely an adaptation of an existing Jain doctrine. Mythic figures such as Yakkha and Nāga seem to be non-Indo-European and borrowed from some pre-existing chthonic animistic tradition. As I have noted (along with Anālayo), it seems very likely that key Buddhist meditation practices (leading to the āyatana states and to śūnyatā-samādhi) were also borrowed from a pre-existing tradition. There is also Michael Witzel's idea, articulated in my article (2012), that Buddhists borrowed from Iranian tradition.

This criterion will be much more difficult to apply in a Buddhist context, simply because the relations between Buddhists and their milieu were complex, but reliable sources don't exist. We are guessing a lot of the time. Moreover, the literature that later traditions produced was huge compared to the Bible.

A potential example is the word paṭiccasamuppāda. This word does not derive from existing traditions, so far as we have evidence of them. Such evidence as we have is extremely sketchy, however, and arguments from absence are weak. However, what about this word insists that the Buddha rather than the followers of the Buddha coined it?

Another problem that I don't think theologians face is that the first literature we meet is already translated, edited, and organised. In other words, the early Buddhist texts are the product of considerable literary effort on the part of early Buddhists. As already noted, the actual dates of the composition of the Pāli sutta are simply guesses based on normative traditions. The archaeology of Pāli texts does not begin until the 6th century CE.

3. Multiple Attestation or Corroboration

This criterion seems fairly self-evident: a historical event is more plausible if it is mentioned in two or more different sources.

A feature of Liu's (2022) thesis about Xuanzang is that the Yancong Biography contains numerous "facts" about him that are not recorded elsewhere. Liu appears to argue that in these cases the Biography is more valuable as an historical source. The criterion of multiple attestation, or what Nattier calls the "principle of corroboration", tells us the opposite. Where details are included in the Yancong Biography but not corroborated by other sources we must judge such details to be less plausible. This is is also Jeffrey Kotyk's (2019) criticism of the use that scholars have made of the Biography.

On the other hand, Kotyk (2019) notes the existence of another source for an event that is recorded in the Biography, namely the gift of a Heart Sutra in gold ink, accompanied by a presentation case (dated 26 Dec 656). This event occurs in a letter (often referred to by Sinologists as a "memorial") sent by Xuanzang to Gaozong to celebrate the first month of life of a royal prince born to Wu Zhao. The same letter is preserved as part of a collection of Xuanzang's letters preserved independently in Japan (seemingly not extant in China). Since this constitutes corroboration, Kotyk is minded to accept this event as plausibly historical. I agree. Furthermore, this is the earliest reliably dated attestation of the Heart Sutra anywhere in the world. The earliest artefact is from a few years later in the form of an Chinese inscription commissioned on 13 March 661.

4. The Criterion of Coherence

This criterion only comes into play after at least a first pass using the first three criteria. The aim is to establish a baseline of style or behaviour that allows us to compare a new passage with our existing understanding and judge how well it fits into the existing whole.

In a sense, this is a particular case of what Hans-Georg Gadamer calls the hermeneutic circle. Understanding the whole requires us to first understand the parts. But in order to fully understand each part, we have to know what contribution it makes to the whole. Each iteration clarifies our understanding of how it all fits together and what contribution each part makes.

The problem for Buddhists is, again, the complexity of Buddhist sources.

5. The Criterion of Rejection and Execution

This one is clearly specific to Christianity. The idea here is that the Jewish and Roman authorities would not have condemned and executed Jesus if they did not have cause. However, I'm not entirely sympathetic to this one. The Romans in Palestine are an example of authoritarian control by an foreign occupying power. Authoritarian regimes don't need logical reasons to persecute groups or individuals. As we know, such regimes routinely carry out show trials in which innocent individuals are convicted of crimes and punished. The idea that the Romans executed Jesus for cause is part of Christian mythology, but it's not corroborated by Roman sources. Tacitus is often mentioned as a source, but his mention of Jesus being executed by Pilate was written in 116 CE, a century after the fact. And Tacitus does not mention his sources.

A parallel here might be discerned in Buddhist discussions of the Buddha's last meal. The Parinibbāna Sutta appears to show him eating some dish which gave him food poisoning and led to his death. The criterion of embarrassment makes this detail seem a little more plausible, since being killed off by food poisoning is a rather ignominious death for a saint.

Secondary Criteria

1. Traces of Aramaic

The New Testament was written in koine Greek. Thus when we see traces of Aramaic, supposedly the language spoken by Jesus, this makes a given statement seem more authentic. As Meier immediately says, there are "serious problems" (178) with this criterion.

We see a similar analysis of Pāli, in which we often see evidence of another Prakrit which is close to the language used for the Asoka edicts in Magadha and for this reason called Magadhī. We see examples of Magadhī case endings in particular.

As in the case of Aramaic features in the Greek New Testament, the occurrence of Magadhī case endings in Pāli has been interpreted as evidence of the antiquity and therefore authenticity of the texts where they occur.

2. Palestinian environment

The idea here is that details in stories about Jesus that reflect the known customs, beliefs, and practices associated with Palestine in the first century, are more likely to be authentic. But it is easy for an author to interpolate such details from their own memory, from oral tradition, or from other sources.

The problem here, for Buddhism, is that we only have normative Buddhist sources for this. There is no contemporary literature that we can draw on. All we have for comparison is scant and fragmentary evidence from archaeology. Buddhist Stories are, for example, set in real cities. We know these cities existed because we have the remains of them. We can still visit the ruins (as I have done).

This criterion is more useful negatively. For example, we may say that the appearance of foreign ideas and customs should make us sceptical. As noted above, some of my speculations about the possibility that certain details of Buddhism might be explained if the Sakya tribe had originally come from Iran. Note that this is not the same as relating the Sakya tribe to the Iranian Saka tribe (no, the Sakya were not Scythians).

3. Vividness of narration

This criterion supposes that the kinds of vivid descriptions that we see of Palestine in some Biblical stories indicate an eyewitness account.

The problem here is that supplying vivid details is something that any competent storyteller does to make their account seem more plausible. The example that immediately comes to mind is the fact that Sherlock Holmes lived at 221B Baker Street, Marylebone, London. This address was not real at the time, although the road has since been extended and now does have a 221. The Sherlock Holmes Museum can be found at 239 Baker St. Again, museums are generally the place where we store and display historical artefacts. But Sherlock Holmes is ahistorical, he's a fictional character.

In his paper Has Xuanzang ever been in Mathura?, Max Deeg notes that a number of vivid accounts in Xuanzang's Record of the Western Regions, are not based on first hand accounts. Xuanzang did not go to Mathura, despite providing a vivid description of it.

4. Tendencies of developing synoptic tradition

Maier Considers this principle "highly questionable". This is because it involves subjective opinions about the internal development of the gospel literature. There were, for example, more Aramaic expressions and syntax in the earlier gospel by Mark, that were gradually eliminated in Matthew and Luke.

There is an exact parallel with Buddhist literature featuring Prakrits gradually being Sanskritised. Early Buddhist texts were composed in various Prakrits, but these tended to become standardised and increasingly Sanskritised. In some texts there is an argument as to whether they reflect a Sanskritised Prakrit, or a Sanskrit with some Prakrit features. And of course there are, by the fourth century, texts composed in Pāṇinian Sanskrit. And after this, Buddhists once again began to use Prakrit. Notably, many dhāraṇī appear to have been composed in a Sanskritised Prakrit, i.e. they have Sanskrit words, with an archaic Prakrit case ending in -e. A good example is the dhāraṇī in the Heart Sutra which uses the Prakrit masculine nominative singular -e ending: gate gate pārasaṃgate bodhi svhāhā. Note that bodhi is also in the nominative singular and svāhā is now considered indeclinable.

The problem for Buddhists, as for Christians, is that constructing such internal chronologies that have no corroboration from external sources is speculative and subjective. All too often Buddhists work on the axiom "older is more authentic" and the people claiming to have discovered internal structures and developments are apologists for a sectarian tradition whose identity and livelihood are tightly bound to the outcome of the inquiry. No Theravādin bhikkhu, for example, is going to undermine Theravāda orthodoxy, since in doing so they risk expulsion from their order or worse, they risk talking themselves into apostasy.

So, for example, some Pāli texts contain elements of a living Prakrit that is similar to the language of the Asoka edicts from Magadha. We call that language Māgadhī. This is the Prakrit that emerged south of the Ganga around the kingdom of Rājagāha, then based in what is now Southern Bihar. We get many hints that there was considerably more linguistic diversity in the second urbanisation period ca. 600 BCE - 200 BCE. Arguments are made that the persistence of elements of Māgadhī in Pāli is evidence of antiquity amongst texts. These are combined with other speculative ideas, such as the idea that the use of certain poetic meters was prevalent in older material.

The problem is that the Pāli that has come down to us is the product of both translation (from various Prakrits into Pāli) and repeated editing (which is inevitable in the process of compilation). The Pāli suttas, for example, may well have relied on older traditions in some form, but they reflect Buddhism at the time the suttas were compiled and, even more than this, they reflect Buddhism at the time the suttas were written down. The trouble is that we don't know when Pāli was written down because the only evidence we have is from normative religious texts that have not been assessed using standardised hermeneutic principles.

5. Historical presumption

The last criterion that Meier considers concerns opinions about where the burden of proof lies. One of the things that Richard Gombrich has said about this issue was: "We have no reason to doubt the traditional accounts." For Gombrich the burden of proof lies with those who distrust the normative sources. But this is problematic. If, for example, we take Gombrich's assertion that the Buddha was an historical person, how can we prove that he was not? We cannot prove a negative. Many religieux take a similar position to the historicity of the Buddha, claiming, for example, that the very existence of Buddhism or the literature of Buddhism, is all the proof we need that someone called Buddha lived.

However, Meier notes the standard objection to this approach: "The burden of proof is simply on anyone who tries to prove something" (183). For example, anyone who wishes to assert that the Buddha was an historical person bears the burden of proof.

And, on the contrary, we have the argument by David Drewes that the Buddha has never been linked to any historical fact or event. There is no archaeology associated with the Buddha, and precious little even from that time beyond some sherds of pottery. The argument David Drewes makes is not that the Buddha was not historical. You cannot prove a negative. Drewes argues that there is simply no basis on which to make a conjecture because he does not consider the Pāli suttas to be a source of historical information. The suttas are normative religious documents that primarily reflect the religious ideas and ideals of the authors of the stories. If they are supposed to reflect the religious ideas of someone other than the author, then let us see that evidence. And let us evaluate the plausibility of that evidence according to standardised criteria.


While not all of Meier's criteria are directly applicable to reading Buddhist texts, we have here a coherent account of how some scholars have tried to work with the literature of early Christianity to divine the historical Jesus. In most cases I can see how I would apply such criteria to working with Buddhist normative texts. And have given examples.

These essays of mine are really just my notes typed up into a coherent form. The field of New Testament hermeneutics has thrown up many different lists of criteria and many arguments about which is applicable and to what extent.

The benefits of standardising our evaluations of the historical value of our sources have obvious advantages for scholars amateur and professional, because we have a common basis of evaluation. To date, a lot of the work on historicity is piecemeal; each fact is presented and justified in an isolated fashion so that a refutation of the whole requires a refutation of each fact individually. This makes the arguments for the historicity of the Buddha or the Pāli suttas very arduous and it makes the claims for "authenticity" sketchy, at best. This is the principle advantage of eschewing methods.

The rejection of standardised methods seems to be a conscious thing in Buddhist Studies as though our methods are value-free. A key social studies criticism of science, which also purports to be value-free, or at least value-neutral, is that it is theory laden. Science proceeds from certain axioms and we cannot pretend that those axioms don't matter. This criticism is entirely fair and identifying those axioms has helped to improve the scientific process and make scientific inquiry more rigorous.

It's hard to imagine a more theory-laden scenario, however, than a religious (often one who has given up sex for their religion) writing an apologia for one of the central tenets of their religion. The idea that their writing is value-free or that it is not theory-laden is simply laughable. As Upton Sinclair said: “It is difficult to get a man to understand something, when his salary depends on his not understanding it.” How much more so when that person, on the basis of their religious convictions, lives in extreme self-denial, is accorded exalted social status, and is praised sycophantically by followers.

And this is the problem that hermeneutic criteria were meant to solve. How can we be confident that anything we read in a religious text has any connection to history in the complete absence of any contemporary archaeology that might shed light on the situation. We examine claims, weighing them against our stated criteria, and then highlight what seems to be the most plausible explanation or narrative.

Such a procedure does not produce certainty, however. We have to treat it in a Bayesian way: using information gleaned from sources to assign finite probabilities to each scenario, and then adjusting the probabilities as new information comes in. Meier emphasises that no single criterion used in isolation would be useful. Rather, we have to view each new piece of information in relation to all the criteria.

Again, this process of going from part to whole and back to the parts in an iterative cycle was identified by Gadamer as the historiographical process par excellence. This is our best way forward in trying to reconstruct history from partial and fragmentary information produced by a community of religieux. And this process is never finished. New information is emerging all the time. And very often this means that we have to weigh up everything anew.

For example, the Chinese origins of the Heart Sutra now seems certain because there is a mountain of evidence for the Sanskrit text being a back-translation from Chinese. How does this knowledge affect other conclusions that we have about the Heart Sutra? If the text is not authenticated as a genuine, Indian, Buddhist text, then on what basis can it be authenticated? Or if we are being more provocative, we might ask, Is the Heart Sutra an authentic Buddhist text at all? If we don't have clear and agreed upon criteria for having such a discussion, then it tends to be a waste of time.

Of course, even within New Testament scholarship the criteria themselves are contentious and contended. Meier's work is widely cited in their literature. Meier values some principles (embarrassment, discontinuity, corroboration) and devalues others (historical presumption). Other authors come to different conclusions. But at least they know which argument they are having and what is at stake. My sense is that parallel discussions in Buddhism and Buddhism Studies are largely incoherent and slaved to sectarian orthodoxy. I think we can do better than a series of disconnected, idiosyncratic asseverations of faith. Certainly, in amateur and professional Buddhism Studies we need to pay more attention to methods and methodology (the study of methods).



Attwood, J. 2004. "Suicide as a response to suffering." Western Buddhist Review, 4.

Attwood, J. 2012. "Possible Iranian Origins for Sākyas and Aspects of Buddhism." Journal of the Oxford Centre for Buddhist Studies, 3, 47-69.

Liu, Shufen. (2022). “The Waning Years of the Eminent Monk Xuanzang and his Deification in China and Japan.” In Chinese Buddhism and the Scholarship of Erik Zürcher. Edited by Jonathan A. Silk and Stefano Zacchetti, 255–289. Leiden: Brill.

Meier, John P. (1991). A Marginal Jew: Rethinking the Historical Jesus. 4 Vol. New York: Doubleday.

Nattier, Jan. (2003). A Few Good Men: The Bodhisattva Path according to The Inquiry of Ugra (Ugraparipṛcchā). Honolulu: University of Hawai’i Press.

Related Posts with Thumbnails