29 June 2018

Critical Thinking in Buddhist Historiography

In this essay I'm going to try to show some of how I evaluate sources of information using a case study of the two mentions of the Heart Sutra in the standard biography of Xuánzàng written by Huìlì and Yàncóng. This involves a detailed assessment of such qualities as authenticity, veracity, accuracy, precision, and reliability. What knowledge can we obtain from a text, using which kinds of methods? What kinds of caveats apply to this process?

On the 5th day, 12th Month of Yǒnghuī 6 (永徽六年十二月五日) i.e., 6 January, 656, Xuánzàng sent a letter to the emperor celebrating the birth of a new prince the month before. He wrote in a letter: "I dare to offer a copy of the Prajñā Heart Sutra in gold letters, one scroll and a case." (輒敢進金字《般若心經》一卷并函 T 50.272b.12). This piece of information is quite a big deal because this is the earliest literary reference to the Heart Sutra with a precise date that I know of. It is five years before the earliest physical evidence (661 CE) and definitely before Xuánzàng began translating Prajñāpāramitā texts in ernest (660 CE). But this important date is seldom, if ever, discussed in the literature on the Heart Sutra

The information comes from the 《大唐大慈恩寺三藏法師傳》, a Tang Dynasty (唐) biography (傳) of Xuánzàng (aka "the Dharma Master 三藏法師傳 of the great Ci'en Temple 大慈恩寺") by 慧立 Huìlì and 彥悰Yàncóng from about 688 CE (henceforth, Biography). I say "biography", but as Xuánzàng never puts a foot wrong and is portrayed as conforming to an ideal, the text is clearly part of a trend to elevate him to the status of Buddhist saint. As such, we might better refer to it as a hagiography. Chinese Buddhist saints are quite different in character from Indian Buddhist saints, which is something that requires its own study (and I don't propose to get into it here). Unlike some hagiographies, the Biography was composed quite close to the subject's lifetime, in a literate society that kept good records in both the religious and imperial spheres, and partly by someone who knew the subject personally. So Xuánzàng is not presented as doing many miracles, but more as behaving in an exemplary manner in social, political, and religious spheres. He is, in short, the archetype of a good Buddhist living in a fundamentally Confucian society.

The Biography is included in the Taishō Edition of the Tripiṭaka as text no. 2053. T2053 was translated by Samuel Beal in 1911 (I use a 1914 reprint), and more recently by Li Rongxi  for the BDK English Tripiṭaka (1995). Huìlì was a younger contemporary of Xuánzàng, who knew and worked with him. Yàncóng recounts the  story of how the Biography came about in a preface. Having written down the the Biography, Huìlì lost confidence and buried them "fearing that the contents of his writing might be incomplete and inadequate for the public" (Li 1995: 3). However, on his death bed he asked his students to dig up the manuscript. Within a few years it had become divided and scattered. Some parts were lost. Around 688 CE, the students managed to collect up the remaining works and commissioned 彥悰 Yàncóng to edit it into a book. What remained of Huìlì's work amounted to five scrolls, and Yàncóng added five more. The result is the now famous biography. 

However, there is a slight difference between how Li tells the story in his translator's introduction and how Yàncóng tells it. Li suggests that Huìlì's account ended with Xuánzàng's return to China in 645 CE. If Yàncóng does not supply this detail, then who does? It probably originates from a seemingly off-the-cuff comment by Beal in his introduction:
The five chapters added by [Yàncóng] are probably those which follow the account of [Xuánzàng]'s return from India, and relate to his work of translation in China. (1914: xix)
As far as I can tell, there is no basis for this "probably". Beal is guessing. What Yàncóng says in Beal's own translation is that he was engaged to "to re-arrange and correct the leaves which their master had written and hidden in a cave." (xix).  But compare Li "Then I mixed the original text with supplementary annotations, extending it into a work of ten fascicles" (1995: 9). What this suggests to me is that Huìlì provided the framework for the whole text and Yàncóng expanded on it in general.

Li is not beyond adding little details to pad out the story. Take the example of the famous story of the sick man giving Xuánzàng the Heart Sutra, which occurs earlier in the Biography:
慜將向寺施與衣服飲食之直。(T 50.224b.8-9)
Li: "With a feeling of pity, he took the man to his monastery and gave him money to purchase clothes and food." (26).
However, when I parse this sentence what I get is:
慜 benevolent 將 will 向 to 寺 temple 施 bestow 與 give 衣服 clothing 飲食 food and drink 之 going 直 straight 。
i.e.,  feeling benevolent, [Xuánzàng] took [the sick man] straight to the temple, and gave him food, drink, and clothing. [my trans]
There is no mention of "money" or "purchasing" in the text. And, given that we generally understand monks as not handling money, this added feature is incongruous, although, of course, some monks did handle money (and I haven't checked which kind Xuánzàng was). So the translation here is more precise than the source text (it supplies extra details), but it appears to be inaccurate (because the details don't arise from the source text).

I started out assuming that Li must be fairly reliable to be employed as a translator by the prestigious BDK organisation. Based on assessing these two details, I conclude Li is less reliable than my initial assessment of him. Samuel Beal's translation is, "Pitying the man he took him to his convent, and gave him clothing and food" (1914: 21). Beal, writing in a different era, uses language with an archaic feel to it. We might also wonder why he chose "convent" (usually associated with nuns in English) for 寺. On the whole, however, his translations seems more more accurate, and thus more reliable than Li at this point.

How does anyone assess when a translator is more or less reliable without reference to the source text in the source language? Very often my friends and colleagues make a kind of aesthetic judgement on the grounds of which one reads better. Often the more "poetic" a translation is the more folk like it. Which seems like nonsense to me. Is it fair to judge a translation on a handful of sentences? When we are only interested in a handful of sentences, the importance of them with respect to the text as a whole is magnified.

In this case Li's small amendment to include a reference to giving money has a major impact on my study. Had I taken Li seriously, I might have decided that I needed to spend time looking into the issue of use of money by monks in medieval China. This is a potentially fruitful avenue to go down if it relates to my task at hand, but the fact is that it doesn't. It's just a lapse on the part of a translator. Fortunately, I like to try to see how source texts look before taking translations seriously. Any reader who does not check the source text (for whatever reason) is always at risk of being misled.


譯 = Translator?

One of the interesting things about the Biography is the style of the attribution. In some past explorations of the attribution of the Heart Sutra we've seen that 譯 means "translator". Dictionaries are quite unequivocal on this point. However, in the Biography the attribution is:
沙門慧立本 譯彥悰箋.  
"Originally composed by Monk Huìlì, edited by Yàncóng, with annotations."
Here 沙門* means that Huìlì (慧立) was a Buddhist monk. The character 本 means "origin, root" and tells us that the text originates with Huìlì.  Next, 箋 means "annotation" or "commentary". We know from Yàncóng’s (彥悰) preface that his role was that of editor of Huìlì’s manuscript and notes, 譯 here must mean “editor/edited”, rather than “translator/translated”. However, this work  was composed in Chinese, so 譯 cannot mean "translator" as nothing was translated. This is an important detail because it contributes to another aspect of my work on the Heart Sutra: the relationship between Xuanzang and the Heart Sutra. When I checked I found that Xuánzàng is credited with being the 譯 of his own travelogue, also composed in Chinese. 
* 沙門 is pronounced like 'shaman', deriving from Prakrit ṣāmana (Sanskrit śrāmaṇa). There is a possible link with English shaman: Sanskrit śrāmaṇa → Prakrit ṣāmane → Middle Chinese 沙門  ʃamuən → Siberian/Tungus šamān → Russian shamán → German schamane → English shaman (attested 1698).
It might be fair to assume that if many texts that are translations refer to a person as the 譯 in relation to it, then that person is the "translator" and that 譯 must mean "translated [by]". But we have two examples of the character being used in contexts where it cannot mean "translate". The dictionary  definition seems to be incomplete. There is another sense which is something like "worked on", which is distinguished from "authored".

Keeping track of such small details is integral to this kind of work, because the accumulation of details is what adds up to a case. The only problem is that some intellectuals tell us that no accumulation of details adds up to a case. 


Critical Thinking and Historiography

Advocates of critical thinking sometimes suggest that there is only one rational way to go about seeking knowledge, i.e., through refutation. This is supposedly based on Karl Popper's principle of conjecture and refutation. In this view, there is nothing to be gained by looking for confirmation or, in my case, the accumulation of details. We can never confirm a theory; all we can hope for is to refute it. This, they say, is because of the so-called black swan effect. The story goes that Aristotle, when formulating his outline of logic, took it as axiomatic that "all swans are white". This allowed him to confidently construct deductive syllogisms like
All swans are white.
Bruce is swan.
Therefore, Bruce is white. ✓
Until one discovers that Bruce is Australian and he is actually a black swan and the deduction is false. The problem here, and with all deductive reasoning, is that it all revolves around axioms, i.e., propositions that one accepts as truth before proceeding to infer some new fact by deduction. If the axiom is false, then all inferences from it are also false. The argument proceeds to say that it doesn't matter how many white swans you meet, you can never be certain that all swans are white. It only takes one black swan to disprove the axiom that all swans are white and all inferences from the axiom fall apart.

The "black swan" argument is that you can never arrive at the truth through seeking confirmation of an axiom. Indeed, proving that any axiom is true is a very difficult thing to do. It is possible in mathematics. However, in any system of mathematics it is also possible for something to be true or false, but for this to be impossible to prove. So the search for truth quickly gets bogged down. And this is why scientists tend to avoid the idea of truth, and instead seek accurate and precise descriptions of reality (i.e., the day to day focus is on the epistemic aspects of their work). The attitude is "let the philosophers argue over the nature of that reality, as long as we can predict how it's going to behave". Scientists and philosophers are often dismissive of each other, largely because scientists stray into the area of speculating about the (ultimate) nature of reality (metaphysics) and because philosophers often speculate without reference to empirical knowledge - which is far and away the most reliable form of knowledge.

This critical thinking approach, call it hyper-critical thinking, leads to an impasse. It seems as if all claims to truth are either false now or soon will be. And thus it may seem that there is no point in even seeking knowledge, because in common sense and classical philosophy we equate knowledge and truth. Meanwhile, in the real world, very general rules of thumb turn out to be surprisingly useful in day to day life. We mainly get by on heuristics, or generalised approaches to solving problems that are good enough. Truth, as an abstract or an ideal, turns out to have surprisingly little practical value. Law courts, for example, use the heuristic of establishing something beyond reasonable doubt, which may be very far from the ideal of truth.

The hyper-critical approach to knowledge is a more or less useless strategy for studying history. History is always written from a point of view. That point of view includes the axioms that the historian explicitly accepts about history and historiography (the writing of history) as well as the implicit axioms they accept uncritically (bias, prejudice, cultural conditioning, etc.). In Justin L. Barrett's terms, historians, like everyone, have  reflective and non-reflective beliefs. And by now historians all know this. Very few historians, especially trained historians, ignore these problems. But just in case they do, few serious readers of history ignore these problems. We know that history is not "truth", but that doesn't matter. No one is much interested in truth in the absolute sense. History provides us with an understanding of events, from a point of view. Historians and readers alike know that multiple points of view are available. History is not science, much less abstract philosophy.

Equally, historians are aware that new information surfaces all the time. A history written in 1900 or 1950 is likely to be out of date for this reason. We would usually like our records to have been recorded as close to the events as possible, and our histories written as close to the present as possible. But the fact that there are always going to be new takes on history should not, and does not, paralyse historians, or prevent them from publishing. The black swan effect is a given. Two years ago I blogged an essay about "the oldest dated Heart Sutra" unaware that in Chinese academic circles an older version had been common knowledge for almost sixty years. Unaware of the fact, I continued to suggest that the oldest Heart Sutra was dated 672 CE right up until the last couple of weeks.  History is not only written by the winners, but it is rewritten by the better informed amongst the winners' descendents.


Approaching the subject

Coming back to the to passage from the Biography that I started our with: "Xuánzàng presented the Emperor Gāozōng with a copy of the Xīnjīn on 6 Dec 656." There are two ways to approach a statement like this.

On one hand, we may doubt the authenticity and veracity of the statement and look for ways to refute it. We may, for example, check that the dates coincide with other records of the reign of Gāozōng and the birth of the prince. We could check if there is any record of the Emperor receiving such a gift in the imperial records. Some documents from that time still exist in some form. We might query whether the conversion to the Gregorian calendar is accurate (since I used an online black-box converter this would be a good question). In this approach we think of Huìlì and Yàncóng as unreliable, motivated witnesses and we interrogate them like prosecuting attorneys. We try to pick apart their story. Some might argue that such a procedure is the only way to deal with historical sources. The Greek historian is known both as The Father of History and The Father of Lies. Pre-modern historians were not always critical when it came to their sources.

The other approach is to assume, in the absence of evidence to the contrary, that Huìlì and Yàncóng are at least sincere in their statements, that they have not set out to deceive us. They may themselves have been deceived, but Huìlì was contemporary with the subject and Yàncóng only one generation removed. Many living people who knew Xuánzàng would have been available to Yàncóng as witnesses. Also, one or both of them seem to have had access to official records of both state and religious institutions. Apparently, one or both had access to the correspondence between Xuánzàng and Gaōzōng (in the days before carbon paper). In this approach, we may look for corroboration of dates. In doing so we may turn to the very same sources as those who set off looking to refute the statement. We may look for a state record of receiving the gift, or a letter of acknowledgement from the Emperor. Such a letter is reproduced in the biography, but does it occur anywhere else.

The trick is not to ask is it true or false. We know at the outset that what we are seeing is not the truth in any abstract sense. We understand that someone is expressing their values through the medium of a biography (or hagiography). So we know to look at a text like the Biography as an anthropologist might. What interests us as historians is how reliable are our witnesses? What level of confidence should we have in them? What biases do they have? In this sense, good history is naturally Bayesian in its approach. We look at the givens and we make an initial assessment of the veracity. The different scenarios from complete falsification to existentially accurate and precise. Then we look into the matter, gather evidence and see how that affects our perceptions of the possibilities.

We will never establish an existential truth beyond the actual existence of the text we are studying. There is a biography and it is text no. 2053 in the Taishō. The accuracy of the authorship, the date, and provenance of it are all matters of conjecture. What we seek, rather, is a plausible account that, ideally, fits all the facts. If, for example, we know for certain that the Heart Sutra is not a translation, then we need to account for the stories that state Xuánzàng translated the text. This may involve factors such as the ambiguous use of language and the pious desire to connect Xuánzàng with the Sutra.


Precision vs Accuracy

One of the problems that we face is that the biography gives us a precise date: 永徽六年十二月五日. The precision is admirable and can, with some effort, be translated with equal precision into the more familiar Gregorian Calendar, 6 January, 656. Precision to the day might be undermined if all the other references to the event are less precise. If a dozen other texts say "sometime in 656" then the precision of Huìlì and Yàncóng might seem suspicious. In general, however, Chinese sources do keep track of events to this level of precision quite routinely. Two prominent exceptions are the commentaries on the Heart Sutra attributed to Xuánzàng's successors, Kuījī and Woncheuk. But since these almost certainly post-date the date of 656, they don't really matter in terms of establishing the provenance (except that Woncheuk appears to refer to a Sanskrit text). 

Other scenarios include the whole event being made up, i.e., high precision but completely inaccurate. It it might be that they got the day, month, or year wrong, causing inaccuracy at different orders of magnitude. A lot of history is written about at the level of precision of the year. For example, we often cite the year of birth and death for a historical figure: Xuánzàng (602-664). On the other hand, earlier in the Biography the authors suggest that the prince in question is born on the afternoon of the Month 1, day 5 and that Xuánzàng's gift was given in Month 2, day 5 giving us precision to the day. 

So, given a precise date, we have to think about how precise it is and how plausible that level of precision is; and how accurate it is. How would we know? Again, the approach is to look for either refuting or corroborating evidence: which either lowers or raises our level of confidence. What critical thinking does, is make it more likely that our confidence will fall to 0% than that it will rise to 100%. We can more easily be convinced that something is false, than that it is true. But most of the time we will be somewhere in the middle.

For example, I think it unlikely that Xuánzàng translated the Heart Sutra from Sanskrit into Chinese. The preponderance of evidence suggests that the Heart Sutra is not a sutra at all, but a sutra extract. The practice of copying out extracts is distinctively Chinese. Also, Nattier has shown that extraction was done in Chinese, from Kumārajīva's translation. The story in the Biography makes it seem likely that Xuánzàng received a Chinese text, before he left for India and learned Sanskrit. And the date of 656 CE from the Biography suggests that he had the text before he started to translate the Prajñāpāramitā texts in 660 CE. There is a story that he translated the Heart Sutra in 649 CE, but this first appears some centuries later and is quite obviously apocryphal. So any story we tell about the man and the text, has to fit all these points. And we must ignore that fact that many uncritical authors have told other stories (the 649 CE date is repeated as a solid fact uncountable times). 


Doing Historiography

So, despite what critical thinking nerds might say, it absolutely makes sense to look for confirmation as long as one does it in the right spirit. As historians, we pile up evidence  and then try to weave a narrative in which all the evidence is accounted for. We tell stories in the full knowledge that next year or tomorrow, some new piece of evidence may turn up that changes the story. And we (mostly) acknowledge our biases. No one is pretending that History is a science, though sometimes it may approach being a kind of philosophy.

Histories are always constructed on partial information. The historical record is patchy, though it is often better in China than almost anywhere else because the Chinese were literate and kept records. Knowledge is always partial in any case, but as the centuries pass such records tend to degrade. So while we have what we think are reliable copies of the Biography composed by Huìlì and Yàncóng, the kinds of records against which we might look to evaluate the biography often don't exist (such as the correspondence between Xuánzàng and the Emperor). Which is not to say that evidence never existed, although sometimes this may be the case. As the saying goes "Absence of evidence, is not evidence of absence." 

Under these conditions, sitting on the fence and being a tooth-fairy agnostic is not interesting at all. To participate, one has to get off the fence and join the discussion. This is why historians write histories with conviction. As Mercier and Sperber have observed, when making a case, it is natural, reasonable, and rational to make the best case possible and then see what others say. History is not a solitary, one-time occupation, it is an ongoing, collective effort. At any given time a small number of people will be putting forward stories constructed as the strongest case they can make (harnessing confirmation bias) and a majority will be sitting back and arguing over the alternatives. Fundamentally, reason is both collective and argumentative. And so is history. 

Another problem is the motivation of the witnesses. The donor of the Fangshan stele states that his desire is for his family to attain awakening by donating the merit of making the stone sutra to that end. We can probably relate to this. However, what was the motivation of the stone carver, or the monastery who employed him? We don't know. We do know that Chinese monasteries were often extremely wealthy as donors sought to mitigate misfortune or buy their way into Heaven. These carved sutras with donor inscriptions are a bit like the Roman Catholic Church selling indulgences -- make a big enough donation and your sins will be forgiven. Monasteries also engaged in usury, farming, and manufacturing to generate income. Do these motivations give us any reason to doubt the details of the artefact or the biography? Sometimes the adage, "follow the money" is apposite in historiography.

Is the association of the Heart Sutra with Xuánzàng historical or is it legendary? We might want to ask the question, who benefits from the association? 

One thing that is clear is that, in 7th Century China, insisting that Heart Sutra was a translation from an Indian text would have added an air of authenticity. The implication was that a sutra from India was ipso facto the words of the Buddha. In the story about Xuánzàng receiving the Heart Sutra from a sick man, we are not told what language the sutra is in. But if we look at inscriptions from the period, they are almost all in Chinese, not Sanskrit. A few Sanskrit inscriptions exist, but only a handful of people could read them (a situation analogous to the present). It's unlikely that Sanskrit was heard outside the monasteries in which translations projects were carried out. 

It seems very likely that there was a conscious effort to promote the Heart Sutra from sutra extract (抄經) to sutra (經). And to focus on the name 《心經》 (Hṛdayasūtra) rather than any of the alternatives such as 《大明呪經》(Mahāvidyasūtra). The assigning of a translator (譯) would have been an essential part of this process, though it may have exploited an existing ambiguity in which Xuánzàng was an editor (譯) of the text. It is so tempting to see T251 as a edited version of T250, attributed to Kumārajīva, that we might not fault Tanahashi for referring to is as the α-version. Actually, we do not know the provenance of T250, though we do know that the evidence for it is later than evidence for T250. 


Questions, questions

In writing up my notes on the Fangshan Stele I was left with a number of questions:
  • Are the precise dates I have accurate? 
  • Are the 7th Century sources reliable? 
  • And how would I know?
  • Where can I find accurate geographical information on Tang China?
  • Do my observations about 譯 add up to anything?
  • What was Xuánzàng's involvement in the Xīnjīng?
I'm puzzled that many experts have transcribed the colophon of the Fangshan Stele without commenting on the words in it, especially the place names and military titles. Or is it just so obvious to experts that they didn't think it needed commenting on? When the experts in epigraphy don't do their job, then historians struggle to know what to make of such inscriptions. I'm also puzzled as to why so little has been made, by other historians, of the clear and dated reference to the Heart Sutra discussed in the Biography. If Xuánzàng gifted a copy of the Heart Sutra to the Emperor in 656 CE, then this really does change the narrative. 

The important point is that historians cannot afford to take witnesses at face value. Questions must be asked. Whether we seek to refute or confirm, we have to evaluate sources. Careful historiography is often our best defence against religious bias. History often reveals the weaknesses of religious stories precisely because it evaluates and compares sources. As a historian of ideas, I am fascinated by how doctrines that some religieux treat as articles of faith have been quite changeable over time. And, in particular, by how historical arguments about doctrine reveal weaknesses visible even in antiquity (without the need to invoke modernity or science). I hope to inspire friends, colleagues, and fellow religieux to be more careful in their use of historical sources, to cast a wide net, and above all to critically evaluate sources. 

~~oOo~~


Bibliography

Chinese texts from the Online CBETA Reader.

Beal, S. (1914.) The Life of Hiuen-Tsiang. By the Shamans Hwui Li and Yen-Tsung. London: Kegan Paul, Trench, Trubner & Co.

Li Rongxi (1995) A Biography of the Tripiṭaka Master of the Great Ci'en Monastery of the Great Tang Dynasty. Numata Center for Buddhist Translation and Research.