[5/7] 南傳佛教研究中的“經典”與“理性”問題

[Bilingual, Chinese & English] 

§5.

對於書面經典現狀誤解部分歸咎於研究一手源文本的學者們:我們未有簡明地讓民族志學者理解我們所指代的“南傳佛教經典”。在接下來的兩節中,我將在此一節(§5.)中描述來自(許多)存世經典之間關係的困惑,而後(在§6中)解釋困惑內在于我們對南傳佛教經典的理解之中。希望還能提出一些建設性的建議來減輕民族志學者和其他感興趣學者產生的困惑,即便本文未能詳盡備至地滿足巴利文專家的要求。

困惑的第一個原因僅僅出於偶然,即佛典零散瑣碎地呈現給現代學者;文本“再發現”(和再印刷)的順序較隨意。來自厄斯艾普(Urs App)的近期研究探討了隨著第一批(低品質翻譯的)經文到達歐洲並出版而招致的學術反應;[例如 App, 2010] 因為是在從事接收和評估這些文本的研究學者小圈子內得知,所以比同時期的歷史更富戲劇性。[例如 Yuyama, 2000]

當歐洲人第一次收到有關佛教的漢語文本時,沒有明確的指導原則來說明對於經典傳統何為明確重要同時又何為次要;漢語來源有其所聲稱的權威性,這是基於其(所認為)傳承於更古老的梵文原本,而歐洲人在甄別這些聲稱上沒有專門的優勢。作為一個啟示性的例子,艾普(App)指出“……幾個世紀以來,《四十二章經》一直被譽為佛教早期文本之一,而且作為首部梵文經文進入中國,並被譯成漢語,[但是]這部經書實際上是西元五世紀時來自中國的產物。” [App, 2010, 第10頁] 這一情況,中國人在許多世紀前就被蒙在鼓裡,而後當文本的傳播跨越大陸後,歐洲人又再次被完全愚弄了(該文本頗具影響力:出於巧合,成為被譯成歐洲語言的首部佛教文本)。中國是一個重要的經文來源,儘管它們的重要性源於它們同始于印度梵文來源的(脆弱)聯繫;西方學者沿用了這些漢語文本,同諸多他們幾乎不知如何質疑的文化假設“捆綁起來”。

還有一個例子,佛教詩人馬鳴菩薩(Aśvaghoṣa) 於1893年(由E.B. Cowell)發行了他的首部歐洲版作品,後因被薩繆爾比爾(Samuel Beal)從漢語譯成英語而頗具大眾影響力(作為馬克斯米勒(Max Müller)最暢銷的“東方聖書”系列推廣於1899-1900年間)。問題並不在於該文本從梵文譯為漢語,而後又由漢語譯成英語;而是西方人一直對先到歐洲,後因翻譯而流行的作品至於明確的重要位置這一問題上較為輕信。100年前馬鳴菩薩作品的暢銷(英譯版)實出於偶然,但這似乎足以扭曲人們對這部有著一千多年重要性之文本的假定(提出問題,再者,有關其當下重要意義等等)。歷史有多流行化較之純粹的歷史現實更時興。《四十二章經》與馬鳴菩薩詩作這兩者的譯作在現代都頗具影響性,卻缺乏這些文本是怎樣對經典的歷史發展產生關係的任何理解(或者,我們可以說,毋論他們與經典關係的脆弱性)。

總的來講,在19世紀和20世紀,對漢語經典佛教的研究比同期的南傳佛教學術研究要深入得多,但漢語典籍本身只含有對南傳佛教經典的呼應,隨著這些來源是如何彼此相互關聯的困惑也(自然而然地)接踵而至。接下來的觀點,我發現,距今已超過80年,但依然具有代表性,即(據我個人經驗)許多研究人員仍認為其正確:

“巴利文大藏經在漢語中並不代表一個整體。只有少數幾份翻譯還原回了巴利文原件。漢語三藏主要由梵文和普拉克利特語翻譯構成,但古代梵文經典本身並未在中文裡作為一個個體被發現,儘管似乎多以分散的形式存在。” [Clark, 1930, 第127頁]

中文研究者會對他們自己的經典及其古印度前著作的“分散”對應感到困惑,這並非不可思議。在漢語裡,所有這些經典都被不準確的稱作Sān Zàng (三藏,即,“The Tipiṭaka”);詞語Ā Hán (阿含,“Āgama”)仍然令人費解*4。

西方人沒曾有過多少優勢,部分原因是由於我們過多依賴中文學術作為媒介:漢語經典先于南傳佛教經典而被人們接觸到,但更重要地是在我們日後理解進展的每一步中,我們曾有許多能夠閱讀古漢語的學者(亞洲和歐洲都有)但讀懂巴利文的寥寥。古漢語現今在全世界已成為大學的主流學科,而南傳佛教經典的語言仍舊為學術研究的模糊地帶。儘管阿含文學中應用過的古漢語形式是研究處於特殊領域(也就是說,只有少數古漢語學者可以閱讀公認對等在巴利文大藏經中的那些文本),這一分支學科比南傳佛教經典研究成果豐富。如今在亞洲任何地方,甚至很少有人瞭解“巴利”一詞;在東南亞和中國國內,對何為巴利文與梵文之區別仍存普遍困惑。有幾次,中國教授們自信地(而錯誤地)為我講解,說將“巴利”作為“文”是不恰當的,因為他們認為巴利僅僅是口頭方言而無其自身的書寫習慣(且無文獻著作)。當然這比起歐洲的佛學教授對我所講的同樣話題來說並無可厚非;在諸多方面,不論東西方,我們對巴利文大藏經的介紹還處在“第一階段”。

引發進一步困惑的,是西方學術對已經歷過許多時尚與潮流的經典所持有的態度;其中一些是針對著新證據的出現,還有些則是關於研究方法的改變。在過去,曾有些錯誤的希望認為現代學者將從中國與印度來源的比較研究中重建一個 “純粹”的經典。這一愚昧的擔當被歐洲人認為是“……重新建立,盡可能深遠的,古代梵文經典”,並進一步希望對經典間的比較會揭示“先經典時期的佛教”脈絡。[Clark, 1930, 第138頁]

這一方法,或隱晦或公然地試圖認為(非現存的)梵文經典要先于南傳佛教經典;該假設是錯誤的。更糟的是,認為學者能夠在兩種文本間做出比較,來揭示出較前兩者更為古老的第三種文本的想法通常是黃粱一夢:在比較完整的南傳佛教經典和來自中亞的殘存經典中,就出土的殘存文獻而言,我們初步瞭解了關於中亞地區同一時代的語言與文化。換言之,當我們比較不同版本的經典時,我們並不深入往回去探究經典構成時的歷史,而是相反去向前邁進至其後期傳播時的歷史。通過“中亞”,我們指現在涵蓋阿富汗,土庫曼斯坦,以及大部分古代絲綢之路與中國相連的地區,但是相同的爭論也出現在其他例子中。

一些西方學者已經能夠“三方綜合”來自以下各方面的殘缺文獻,

  1. 非南傳佛教傳統(多以梵文,古印度帕拉克裡語,犍陀羅語等等記錄,還有許多殘存文獻以前伊斯蘭時期中亞地區的語言記錄,如和闐語,粟特語以及吐火羅語),還有,
  2. 巴利文大藏經,再有,
  3. “零散的”經典(上述)部分保存在漢譯版本中。

有一些關於這類非常詳細研究的深刻例子[例如,Stache-Rosen, 1968; de Jong, 1979] , 但我要告誡讀者的是,這些研究無法回答有關大多數人所感興趣的哲學和文化方面問題;我認為可以說來自這些研究的發現與當代民族志學無關(儘管我個人對它們很感興趣)。在de Jong對存于某些經典文本殘存文獻間細小變化的詳細探討中,或許民族志學者感興趣的唯一發現就是文本本身的結論,“……一定已經成為最為廣受青睞的佛教經部之一。” [de Jong, 1979, 第252頁] 這一發現並非微不足道*5,而是同通過比較研究來“重新建立”(如我所引自的克拉克(Clark),1930,見上)這一巨大的期望形成了對比。對於大部分,這些經典間的比較揭示了極其相似而(數位)順序不同的內容,並有大量證據顯示古代譯者努力的艱難。從這些差異的研究來看,我們(在發掘殘存文獻的地點和時間處)能瞭解到許多關於翻譯和傳播的文化,而鮮有關於佛教方面的內容。

最近在阿富汗發現了大量非南傳佛教經卷手稿(自1999年後只有學者瞭解)[Hartmann, 2004],能夠通過對來源的“三方綜合”開啟研究的嶄新時代。相反,我推測這也會令人對過去幾百年來堆積的非南傳佛教來源感到些許失望。發現的梵文所對應的巴利文經藏長部對許多西方學者是“一個實現了的夢想”,也同時應照了中文經部中的長阿含經。嗟夫,真實的現狀卻鮮能滿足對它預期的幻夢。與克拉克(Clark)[1930, 以上引用] 樂觀主義相悖的是,中亞殘存的文獻首先教給了我們關於在中亞傳播佛教的文化。其二,我們通過對比同一文本(均為同一文本)的兩種不同校本所收集的資訊類型僅僅證實了保存一個相同故事的兩種不同“策略”;我使用“策略”一詞是因為這些(依我個人觀點)被身為人類的作者們有意圖地運用著。比如有一個例子,在某種情況下南傳佛教經典將兩個故事保留為兩部獨立的經,一部接著另一部,但新近發現的梵文版本則把兩個故事合為了一篇文本*6;如果我們承認這些就是敘事者們採用的兩種策略,那麼區別就顯得平淡無奇了,而且不能為任何方式的重新構建提供材料。這阻礙了那個(愚昧的)以細微且無處不在的方式獲得更大真實性與權威性的訴求。再有一個例子,我確實發現有趣的是梵文版本中將《梵動經》同《十上經》和《眾集經》進行了調換,“……而且很難避免因其中章節和文本順序被故意轉換所造成的印象……”當對比了梵文和相應的巴利文之後;[Hartmann, 第 4–5頁] 然而,這種故意的變動只是說明了非南傳佛教學派在考慮到南傳佛教經典的時候,曾修改過其自己的經典,似乎如此一來衍生出了更多內容。因為梵文語料庫為巴利文的衍生,研究其中的區別便有些無足輕重了;的確,即使當梵文起源於一些其他先期典籍(早于巴利文但與之同源),我們仍沒有“權威性”和“真實性”的優勢(但相反,只是不同的一類謬誤)。其三,如果我們對兩種未保留相同故事的文本進行比較,我們便無論如何都沒了對比的基礎,且如此以來所有同樣(有關“重新構建”等等)的希望與主張將化為泡影。儘管對一個“新”故事的發現會受到普通讀者的歡迎,但如果我們發現的一部經典文本缺乏任何其他經典的相應佐證,則落入了文獻學中的死胡同。

在該類對各來源進行“三方整合”中所描述的同樣問題,限制了通過(如)Cheng Jianhua[未注明出版日期]主持的一個非常仔細的巴利文與漢語版本《梵動經》比較研究這種“雙面”法所獲得的發現。對佛教的漢語翻譯(及改寫)造成的歷史問題,我們既能以“前瞻性”的興趣來閱讀這種比較而獲得的發現,也可通過南傳佛教大藏經,以“回顧”的興趣簡單罷手(因為再無更早的證據可考)。若是回顧,我們最多可以期待的發現則是在非南傳佛教學派間的宗派差異,表現于文本的修訂上;這樣的證據比南傳佛教經典的產生更晚(對比“……我們必須承認該證據反映了一個充分發展的文化傳統,而不是存在於佛教構成之初。” [Wynne, 2004, 第120頁])。反之,這些發現對非專業者(包括民族志學者)不會有何興趣,因為這些非南傳佛教學派現已絕跡。

我們不能從傳播與翻譯後期產生的差異來推斷關於經典的首次構成情況。首次構成與後期傳播之間的差別涉及幾個世紀,以及,大多數情況下,數百公里的跨度。從文化角度講,這種差異描述了從作者身份態度到守護者身份態度的轉變。守護者的文化就是,在“詞以精准而存”的理念下,“採取的每一項措施,在當時的情況下,都(應)保證早期文學內容盡可能的固定與精確。[Wynne, 2004, 第122–3頁; 比較102–4頁]

在南傳佛教與非南傳佛教經典的對比中,我們僅僅研究了一些用於保存經典佛教的一些不同策略;該比較法只有當這些是相同先期典籍的衍生時(或當它們試圖講述相同故事時)才有效用,而不在於是否以梵文形式書寫。由於所有這些原因,比對經典不能揭示任何關於“先經典時期”佛教的內容。這既是基本現狀,且或許在同時“嶄新”與“古老”的證據搜集熱情之中所遺忘(正如新近發現於阿富汗的手稿)。

任何西方人曾經希望重新構建一個早于南傳佛教經典的經典如今都破滅了(儘管為這一愚昧訴求提供的現有資料比以往還要好)。從積極角度看,這也意味著對南傳佛教經典本身的研究比以往任何時候都更重要(也就是說,與§2和§3引用來源所對立的態度)。

近期的一份刊物報導了“重構先經典佛教的事業”依然在日本進行著,並且仍基於不同經典進行比較的方法,儘管這在西方已經過時很久。[Cho, 2002, 第435頁] 趨勢間的同樣區別似乎在日本人有關佛陀歷史日期辯論的貢獻方面也很明顯。[Prebish, 2008, 尤第12頁及以下等等] 在諸多信奉大乘佛教國家中,南傳佛教經典與漢語版本比較的主要影響之一,已重新引發了對以巴利文相應典籍所認為較古老與純粹篇章的興趣(而其他文本,如之前提及的《四十二章經》以明顯被現代學術所貶抑)。據報導,在現代韓國,“阿含”文本突然從漢文經典的困境中嶄露,並呈現在學術論文中,因為現代南傳佛教學術研究證明了其與歷史上的佛陀存在著聯繫(較之于後來起源的大乘文本,等等。)。[Cho, 2002, 第427頁]*7 沒有哪一佛教經典是依照編年體的順序安排的;隨著將最古老文本按歷史層次進行分離這一令人振奮的可能性,南傳佛教學術研究喚起了人們對大乘經典的興趣。這亦是無法持久的潮流(儘管或在日本,韓國等地時起時落)。

我認為,許多學者傾注于非南傳佛教材料的研究,希冀發現一些比南傳佛教經典“更古老”和“更純粹”的內容(即使他們為了一部僅存于現代想像之中的“元經典”而搜集著);最後,我懷疑,對於日本與西方學者,這樣的訴求,較之於所有在“重新構建”有關前經典的(失敗)嘗試,將不經意間證明南傳佛教經典本身的獨特文獻學重要性。在我所提到的所有事例中,文本間比較的艱巨工作是值得褒揚的,但其並不說明上述引自克拉克(Clark),1930,的該期望,且依我個人觀點,甚至也不能導出諸如歷史上佛陀日期之類的簡單問題。[參見 Prebish,2008]

在這一部分,我已努力勾畫出為何“經典”這一簡單概念始終在專家之中存在爭論的原因,以及為何它持續成為令非專家們沮喪和困惑的來源。對此情況的一種反應曾是一些學者對“經典”一詞的抵制,儘管其廢止的提議較為溫和且時而帶有詼諧之感。哈裡希(Hallisey)[1993]以術語“據稱經典的”和“據稱非經典的”,或許暗示我們應當克制因缺乏證據做出(關於何為經典)的判斷;同樣,柯林斯(Collins)[1990]寫了一篇文章,以略帶滑稽的標題,邀請我們面對問題,“……巴利文大藏經之特有理念”。這種正規和諧的缺乏可視為積極的一面;而某種程度上,或許反映了當新證據出現時(且舊假設不再時興)的一種懷疑精神。遺憾的是,許多針對同樣不諧的回應則是或完全避諱經典,或對他們研究的突出成就加以否定,或通過提出觀點認為經典某種程度上是西方學術爭議的產物(如在§2–3, 見上, 和 §7, 見下提到的)。

圖例翻譯(下邊三個解釋框,三段自上而下,順序依次為左,中,右)

此《律藏》文學歸類為“准經典的”
(1)仍排除所有晚期視為非經典(non-canonical)的著作評論,且
(2)認為現今在《波羅提木叉》(Patimokkha)與 《經分別》(Suttavibhanga)中發現的律戒衍生於一個共同的起源。現存形式的律戒為邊緣經典的(peri-canonical);其後來添加的層次例子(等等)作為准經典的(quasi-canonical)

文章中有談到,諸多這些文本所包含的(作為“經典的”)在古跡文獻中存有爭議;再有四部文本,見上右圓圈,只在緬甸被奉為經典。

還有一個巨大數量的(歷史晚期)非經典的(non-canonical)文學(故事,詩歌以及評述)英文稱為“Jātaka ”(《本生經》),引起困惑。


§5. Part of the blame for the confusion over the status of the written canon should be apportioned to scholars of the primary source texts: we have not made it easy for ethnographers to know what we mean by “the Theravāda canon”. In two sections, I would here (in §5) describe the confusion arising from the relationship between the extant canons (in the plural), then explain (in §6) the confusion that is internal to our understanding of the Theravāda canon. Hopefully, I can also offer some constructive suggestions to alleviate this confusion for ethnographers and other interested scholars, even if this essay cannot be detailed enough to suit Pali specialists.

The first cause of confusion was the mere happenstance whereby the Buddhist canons were revealed to modern scholars in bits and pieces; the order of the “rediscovery” (and re- printing) of the texts was arbitrary. Recent studies by Urs App discuss the intellectual reactions that accompanied the arrival and publication of the first (poorly translated) canonical texts in Europe; [e.g. App, 2010] this is much more dramatic than the history of the same period as told within the small circle of textual scholars whose job was to receive and evaluate these texts. [e.g. Yuyama, 2000]

When Europeans first received Chinese texts about Buddhism there were no clear guidelines as to what was of definitive importance and what was merely peripheral to the canonical tradition; the Chinese sources had their own claims to authority based on their (supposed) descent from more ancient Sanskrit (梵文) originals, and Europeans had no special advantage in scrutinizing these claims. As an instructive example, App points out that “…the Forty-Two Sections Sutra [四十二章經] had for many centuries been hailed as one of Buddhism’s earliest texts and as the first Sanskrit scripture to reach China and to be translated into Chinese, [but] this sutra is in reality a product of fifth-century China.” [App, 2010, p. 10] In this case, the Chinese were fooled many centuries ago, and then Europeans were fooled all over again when the text was transmitted across continents (and the text was influential: by happenstance, it was one of the first Buddhist texts to be translated into European languages). China was an important source of canonical texts, even if their importance derived from their (tenuous) connections to Sanskrit sources originating in India; Western scholars inherited these Chinese texts “bundled up” with many cultural assumptions they could hardly know how to question.

As another example, the Buddhist poet Aśvaghoṣa had his first European edition in 1893 (by E.B. Cowell) and then became massively influential due to a translation from Chinese into English by Samuel Beal (promoted as part of Max Müller’s popular “Sacred books of the East” series, starting in 1899–1900). The problem is not that this text went from Sanskrit into Chinese and was then translated again from Chinese into English; the problem is that Westerners have been credulous in assigning definitive importance to works that happened to arrive in Europe first, and then happened to become popular in translation. It is really just happenstance that Aśvaghoṣa became a best seller (in English translation) 100 years ago, but this seems to powerfully warp people’s assumptions about the text’s significance more than 1000 years ago (raising questions, moreover, about its significance here and now). Fashions in how history has been popularized are more powerful than bare historical facts; the translations of both the the Forty-Two Sections Sutra and Aśvaghoṣa’s poetry became influential in the modern era, without any understanding of how these texts related to the historical development of the canon (or, we could say, despite the weakness of their relationship to the canon).

In general, researches into Chinese canonical Buddhism were much more advanced than contemporaneous Theravāda scholarship in the 19th and 20th centuries, but the Chinese canon itself contains only an echo of the Theravāda canon, with confusion (naturally) ensuing as to how all these sources relate to each other. The following opinion, I note, is now more than 80 years out of date, but still (in my experience) representative of what many researchers presume to be true:

The Pali canon is not represented as a whole in Chinese. Only a few of the translations go back to Pali originals. The Chinese Tripitaka consists mainly of translations from Sanskrit and Prakrit, but the old Sanskrit canon itself is not found in Chinese as a unit, although most of it seems to be present in a scattered form. [Clark, 1930, p. 127]

It is not surprising (in any era, modern or ancient) that Chinese researchers would be confused about the “scattered” correspondence between their own canon and its antecedents in ancient India. In Chinese, all of these canons are imprecisely called Sān Zàng (三藏, i.e., “the Tipiṭaka”); the term Ā Hán (阿含, “Āgama”) is no less confusing. 4

Westerners have not had much of an advantage, partly because we relied so much on Chinese scholarship as an intermediary: the Chinese canon became accessible before the Theravāda canon, but, more importantly, at every stage of the development of our understanding thereafter, we have had many scholars who could read classical Chinese (in both Asia and Europe) but very few who could read Pali. Classical Chinese has now become a mainstream subject in universities around the world, while the language of the Theravāda canon has remained an obscure area of study. Although the form of archaic Chinese used in Ā Hán (阿含) literature is a special area of study unto itself (i.e., only a minority of Chinese classicists can read the texts that have putative analogues in the Pali canon) this sub-discipline is copous in contrast to Theravāda canonical research. Remarkably few people even know what the word “Pali” means in any given part of Asia today; both in Southeast Asia and within China there is still widespread confusion as to what the difference between Pali (巴利文) and Sanskrit (梵文) is supposed to be. Several times, Chinese professors have self-confidently (and incorrectly) instructed me that it was inappropriate to refer to Pali as a wén (文) because they assumed that Pali was merely a spoken dialect with no written tradition (and no literature) of its own. Of course, this is neither more absurd nor less absurd than what European professors of Buddhism say to me on the same subject; in many ways, we remain at “step 1" in introducing the Pali canon to both the East and the West.

Fomenting further confusion, Western academic attitudes toward the canon have gone through many fads and fashions; some of these have been responses to the emergence of new evidence, and some relate to changes in research methods. In the past, there had been some false hope that modern scholars were going to reconstruct a “pure” canon from the comparative study of Chinese and Indian sources. This benighted duty that Europeans assumed was “…reconstructing, as far as possible, the old Sanskrit canon”, with the further hope that the comparison of canons would reveal traces of “precanonical Buddhism”. [Clark, 1930, p. 138]

Either implicitly or explicitly, this approach tended to assume that a (non-extant) Sanskrit canon pre-dated the Theravāda canon; this assumption is false. Worse, the idea that scholars can compare two texts to reveal a third one that is more ancient than either of the first two is usually a delusion: in comparing the intact Theravāda canon to the fragments of canons from Central Asia, we primarily learn about the languages and cultures of Central Asia in the same era as the unearthed fragments concerned. In other words, when we compare different versions of the canon we do not probe any further backward into the history of the composition of the canon, but instead move forward into the history of its later dissemination. By “Central Asia”, we mean the area that now includes Afghanistan (阿富汗), Turkmenistan (土庫曼), and most of the old Silk Road (絲綢之路) linking this area to China, but the same argument could apply to other examples.

Some Western scholars have been able to “triangulate” fragments of,

  1. Non-Theravāda traditions (variously recorded in Sanskrit, Prakrit, Gandhārī, etc., with many fragments in languages of pre-Islamic Central Asia, such as Khotanese, Sogdian and Tokharian), with,
  2. The Pali canon, and,
  3. The “scattered” canons (as described above) partially preserved in Chinese translation.

There are some impressive examples of very detailed studies of this kind [e.g., Stache-Rosen, 1968; de Jong, 1979] but I would warn the reader that this research does not answer the type of philosophical and cultural questions that most people are interested in; I think it is fair to say that the findings arising from these studies are irrelevant to contemporary ethnography (although they are interesting to myself personally). In de Jong’s detailed discussion of minor variations between fragments of a certain canonical text, perhaps the sole observation of interest to ethnographers is the conclusion that the text itself “…must have been one of the most popular of Buddhist sūtras.” [de Jong, 1979, p. 252] This type of finding is not trivial,5 but it is a contrast to the grandiose expectations of “reconstruction” (such as I’ve quoted from Clark, 1930, above) through comparative study. For the most part, these inter-canonical comparisons reveal extremely similar contents arranged in different (numerical) sequences, along with a great deal of evidence of the difficulties that translators struggled with in the ancient world. From the study of these differences, we can learn a lot about the culture of translation and transmission (in the place and time of the excavated fragments), but very little about Buddhism.

The recent discovery of a large volume of non-Theravāda manuscripts in Afghanistan (only known to scholars from 1999 forward) [Hartmann, 2004] could have started a new era of such research through the “triangulation” of sources. Instead, I surmise that it will affirm the sense of disappointment with non-Theravāda sources that mounted during the last hundred years. The discovery of a Sanskrit analog to the Pali Dīgha Nikāya is “a dream come true” for many Western scholars, as both are analogous to the Cháng Ā Hán (長阿含) in the Chinese canon. Alas, palpable reality can rarely satisfy the fantasy that anticipated it. Contrary to the optimism of Clark [1930, quoted above], Central Asian fragments primarily teach us about the cultures that transmitted Buddhism in Central Asia. Secondarily, the type of information that we glean from comparing two different recensions of the same text (where they are the same text) only demonstrates two different strategies to preserve the same story; I use the term “strategy” because these are (in my opinion) intentionally employed by human authors. As an example, in one instance the Theravāda canon preserves two stories as two separate suttas (經), one after the other, whereas the newly-discovered Sanskrit version has the two stories combined as one text;6 if we accept that these are simply two strategies employed by storytellers, the difference becomes banal, and cannot provide the materials for any kind of reconstruction. This stymies the (benighted) quest for greater authenticity and authority in a subtle but pervasive way. For one further example, I do find it interesting that the Sanskrit version has transposed the Brahmajāla-sutta (梵動經) with the Dasuttara and Saṅgīti (十上經 & 眾集經), “…and it is difficult to avoid the impression that the order of sections and texts within them has been reversed intentionally…” in contrasting the Sanskrit to the Pali equivalents; [Hartmann, p. 4–5] however, this type of intentional change only demonstrates that the non-Theravāda schools were revising their own canon with an awareness of the Theravāda canon, seeming all the more derivative in so doing. Inasmuch as the Sanskrit corpus is derivative of the Pali, the study of the differences is banal; indeed, even when the Sanskrit derives from some other antecedent (prior to but sharing origins with the Pali) we still have no advantage in “authority” nor “authenticity” (but instead, just a separate set of errors). Thirdly, if we compare two texts that do not preserve the same story we have no basis for comparison whatsoever, and so all of the same hopes and pretensions (of “reconstruction”, etc.) collapse. Although the discovery of a “new” story might be welcome for the casual reader, if we discover an allegedly-canonical text without any parallel in the other canons, it is a philological dead-end.

The same type of problem described in this type of “triangulation” of sources limits findings that are possible through the “two sided” approach of (e.g.) a very careful comparative study of the Pali and Chinese versions of the Brahmajāla-sutta (梵動經) undertaken by Cheng Jianhua [n.d.]. We can either read the findings of such a comparison with a “forward looking” interest in the historical problems of the Chinese translation (and adaptation) of Buddhism, or else with a “backward looking” interest that simply stops with the Theravāda canon (because there is no earlier evidence to consider). In looking backward, the most we can hope to discover are the sectarian differences between the non-Theravāda schools showing up in revisions to the text; evidence of this kind is very much posterior to the creation of the Theravāda canon (cp. “…we must admit that this evidence reflects a well developed literary tradition, and not the conditions which existed at the beginning of Buddhist composition.” [Wynne, 2004, p. 120]). Conversely, these discoveries are of little interest to non-specialists (including ethnographers) because these non-Theravāda schools of Buddhism are now extinct.

We cannot infer anything about the canon’s first composition from differences that arose in later stages of transmission and translation. The difference between first composition and later transmission involves a space of centuries and, in most cases, hundreds of kilometres. Culturally, this difference describes a shift from the attitudes of authorship to the attitudes of conservatorship. The culture of the conservator was that, “Every measure [should be] taken to ensure that the early literature was as fixed and accurate as it could be under the circumstances…” with the ideal of “word for word accuracy” in transmission. [Wynne, 2004, p. 122–3; cf. 102–4]

In contrasting the Theravāda and non-Theravāda canons, we are merely studying slightly different strategies that were employed to preserve canonical Buddhism; the comparative method is only useful where these are derivative of the same antecedents (or wherein they attempt to tell the same story), regardless of whether they are written in Sanskrit or not. For all of these reasons, contrasting canons cannot reveal anything about “precanonical” Buddhism. These are fundamental facts that are, perhaps, forgotten in the enthusiasm for evidence that is simultaneously “new” and “ancient” (such as the newly discovered manuscripts from Afghanistan certainly are).

Any hope that Westerners formerly had of reconstructing a canon older than (and more authoritative than) the Theravāda canon is now over (even though the materials that are now available for this benighted quest are now much better than ever before). Viewed positively, this also means that the study of the Theravāda canon itself is more important than ever before (i.e., an attitude very much the opposite of the sources quoted in §2 and §3).

A recent publication reports that “the enterprise of reconstructing precanonical Buddhism” is still underway in Japan, and is still based on the method of comparing different canons, though this has been out-of-fashion for a long time in the West. [Cho, 2002, p. 435] The same disparity between the trends seems to be evident in the Japanese contribution to debates about the date of the historical Buddha. [Prebish, 2008, esp. p. 12 et seq.] In many Mahāyāna countries, one of the major effects of the comparison of the Theravāda canon to the Chinese versions has been a renewed interest in the passages that were identified as relatively ancient and authentic by their correspondence to the Pali (whereas other texts, like the aforementioned Forty-Two Sections Sutra, were effectively demoted by modern scholarship). Reportedly, in modern Korea the Āgama (阿含) texts suddenly emerged from the confusing morass of the Chinese canon to come to the fore of scholarly discourse because modern Theravāda scholarship demonstrated their connection to the historical Buddha (in contrast to Mahāyāna texts of later origin, etc.). [Cho, 2002, p. 427]7 None of the Buddhist canons were arranged in chronological order; Theravāda scholarship revived interest in the Mahāyāna canon with the exciting possibility of separating the most ancient texts into historical strata. This, also, is a fashion that cannot last long (though it may come and go at different times in Japan, Korea, etc.).

I surmise that many scholars plunged into the study of non-Theravāda materials hoping to discover something “more ancient” and “more authentic” than the Theravāda canon (even if they are searching for a “proto-canon” that only existed in the modern imagination); in the end, for Japanese and Western scholars alike, I suspect, this quest will inadvertently demonstrate the unique philological significance of the Theravāda canon itself, in contrast to all of the (failed) attempts at the “reconstruction” of something pre-canonical. The hard work of textual comparison is praiseworthy in all of the examples I’ve mentioned, but it does not answer the type of expectation quoted from Clark, 1930, above, nor, in my opinion, can it even broach simple questions such as the date of the historical Buddha. [cf. Prebish, 2008]

In this section I have tried to sketch out the reasons as to why the simple concept of “the canon” has remained contentious amongst specialists, and why it continues to be a source of frustration and confusion for non-specialists. One reaction to this situation has been the rejection of the word “canon” by some scholars, though it is an abolition proposed in a genial and sometimes jocular mood. Hallisey [1993] uses the terms “allegedly canonical” and “allegedly non-canonical”, perhaps implying that we should refrain from judgement (as to what the canon is) due to lack of evidence; similarly, Collins [1990] wrote an article with a slightly droll title, inviting us to question, “…the Very Idea of the Pali Canon”. This lack of consonance about canonicity can be seen in a positive light; for some, perhaps, it reflects a spirit of skepticism as new evidence becomes available (and as old assumptions go out of vogue). Unfortunately, many respond to the same discord by eschewing the canon entirely, or by denying its salience to their research, or by suggesting that the notion of the canon is somehow a fabrication of Western scholarly debate (as discussed in §2–3, above, and in §7, below).


4. 與一些期待相反,該詞語 (阿含) 不能認定一部文本為非南傳佛教的;同樣的詞 (Āgama) 還用在巴利文評論裡,等等。Contrary to some expectations, this term (阿含) does not identify a text as non-Theravāda; the same term (Āgama) is used in the Pali commentaries, etc.

5. 的確,對民族志學者來說,要知道佛教經典的盛裝服飾有何要素(在任何文化中,任何特定的歷史時期中)實際受到觀眾歡迎,而不用參考這類“生硬的”文本(及考古)研究是件非常困難(或許不可能)的事。Indeed, it is very difficult (or perhaps impossible) for ethnographers to know what elements of the Buddhist canonical panoply were actually popular with audiences (in any given culture, in any particular historical period) without reference to “hard” textual (and archaeological) research of this kind.

6. 這裡所涉及的兩部文本是Mahā-Parinibbāna-sutta (類似於《大般涅盤經》) ,然後是Mahā-Sudassana-sutta (即《長阿含經》(Dīgha Nikāya)中的 第16部和第17部文本); 依據Hartmann [2004, 第4頁] 在梵文校本中是將兩個故事合為一個連續的講述來討論的。The two texts alluded to here are the Mahā-Parinibbāna-sutta (analagous to the 大般涅盘经) followed by the Mahā-Sudassana-sutta (i.e., the 16th and 17th texts in the Dīgha Nikāya); according to Hartmann [2004, p. 4] the Sanskrit recension discussed presents both stories as one continuous narrative.

7. 我注意到,Cho本人是作為該傾向的一名反對者去寫作;他從許多韓國學者那裡引用,並描述著這一歷史反應,但他卻對關於佛教所導致的態度與假定持批評立場。I note that Cho himself is writing as an opponent of this tendency; he quotes from various Korean scholars and describes this historical reaction, but he is a critic of the resulting attitudes and assumptions about Buddhism.