[Bilingual, Chinese & English]
索斯沃德的論點並不完全抵制南傳佛教書面經典，但尤其抵制歐洲學術界所瞭解的書面經典。這既是他稱為的“涅槃佛教”（見上）並嗤之以，“在很大程度上主要是西方學者的產物”。[Southwold, 1987, 第448頁] 對於只通過歐洲譯本來接觸這些文本的歐洲人來說，摒棄這些翻譯就是摒棄了經典。同時索斯沃德承認這些文本經典早在歐洲人開始翻譯前就已經存在，“西方學者由於自詡其成就與權威性而曲解了經典”。[同上]
“以文本為中心開展[歐洲人在南傳佛教經典上的]研究所創立的佛教充滿著理性和人文氣息，同時借助西方學術研究機制取得成效，並以Gautama*2 歷史事實為中心，且毫不掩飾地異于佛教徒的實踐。正如T. W. Rhys Davids自己寫道，“巴利三藏佛教並不僅僅區別於迄今為止通常接受的佛教，而是與之勢不兩立。” [Snodgrass, 2007, 第198頁]
儘管我對歐洲譯者在經典上之偏頗和曲解的大致抱怨表示同情（我也曾在此議題上屢次發文），對此問題我的回答就是我對有著原始語言的古代一手來源進行研究（也就是說，我教會自己去讀懂巴利文）；這便克服了二手來源的偏頗，而且使我與其他人存在異見時能質疑既定假設。在21世紀，極少數的歐洲人能閱讀理解巴利文（更鮮有人能將巴利文，這一失傳語言的專業知識同任一現有傳統，如柬埔寨，緬甸，斯里蘭卡的，等等，的知識相結合）；對於不能閱讀巴利文的那些人來說，尋求具備這方面能力的人的建議與幫助，似乎是情理之中的讓步。不幸的是，索斯沃德那毫不妥協的研究方法（當下流行）為研究者提供了來忽視這些古代來源，或至少削弱這些來源的重要性的口實；索斯沃德有一相當明顯具削弱性的論斷，即書面經典的“涅槃佛教”乃“實際上的小眾傳統”。 [Southwold, 1987, 第 448頁]
將一種作為“小眾”的（以失傳語言承載的）書面經典同人類學家所面對的集體“大眾”化了的多數（以現有語言和當代文化承載的）地方傳統進行對比，具有誤導性；由於閱讀的人不多，且能讀懂的人更是寥寥，因此文本便不具備社會權威性；相反，當（較多）人員參閱並遵從這些經文時，其權威性即可彰顯。 我們所要考慮的問題應該是何時以及怎樣參閱經文（某些因文化差異而不同的東西，且對比於接下來時期的某一歷史時期。）小林（Kobayashi）在柬埔寨農村曾觀察到一個有趣的例子。[2005, 第495–501頁] 在一座寺廟裡，他報導稱，僧侶們在查閱佛典的偈頌後抵制了某一飼鬼的儀式；而這個儀式曾在其它研究人員調查過的當地寺廟中舉行過，且至今在整個柬埔寨依然被普遍接受。以下我引自小林從寺廟住持那裡報導的摒棄該儀式的理由：
“若寄功德予亡靈，米施僧人作媒因。在大藏經裡佛祖的聖諭中，我們不能找到關於 bân baybin [即這一飼鬼儀式的柬埔寨名稱] 的任何解釋。該儀式毫無意義， 因為功德須經佛教僧侶傳遞。而狗于地裡吃米便徒勞枉費了。” [Kobayashi, 2005, 第501頁]
傳統本身的力量存在於阻止這些問題產生的意識習慣之中；文化的研究在於探尋未解的問題集合。無論何時存有疑惑，文本優於實踐的重要性便在一次又一次這樣微小的“改進”中得以驗證：改進或會起于某位僧人的探詢，之後或止於某一寺廟內儀式的細小變化。同時，無人質疑寺廟牆壁上所描繪的那頭上長滿發的佛祖，但古代經文一致地將他描述成削髮剃度的形象（此例曾在我另外一篇，題為 The Buddha was Bald （直譯為《佛祖光著頭》）的文章中探討過）。這樣無可置疑的矛盾例子可以找到很多，而且這就是我們所謂的文化的實質 （似乎有別于信念或知識）。同時，“將殊榮賦予東南亞大陸的僧伽羅南傳佛教意味著當地傳統和文本會經常地對照巴利文大藏經來進行評價，而且一旦發現缺漏，或加以改進或刪節。” [Guthrie, 2004, 第 3–4頁]
甚至更糟（卻不易察覺）的問題是，忽視經典的研究者將不能對近期的革新與（純粹）的古代傳統做出區分。我在閱讀裘蒂 萊格伍德（Judy Ledgerwood）有關一種儀式（布薩）在如今每個月的同一天進行，且與一百年前的情況完全相同這一觀察報告時感到吃驚。事實上，（她所評述的）日期的重要性已經有兩千多年的歷史。除南傳佛典最古老的“核心”證據外，宗教曆法中的這一方面在古印度先于佛教的傳統中亦很明顯。[Nyanatusita, 2008, 第64頁] 萊格伍德針對（本應有）一百年的古老傳統提出她的看法，而沒有參考任何書面經典；若這也視為行之有效，則在她自己的觀察報告與來自一個世紀前另一西方人的觀察報告比較後，完全忽略了經典這一方面來看則至少不妥。這類觀察報告的出現（即缺少文獻學參與的民族志學）的確將關鍵任務置於未竟之地，而且會誤導眾多讀者。缺乏對古代文本的興趣造成的局限會令觀察過程中的研究人員，以及得出結論的讀者忽視重要方面的內容。
Southwold’s argument does not reject the written Theravāda canon entirely, but specifically rejects the written canon as it is known through European scholarship. This is what he calls “nibbanic Buddhism” (above) and he dismisses it as, “largely the product of mainly western scholars”. [Southwold, 1987, p. 448] For Europeans who only have access to these texts through European interpretations, dismissing the interpretation is tantamount to dismissing the canon. While Southwold admits that this written canon existed before Europeans started to translate it, “Western scholars distorted it by exaggerating its prominence and authority”. [Ibid.]
The problem is that Southwold continues to make comparative judgements about canonical sources in reliance upon the same tradition of European scholarship that he reviles as “distorting”. Southwold is not alone. These views seem to be especially fashionable amongst European scholars who study Theravāda Buddhism but who cannot read Pali themselves:
The Buddhism created by the text-centered study [of the Theravāda canon by Europeans] was rational, humanistic, validated by the apparatus of Western scholarship, and centered on the historical actuality of Gautama2 the man and was unabashedly different from Buddhist practice. As T.W. Rhys Davids himself wrote, “The Buddhism of the Pāli Pitakas is not only a quite different thing from Buddhism as hitherto commonly received, but antagonistic to it.” [Snodgrass, 2007, p. 198]
Although I sympathize with the general complaint that European interpreters have been biased and have misrepresented the canon (and I have published on this issue repeatedly), my response to this problem has been to study the ancient primary sources in their original language myself (i.e., I taught myself to read Pali, 巴利语); this overcomes the bias of secondary sources and enables me to challenge established assumptions when I differ from them. An extremely small number of Europeans have reading comprehension of Pali in the 21st century (and an even smaller number can combine this expertise in Pali as a dead language with knowledge of any one of the living traditions, such as Cambodia, Burma Sri Lanka, etc.); for those who cannot read Pali themselves, asking for advice and assistance from those who can might seem like a reasonable compromise. Unfortunately, Southwold’s uncompromising approach (now in vogue) provides an excuse for researchers to disregard these ancient sources or at least to demote their significance; there is a very significant degree of demotion in Southwold’s claim that the “nibbanic Buddhism” of the written canon is “actually a minority tradition”. [Southwold, 1987, p. 448]
It is misleading to contrast a written canon (in a dead language) as a “minority” to the plurality of local traditions (in living languages and contemporary cultures) encountered by anthropologists as a collective “majority”. Texts do not have social authority because of the number of people who read them, nor even because of the number of people able to read them; on the contrary, the authority of the scriptures extends through the (much larger) number of people who defer to them if and when they are consulted. The question we should consider is when and how the texts are consulted (something that differs from one culture to another, and in one historical period in contrast to the next). An interesting example was observed in rural Cambodia by Kobayashi. [2005, p. 495–501] At one temple, he reports, the monks rejected a particular ghost-feeding ritual after consulting the writ of the Buddhist canon; the ritual they decided to reject was conducted at other temples the researcher had surveyed in the same area, and is still generally accepted throughout Cambodia. I quote the justification for abolishing the ritual that Kobayashi reports from the head monk of the temple:
“If one wishes to transfer merit to the dead, rice should be offered to a monk as a source of merit. In Buddha’s sacred words in the Tripitaka, we could not find any explanations about bân baybin [i.e., the Cambodian name of this ghost-feeding ritual]. Such practice is really meaningless, because merit must be transferred through Buddhist monks. Dogs eating rice on the field can’t help anything.” [Kobayashi, 2005, p. 501]
The final statement refers to the real outcome of throwing rice through the air in the ritual alluded to: regardless of personal religious beliefs, when the ritual is finished, the rice that was thrown through the air falls to the ground, and is often eaten by stray dogs. Although it is anecdotal, this is a useful example of the interaction between textual authority and religious tradition that is ongoing in Theravāda cultures. Although there may be a small minority of people who are able to consult the ancient texts (and an even smaller minority may be motivated to do so) the written canon remains an open resource for anyone who would question or challenge Buddhism as it merely exists.
The power of tradition exists in the habits of mind that deter such questions from arising; the study of culture is research into the sum of questions that are never asked. Whenever such doubts should arise, the priority of the texts over practice is proven, again and again, in tiny “reformations” of this kind: the reform may begin with a single monk’s inquiry, and may end with a single temple’s minor change in rituals. Meanwhile, nobody questions the fact that the Buddha is depicted on the temple walls with a full head of hair, whereas the ancient scriptures uniformly describe him as shaven-bald (an example I’ve discussed in a separate essay, titled The Buddha was Bald). Innumerable examples of this type of unquestioned contradiction could be offered, and this is the substance of what we call culture (as something distinct from belief or knowledge). Meanwhile, “The pre–eminence given to Sinhalese Theravāda Buddhism in mainland Southeast Asia has meant that periodically, local traditions and texts are measured against the Pāli Canon, and if found lacking, reformed or suppressed.” [Guthrie, 2004, p. 3–4]
In this instance, I note, Kobayashi (the researcher quoted above) did not verify (and did not challenge) the monk’s claim that the ghost-feeding ritual lacked any canonical basis. If we want to move beyond merely “reporting as Buddhism what the Buddhists I knew taught me” we need to be able to interrogate the original texts ourselves, to present original contrasts between precept and practice. Participant observation (參與觀察) is insufficient if all of the participants are equally ignorant of the written canon; conversely, the example we’ve just examined shows that the dynamic relation between text and practice can extend from the most ancient written records down to the minutiae of monastic rituals as they are conducted today. If an informant (interviewed in fieldwork) does claim to have some canonical knowledge and the researcher is merely credulous in reporting their claims then the substance of the work still needs to be verified. In this sense, ethnography without philology leaves the task incomplete.
Even worse (but less readily visible) is the problem that a researcher who is unaware of the canon will be unaware of many questions that are worth asking, and will be unable to distinguish recent innovations from (genuinely) ancient traditions. I was astounded to read Judy Ledgerwood’s [2008, p. 154] observation of a certain ritual (the Uposatha) being performed on the same days of the month now as they were observed 100 years ago; in fact, the significance of the dates (that she comments on) were an established fact of history more than two thousand years ago. In addition to the evidence of the most ancient “core” of the Theravāda canon, this aspect of the religious calendar is also evident in pre-Buddhist traditions of ancient India. [Nyanatusita, 2008, p. 64] Ledgerwood offers her comments on this (supposed) hundred-year old tradition without any reference to the written canon; if that is valid, it is at least incongruous to see the canon totally omitted in the comparison of her own observation to another Westerner’s observations of a century past. The presentation of such observations (as “ethnography without philology”) does indeed leave a most important task incomplete, and would mislead many readers. A lack of interest in the ancient texts entails that important aspects go unnoticed both by the researcher in the course of their observations, and by the readers of their conclusions.
*2. 此為巴利部落名Gotama, 漢語稱為瞿曇。 在使用該詞語時， “Gautama the man”, 斯諾德格拉斯（Snodgrass）將注意力引導到歐洲學者對該部落名稱的偏好上，即視其為一個更方言化的（不太超自然的）方式來提及佛陀（Buddha）. This is the Pali clan-name Gotama, transcribed into Chinese as 瞿昙. In using the phrase, “Gautama the man”, Snodgrass draws attention to the European scholars’ preference for this clan-name, i.e., regarding it as a more vernacular (and less supernatural) way to refer to the Buddha.