Dan Arnold. Buddhists, Brahmins, and Belief: Epistemology in South Asian Religion. New York: Columbia University Press, 2005. 318 pp. $50.00 (cloth), ISBN 978-0-231-13280-0.
Reviewed by David Eckel (Boston University)
Published on H-Buddhism (March, 2007)
By now Dan Arnold's fine book on Indian epistemology will be familiar to many of the readers of the "Buddhist Scholars Information Network," especially to those who take delight in picking their way through what Richard P. Hayes calls "the briar patches of difficult Sanskrit and Tibetan texts." Arnold's book has been an instant success, and it is clear that any future work on Dign?ga, Candrak?rti, and their Hindu rivals will have to take its arguments into account. The book also has won the American Academy of Religion's award for the best "Constructive-Reflective Study in Religion" in the year 2006. This award places the book in very distinguished company, including Jeffrey Stout's Democracy and Tradition (2004) and Robert A. Orsi's Between Heaven and Earth: The Religious Worlds People Make and the Scholars who Study Them (2004). What ties these books together is a willingness to grapple with the normative dimension of religious life--not just to describe what religious people do, but to consider how they justify their beliefs, how they take their beliefs to be true, and how they construct the norms that govern their own lives and the lives of their communities. Dan Arnold's book gains its conceptual strength not just from the careful reading of Indian texts, but from considering the way the views presented in these texts could carry the force of truth.
To some people in Buddhist Studies this approach might seem strange, perhaps even dangerous. What right do we have to say whether Dign?ga's definition of perception is "true"? Is it not enough just to try to understand the basic elements of his argument? Arnold argues that the normative questions are intimately related to the problems of understanding. As anyone knows who has struggled with an unfamiliar argument in a Sanskrit text, it is not enough to identify the dictionary meanings of the words; you also have understand how the author could be persuaded that the words were true. For Arnold this understanding involves a process of rational reconstruction in which we attribute the "best possible arguments" to the author and assume that when the author's words make no sense, the failure belongs to us as readers rather to the text itself. This process of reconstruction is a version of the "principle of charity" that comes into play in any introductory course on Buddhism when students try to imagine why it might be compelling to say "All is suffering" or "Nothing has a self."
Arnold begins his investigation of Buddhist epistemology with an account of "Dign?ga's transformation of the Buddhist Abhidharma." The key point in this chapter has to do with the change of the concept of svalak?a?a from a "defining characteristic" in the Abhidharma to the "unique particular" that, for Dign?ga, functioned as the object of perception. Arnold characterizes Dign?ga's understanding of perception as a form of Buddhist "foundationalism," in which the perception of unique particulars functions as the source of knowledge and the final court of appeal for any claim about reality. Stated in this way, Dign?ga's view is subject to the critique directed by William Sellars at the "myth of the given": for the perception of a particular svalak?a?a to function as an instance of knowledge it needs to be connected in some way with concepts and words, but as soon as it becomes conceptual, it loses its direct, perceptual immediacy. Arnold draws out the implications of this problem by using the work of several other recent philosophers, most notably Gottlob Frege. The effect of this analysis is to show that Dign?ga's views, for all their difficulty, represent live intellectual possibilities. Dign?ga is not a strange artifact from another time, but a potential participant in a contemporary conversation.
Part 2 of the book explores the epistemology of Kum?rila Bha??a, a seventh-century M?m??saka. Kum?rila is known for the doctrine of the "intrinsic validity" (svata? pr?m??ya) of the Vedas, a view that B. K. Matilal dismissed as a form of fundamentalism and seems, to Buddhist eyes, to be a desperate attempt to defend the indefensible authority of scripture. In what can only be called an interpretive tour de force, Arnold shows us not only how this view might be intelligible, but how it might be persuasive. The interpretation turns on William Alston's "doxastic account of justification," in which Alston argues that religious experiences are justified in and of themselves. Alston's argument frees people from having to justify experiences with reference to other, more fundamental forms of knowledge (a process that Kum?rila pointed out would lead to an infinite regress). Whether Vedic injunctions count as self-evident experiences in the sense Alston describes is a serious question, but the M?m??sakas have clearly taken an important step away from Dign?ga's foundationalism, and Arnold has added an important dimension to the understanding of Dign?ga by showing the importance of his M?m??saka opponents.
In part 3 Arnold returns to an internal Buddhist controversy with a careful reading of the epistemological arguments in the first chapter of the Prasannapad?, Candrak?rti's commentary on N?g?rjuna's Madhyamakak?rik?s. Candrak?rti's chapter has been the focus of innumerable debates in Tibet and may be one of the most widely discussed passages in Indian Buddhist philosophy, but Arnold still finds room to develop a novel and challenging interpretation. He reads Candrak?rti's critique of the concept of svalak?a?a as a transcendental argument about the conditions of knowledge. In Arnold's words, transcendental arguments "cut short any appeal to experience by arguing that a condition of the possibility of any experience (any experience such as an empiricist might invoke to justify a belief) is precisely the state of affairs shown by the transcendental argument" (p. 125). In Madhyamaka terminology this state of affairs is "dependent origination" (prat?tyasamutp?da) or simply "emptiness" (??nyat?).
Arnold concludes his discussion of Candrak?rti by taking up the time-honored conundrum of the truth of Madhyamaka claims about emptiness. He asks whether it is really true that everything is empty. This question leads to another: if the Madhyamaka statement is true, how is its truth established? N?g?rjuna raised this question in the form of an objection at the beginning of the Vigrahavy?vartan?: if everything is empty, then N?g?rjuna's own words are empty, and they cannot prove anything. N?g?rjuna's response had two components. First he concedes that his words are empty and can have no real effect, but then, in the following verse and commentary, he argues that, even though his words are empty, they can bring about useful results, just as a cart can be used to carry a load of bricks or a pot can be used to carry water. The two parts of this argument correspond to the two parts of any Madhyamaka approach to language: ultimately the M?dhyamikas's words are no more real than anything else, but conventionally they have a useful effect. Arnold's question is whether they can be used to make a transcendental, metaphysical claim about the nature of reality.
He approaches this question first by disposing of a possible objection from Dign?ga to the effect that his account of perception functions as a form of conventional usage (like a cart carrying a load of bricks). Candrak?rti responds by arguing that conventional usage cannot function in that way: if it depended on that kind of foundation, it would literally be "senseless" (vyartha). Arnold moves on to explore a possible relationship between Candrak?rti's understanding of conventional truth and the ordinary language philosophy of J. L. Austin and P. F. Strawson. Then he considers what Jay Garfield has called "N?g?rjuna's Paradox": the claim that the essence of all things is their essencelessness. Attempts to resolve this paradox (including Paul Griffiths's theory of types) involve a recognition of different levels of discourse: claims that can be made coherently on one level cannot be made on another. Not surprisingly, this discussion leads Arnold into a final consideration of the two truths--the ultimate and the conventional--and what it means for Candrak?rti to "defer" to conventional truth, especially in relation to statements about "essences" and "selves."
It is difficult to do justice to the complexity and richness of Arnold's argument in such a short space, but it should be clear from this brief summary that he has given us a challenging new reading of Candrak?rti's Madhyamaka. To say that Candrak?rti develops a "transcendental" argument and makes "metaphysical" claims will seem strange to many readers. In what way could this interpretation be true? If the purpose of Arnold's rational reconstruction is to develop a better understanding of Candrak?rti and his tradition, the critical question is whether his interpretation opens up aspects of Madhyamaka that otherwise might be closed to us. One place to explore the implications of Arnold's argument might be Tibet. There is no end to the discussion of Madhyamaka rationality in the different traditions of Tibetan philosophy, as has been well demonstrated by Georges B. J. Dreyfus and Sara L. McClintock. But I think that the most helpful and promising sources lie closer to home, in the so-called Sv?tantrika sources of the Indian Madhyamaka. In a historical and cultural sense, Bh?viveka, ?r?gupta, Jñ?nagarbha, ??ntarak?ita, and Kamala??la were Candrak?rti's closest conversation partners. Do they confirm Arnold's reading of Candrak?rti? I think they do, in a striking way.
In his final chapter on Candrak?rti and again in his conclusion, Arnold recognizes that his point about Candrak?rti's transcendental arguments leads to an unexpected result: when Candrak?rti's argument is translated back into the language of the pram??as (which Arnold calls "reliable warrants"), Candrak?rti's final court of appeal is not perception (pratyak?a), as it was for Dign?ga, or verbal testimony (?abda), as it was for Kum?rila, but inference (anum?na). One way to reach this conclusion is simply by process of elimination. If Candrak?rti's argument generates knowledge, and this knowledge does not come from perception or verbal testimony, it must come from inference. For Buddhists, at least, there are no other practical options. But this conclusion also follows from the nature of Candrak?rti's argument, as Arnold characterizes it: "'Emptiness,' if it means simply the possibility and necessity of relationship, can be understood as a logical category as basic as the principle of noncontradiction" (p. 189). Emptiness has to do with "logically entailed consequences (_prasa?.
..ga)." In other words, it has to do with "reason" (yukti_).
Candrak?rti does not draw out this aspect of his argument as explicitly as he might, and for good reasons. There is strong suspicion in the Buddhist tradition about any attempt to use logical reasoning to gain access to ultimate reality. Candrak?rti quotes a s?tra that locates this suspicion in the life story of the Buddha himself: "Not long after his perfect awakening, the Blessed One thought: 'I have attained a Dharma that appears profound and is profound. It cannot be investigated (atarkya) and is not accessible to logical reasoning (atark?vac?ra). It is subtle and can be known only by the consciousness of a sage'" (Prasannapad? 498). In the Tarkajv?l? (Flame of Reason), Bh?viveka shows that the suspicion of logical reasoning played a key role in the debate between the Madhyamaka and Yog?c?ra. A Yog?c?ra critic objects to the Madhyamaka by saying: "Reality cannot be understood by inference, because it cannot be known by logical reasoning (verse 5.104: tattvasy?tarkagamyatv?t tadbodho n?num?nata?)." Bh?viveka responds by saying that "reality is not known as an object of inference, but inference rules out the opposite of the knowledge of reality" (verse 5.107: "ato 'num?navi?aya? na tattva? pratipadyate / tattvajñ?navipak?o yas tasya tena nir?kriy?"). Later in the same text, a M?m??saka compares M?dhyamikas to blind people who try to use inference to run along a dangerous road (verse 9.13-14: d??ayitv? tray?m?rga? hetubhir hetuv?dina? / anum?napradh?natv?t svanaya? dyotayanti ye // p?daspar??div?ndhy?n?? vi?ame pathi dh?vat?m / anum?napradh?n?n?? p?tas te??? na durlabha?) . Bh?viveka returns the favor by saying that the M?m??sakas travel the same road by "just following directions" (?abdam?tra). Unless they have inference to help them investigate and interpret their directions, it is easy for them to fall. For Bh?viveka, true philosophical vision does not come from the pram??as in general, but from the distinctive discriminative capacity of inference.
Bh?viveka's position about the priority of inference had important consequences for later Madhyamaka tradition, especially for the eighth-century philosopher Jñ?nagarbha. In a move that seems at first to be strange, Jñ?nagarbha defines ultimate truth as "correct reason" (rigs pa ji lta ba bzhin nyid ni don dam pa'i bden pa'o). (For this formula and the ones that follow, see Malcolm David Eckel .) Relative (sa?v?ti) truth "corresponds to seeing" (ji ltar snang ba / yath?dar?ana). While this formula is reminiscent of Candrak?rti's suggestion that "the mundane be just as it is seen" (laukikam ev?stu yath?d???am) (Arnold, p. 182), it is clear that Jñ?nagarbha has in mind a specific reversal of the priority of pram??as in Dign?ga and Dharmak?rti. Relative truth corresponds to perception, while ultimate truth corresponds to inference. In addition to this striking distinction between relative and ultimate truths, Jñ?nagarbha gives one of the most clear and concise accounts of the classic eighth-century definition of correct relative truth as "arising dependently," "capable of effective action" (arthakriy?samartha), and "satisfying only when not analyzed" (avic?ramanohara).
Much more could be said about the significance of these ideas in the tradition of Bh?viveka and Jñ?nagarbha, but this should be enough to indicate that Arnold's point about inference has strong precedent in Madhyamaka tradition. The Sv?tantrika sources would make it possible to state the point more clearly and draw out its implications more strongly. They also would allow Arnold to solve some of the lingering problems in his final chapter on Candrak?rti. He asks, for example, "how Candrak?rti can coherently claim always to defer to the conventional while, at the same time, refusing to countenance the one convention that is (particularly given the Buddhist diagnosis of our situation) arguably most central to our ordinary experience" (p. 202). This convention is the idea of a self. The concept of correct conventional truth permits just such a distinction: there is nothing wrong with using words like "self" conventionally (as many scriptural passages do), but this usage cannot be mistaken for the claim that it is possible to search for (or "analyze") and find a "self" in its own right. It also would be more fair to John Dunne's claim that a Buddha "does not see ordinary things in the world" (p. 204) to maintain a careful distinction between ultimate and conventional points of view. While it may be true (as Jñ?nagarbha says) that there ultimately is no difference between the ultimate and the conventional, it is misleading to collapse one into the other. The relationship between the ultimate and the conventional always has an element of paradox, but it is this paradox that makes Madhyamaka a "middle" way.
Arnold's conclusion returns to the normative questions of his introduction. He uses Jeffrey Stout's distinction between justification and truth to argue, once again, for the importance of taking the works of philosophers like Dign?ga, Kum?rila, and Candrak?rti seriously as attempts to determine truth. Otherwise we erase the disagreements between them and fail to grasp what motivates them as scholars. But the differences between their intellectual setting and ours make it possible for us to recognize that their beliefs are rationally held, while we also feel no need to be persuaded by their truth. Whether this line of argument avoids the pitfalls of relativism is unclear, at least to me, but it is clear that Arnold has given us a refreshing new look at some of the central philosophical disputes of the Indian tradition, and for that we should be very grateful.
. Louis de La Vallée Poussin, ed., M?lamadhyamakak?rik?s (M?dhyamikas?tras) de N?g?rjuna avec la Prasannapad? Commentaire de Candrak?rti (Bibliotheca Buddhica 4, St. Petersburg, 1903-1913; reprint, Osnabrück: Biblio Verlag, 1970).
. Georges Dreyfus and Sara L. McClintock, eds., _The Sv?tantrika-Pr?sa?.
..gika Distinction_ (Boston: Wisdom Press, 2003).
. Cf. Kawasaki Shinj?, ed., "The M?m??s? Chapter of Bhavya's Madhyamaka-h?daya-k?rik?--Sanskrit and Tibetan Texts--with the Sarvajña Chapter" (Tsukuba, Japan: Institute of Philosophy, University of Tsukuba, 1976, 1987, 1988). . Chr. Lindtner, ed., "Bhavya's Madhyamakah?daya (Pariccheda Five) Yog?c?ratattvavini?cay?vat?ra," Adyar Library Bulletin 59 (1995): 37-65.
. Malcolm David Eckel, Jñ?nagarbha's Commentary on the Distinction between the Two Truths (Albany: State University of New York Press, 1987).
If there is additional discussion of this review, you may access it through the list discussion logs at: http://h-net.msu.edu/cgi-bin/logbrowse.pl.
David Eckel. Review of Arnold, Dan, Buddhists, Brahmins, and Belief: Epistemology in South Asian Religion.
H-Buddhism, H-Net Reviews.
Copyright © 2007 by H-Net, all rights reserved. H-Net permits the redistribution and reprinting of this work for nonprofit, educational purposes, with full and accurate attribution to the author, web location, date of publication, originating list, and H-Net: Humanities & Social Sciences Online. For any other proposed use, contact the Reviews editorial staff at firstname.lastname@example.org.