Did Nagarjuna Really Refute All Philosophical Views?

·期刊原文

Did Nagarjuna Really Refute All Philosophical Views?

By Richard H. Robinson

Philosophy East & West

V. 22 No. 3 (July 1972) pp. 325-331

Copyright 1972 by University of Hawaii Press

p.325

The treatment Maadhyamika has received from its ancient opponents and modern discussants is, in one respect, as peculiar as the way in which it handled the views against which its criticisms were directed. Far from admitting that they must either refute the Maadhyamika objections or concede defeat, most classical systems either ignored Naagaarjuna's incisive and forceful attacks, or contented themselves with answering one or two specific objections, or tried rather ineptly to discredit the Maadhyamika method of refutation. Unsympathetic modern writers such as A. B. Keith have remarked briefly and categorically that this method is sophistic, but have not attempted to demonstrate their charge in detail. Sympathetic authorities such as T. R. V. Murti have shown clearly and concretely how this destructive dialectic works, but have not subjected to a searching examination its claim to demolish all constructive philosophical views (d.r.s.ti). If this claim is sound, it will be in order to examine why those philosophies to which it has been applied have not considered themselves refuted. Even if the Maadhyamika claim is false and its methods sophistic, philosophy stands to gain from discovering precisely how this form of sophistry works, and why it seems so formally convincing even while arousing suspicion in the observer.

In American country fairs there used to be a well-known game played with three walnut half-shells and one pea. The operator first held up all three shells for the audience to see. Then he turned all three upside down, placed the pea under one shell, and proceeded to shuffle the shells. When he stopped, a member of the audience would try to guess which shell the pea was under. Naagaarjuna's system resembles the shell game in several ways. Its elements are few and its operations are simple, though performed at lightning speed and with great dexterity. And the very fact that he cannot quite follow each move reinforces the observer's conviction that there is a trick somewhere. The objective of this article is to identify the trick and to determine on some points whether or not it is legitimate.

Naagaarjuna has a standard mechanism for refutation, the pattern of which may be abstracted as follows: You say that C relates A and B. A and B must be either completely identical or completely different. If they are completely identical, C cannot obtain, because it is transitive and requires two terms. If they are completely different c cannot obtain, because two things that are completely different can have no common ground and so cannot be related. Therefore it is false that c obtains between A and B.

Several features of this formula excite immediate suspicion. That it can be applied so readily to almost any thesis suggests affinity with a number of well-known sophistic tricks. That it relies on dichotomy calls for caution, since false dichotomies are so easy to: make and are so frequent in philosophizing. And that it seems to contradict common sense ought to arouse distrust, since even in philosophy common sense statements that seem true are not rejected until

p.326

disproved. But before we give way to offhand disbelief, it is proper to note the definitions and presuppositions that underlie this formula and render it more plausible.

The first key term is own-being (svabhaava), which is defined as unmade and not dependent on another (Maadhyamika-Kaarikaas 15.2) ; it is that of which otherwise-being (or change) never occurs (MK 15.8). It does not have another as its condition, it is calm, is not manifested discursively, is non-conceptual, and has no diversity (MK 18.9).

Emptiness (`suunyataa) is defined as equivalent to dependent coarising (MK 24.18). It equals absence of own-being. All entities (bhaava) are dependently coarisen (MK 24.19).

Own-being is defined as nondependent, so own-being cannot arise and so does not exist. It follows that all entities are empty and have no own-being: This is the sole thesis that Naagaarjuna wishes to prove. He asserts that when emptiness holds good all the Buddha's teachings hold good, and that when emptiness does not hold good, nothing is valid (MK 24.18).

The insistence in the refutation formula that A and B must be completely identical or different, rather than partly identical, follows from the definition of svabhaava as not dependent on another. Qualifications such as "some" and "partly" are excluded because the discussion is concerned not with the denial or affirmation of common-sense assertions such as "some fuel is burning and some is not," but with the concepts of own-being and essence. What pertains to part of an essence must pertain to the whole essence. A defining property is either essential or nonessential. If it is nonessential it is not really a defining property of an essence. If it is essential, then the essence can never be devoid of the property.

This set of definitions is clear and consistent. Wherever one of the terms can be applied, the others will follow and refutation occurs. This is not really mysterious, since svabhaava is by definition self-contradictory. If it exists, it must belong to an existent entity, that is, it must be conditioned, dependent on other entities, and possessed of causes. But by definition it is free from conditions, nondependent on others, and not caused. Therefore, it is absurd to maintain that a svabhaava exists.

The validity of Naagaarjuna's refutations hinges upon whether his opponents really upheld the existence of a svabhaava or svabhaava as he defines the term. Those who uphold the existence of a svabhaava are clearly self-contradictory, The possibility remains, however, that any one of the classical dar`sanas could' purge itself of the self-contradictory svabhaava concept and become as unobjectionable as the Buddha-vacana which Naagaarjuna accepts as legitimate. If, on the other hand, some non-Buddhist theories turn out not to be guilty of holding the svabhaava concept in the form Naagaarjuna defines, then his critique will find no mark, and his dialectic will fail to destroy some constructive philosophy.

p.327

A further possibility is that Naagaarjuna's critique, when applied, would fail to damage certain modern metaphysical systems that dispense with the category of essence. A good proportion of the axioms that he presupposes are not considered necessary or even acceptable by many European philosophers. Of course a neo-Maadhyamika might try to adapt the old technique to the presuppositions of the new philosophies, but the success of this attempt could not be assumed before hand. For the present, let us just examine the adequacy of Naagaarjuna's accommodation to the presuppositions of his classical antagonists.

We can concede immediately the soundness of Naagaarjuna's defense of his prasa^nga method (in the Vigrahavyaavartanii). Denial that an entity exists does not imply that its opposite exists. That the words and statements in the refutation are empty does not deprive them of validity. The debater who seeks only to expose self-contradictions in his opponent's arguments need not make any affirmative statements about what exists. He may assume his opponent's axioms for the sake of argument without committing himself to his opponent's truth. But if he uses an axiom or a rule that is not accepted by his opponent, then his argument is disqualified.

Does Naagaarjuna succeed in refuting all views without making any assumptions that are not conceded by the adherents of the particular view under attack? We can list some axioms upon which his arguments depend, and then go on to inquire how widely others accepted them.

(1) Whatever has extension is divisible, hence is composite, not permanent and not real.
(1a) Corollary: An indivisible, infinitesimal thing could have no extension.
(2) To exist means to be arisen; hence existence is synonymous with manifestation, and there is no unmanifested existence.
(3) A real thing would have to be an utterly simple individual which contains no diversity. If it had diversity, it would have extension and therefore would not be indivisible and real.
(4) The perception of arising and ceasing is illusory (MK 21.11, 17.31-33, 7.34).
(5) Only transitive actions and relations are admissible; reflexive actions are disallowed (the eye cannot see itself) (MK 3.2), the fingertip cannot touch itself, etc.), and seemingly intransitive expressions such as "the goer goes" must be recast in transitive form as "the goer goes a distance."
(6) The Buddha's teach truly that two truths or levels of being are to be distinguished, the mundane-conventional (lokasamv.rti) and the absolute (paramaartha) (MK 24.8).

Axiom 1 disagrees with the consensus of all schools, including the `suunyavaada

p.328

of the suutras, that aakaa`sa is ubiquitous and indivisible. Thus there is at least one entity that is not composite, has extension, and is permanent. Naagaarjuna's attempt to demolish the concept of aakaa`sa (MK chap. 8) selects the relation of aakaa`sa to its lak.sa.na as the vulnerable point. But as we will see later, his denial of the entity-attribute relation presupposes his denial of extension and is not admissible until after he has disproved the commonly accepted thesis that aakaa`sa is extended and indivisible.

If it is admitted that there is one extended, permanent and noncomposite entity, then it is not absurd to hold that there are others. And if extension is admitted, then duration must be admitted, too, since the arguments against duration involve the same operations of segmentation as those against extension. What the example of aakaa`sa does not render admissible, however, is diversity within an extended substance. Diversity was commonly considered to belong to the objects that occupied space and occurred in time.

Axiom 2 stands in contradiction to another axiom accepted by all schools and even invoked by Naagaarjuna himself: that the real is that which has never arisen (has no cause), and hence has no beginning or end, is permanent. One must choose between these two axioms. If reality means becoming, then that which is empty is real, that which is real is empty. But if reality means non-becoming, then either there is nothing real (a rather pointless inversion of the ordinary meaning of the word "real"), or there is a plane of being which is free from becoming. Naagaarjuna is not alone among the thinkers of classical India in promiscuously adhering now to one and later to another of these axioms.

Axiom 3 is simply not accepted by the pari.naamavaadins-I`svarak.r.s.na and Raamaanuja, for example. In classical Saa^mkhya, prak.rti has the three gu.nas, hence has diversity as an intrinsic property (SK 11). The gravest difficulty with this position is that there is no ready common-sense example of a variegated and nondivisible entity; apparent examples turn out to be composite. But if lack of a common-sense example militates against diversity-within-unity, Naagaarjuna's Axiom 4 is even more vulnerable on this score. His ostensible examples--a hallucination (maayaa), a dream, a mirage--are all instances that common sense can establish illusions. Naagaarjuna gives no common-sense perceptual criterion for considering all phenomena as maayaa. He does not examine perception empirically, does not attempt to show that the senses are bad witnesses in all cases, and merely asserts dogmatically that perception is marked by vikalpa (conceptualization), which falsifies. This is an empirical proposition, be it noted, and Naagaarjuna's claim is that he does not make empirical assertions apart from those conceded by his opponents. He ought to admit either that he maintains maayaavaada on the authority of the `suunyavaadin suutras, which are not aagama for the Hiinayaanist opponents or for the aastikas, or that he requires an empirically derived theory of error in order

p.329

to exclude empirical arguments from the rest of his dialectic. If he adduces the experience of the Buddhist contemplatives (which he does only obliquely in the Kaarikaas, but which the commentators do explicitly), he is not making an empirical assertion but making one that his opponents are not prepared to concede. Non-Buddhists are not going to accept that the Buddhist, aaryas are authorities, and other Buddhists do not take the experience of the aaryas as evidence for `suunyavaada. There appears to be no way of forcing a realist opponent to concede that common-sense experience is erroneous except to examine experience and to demonstrate empirically that it is delusive. Merely citing well-known varieties of illusion does not prove the point.

Axiom 5 seems sounder than it actually is because the examples are so graphic. It really presupposes Axiom: 1. If an entity possesses extension, then one region of it can act upon or relate to another region of it. This is the commonest sense of reflexive expressions. Naagaarjuna, incidentally, is guilty of a sheer quibble when he says that since the eye cannot see itself it cannot see another (MK 3.2). This is not seriously detrimental to his case, which can quite easily be restated without the quibble. But more serious is his failure either to accept or to disqualify the instances of genuine intransitive action that occur in common-sense experience plus the metaphysical ones that are affirmed by some of his opponents. When he denies that the lamp illuminates itself (MK 7.8), he is simply arbitrarily choosing to consider the reflexive object as if it were a non-reflexive object. He is refusing to allow, as ordinary language does, that reflexive statements are either pseudo-reflexive or pseudo-transitive. "I saw myself in the mirror" and "I scratched myself" are pseudo-reflexives. "Light illuminates itself" and "Water makes itself wet" are pseudo-transitive, better expressed by "Light is inherently bright" and "Water is inherently wet."

This brings us to Naagaarjuna's treatment of the relation between entity and attribute. It has already been mentioned that if extension were granted to entities then attributes might belong to them, but that since Naagaarjuna denies extension he denies the attributive relation. According to Axiom 3, a real thing must be utterly simple; it cannot have more than one property. But if it has even one property, then it consists of two entities--itself and its property--which is incompatible with Axiom 3. The weakness here is that none of Naagaarjuna's opponents really adhered to Axiom 3. The realists' infinitesimals were not utterly without extension and permitted each simple to have one attribute.

An allied but distinct argument against the entity-attribute relation charges that it involves vicious infinite regress. If A is the attribute, B is the entity, and C is the relation between them, then there must be another relation D which relates C to A, another relation E which relates C to B, other relations F and C which relate E to its relata, and so on ad infinitum. Naagaarjuna employs this

p.330

argument against the properties (lak.sa.na) of the momentary dharmas of the Aabhidharmikas-arising, abiding, and ceasing. He says that there would have to be an arising of arising, and an arising of the arising of arising, and so on infinitely (MK 7.3).[1] The Aabhidharmikas simply did not accept the charge. The Abhidharmah.rdaya (Chinese a-p'i-t'an-hsin-lun, Taisho 1550, 811b20ff.), reports the objection and replies that the operation is reflexive and reciprocal; each dharma effects itself and each of the seven others. For example, arising (a) arises of itself, and (b) constitutes the arising of abiding, disintegration, and ceasing and of the four anulak.sa.nas. This implies Aabhidharmikas acceptance of complex entities with multiple functions, and rejection of Naagaarjuna's axiom.

An important example of an intransitive function is the svaprakaa`sa of Advaita Vedaanta. This tenet has been skillfully and extensively defended against heterodox interpretations, and certainly suffices to show that Naagaarjuna's critique does not damage the Advaita Vedaantic position on this point. Thus it appears that the systems under attack commonly defended them- selves by affirming a reflexive, non-transitive or reciprocal operation which obviated vicious infinite regression. The weakness in the defense lies in its arbitrary selection of one, rather than other, place in which to posit non-transitivity.

Axiom 5 concerns the arbitrary introduction of pseudo-subjects and pseudo-objects. This operation is permitted and even encouraged by grammatical conventions in English and Sanskrit. The event which can be reported adequately with the single finite verb "rains" is more usually furnished with a dummy subject "it." Thus we may say "It rains" or "Rain rains," and adding a dummy object in reply to the dubious question "What does rain rain?" we may say "Rain rains rain." That the subject and object are dummies is shown by the fact that rain never does anything but rain, and nothing but rain is ever rained in the primary sense of the verb (excluding the other meaning, as in "Flowers rained down"). English "rain" is ambivalently transitive and intransitive, thus permitting the addition of dummy objects. The option of adding such objects, though, is not a sufficient warrant for transforming all statements to this form, and then treating the objects as if they were not dummies.

Axiom 6 is dogmatic and serves little purpose in the destructive dialectic. The distinction is not admitted by non-Buddhists, as Candrakiirti points out (MKV 27.1), and so the Maadhyamika is forced either to refute Saa^mkhya, for instance, from just one viewpoint, which is irrelevant to the opponent, or from both viewpoints, which is cumbrous for the proponent and irrelevant for the

________________________________________________________________________________
1. See Karl H.Potter,Presuppositions of India's Philosophies (Englewood Cliffs, N,J.: Prentice-Hall, 1963), no. 82-83, for a clear account of anavasthaa.

p.331

antagonist. This axiom is enunciated in Chapter 24 of the Kaarikaas where Naagaarjuna is explaining why he cannot be charged with denying all the Buddha's teachings. The opponent here is a Hiinayaanist who accepts the authority of the Aagamas or Nikaayas but certainly would not accept as Buddhavacana the `suunyavaadin suutras in which Naagaarjuna's doctrine of the two truths is stated, Hence Naagaarjuna departs from his avowed method, makes an existential statement about an exegetical principle, and bases his argument on an axiom not acceptable to his opponent.

The nature of the Madhyamika trick is now quite clear. It consists of (a) reading into the opponent's views a few terms which one defines for him in a self-contradictory way, and (b) insisting on a small set of axioms which are at variance with common sense and not accepted in their entirety by any known philosophy. It needs no insistence to emphasize that the application of such a critique does not demonstrate the inadequacy of reason and experience to provide intelligible answers to the usual philosophical questions.

This critique of Naagaarjuna's critique does demonstrate, however, that critical self-examination is fruitful for philosophy. A similar examination of the axioms and definitions of the other classical dar`sanas would reveal that each depends on a set of arbitrary axioms and hence does not arrive at any non-experiential propositions which all reasonable men must accept. More cogent than Naagaarjuna's criticism of constructive philosophy is that which T.R.V. Murti makes under Naagaarjuna's banner: "By its defective procedure dogmatic metaphysics wrongly understands the transcendent in terms of the empirical modes; it illegitimately extends, to the unconditioned, the categories of thought that are true within phenomena alone."[2]

I may add that dogmatic metaphysics, like the Maadhyamika critique, usually fails to do justice to the categories of thought we commonly employ in thinking about the phenomenal realm. This observable fact furnishes some justification for the Savage in Aldous Huxley's Brave New, World, who extracted from Hamlet's passing remark the definition: "A philosopher is someone who thinks of fewer things than there are in heaven and earth."

___________________________________________________________________________
2. T. R. V. Murti, The Central Philosophy of Buddhism (London: Allen and Unwin, 1955), p. 332.