The Paradox of Causality in Mādhyamika By David Loy
THE PROBLEM of causality is central to all schools of Buddhism, and this is especially true of Mādhyamika. But at first glance there seems to be a contradiction in the Mādhyamika analysis. On the one hand, causal interdependence is clearly a crucial concept, so important that Nāgārjuna identifies it with the most important concept, śunyatā: "We interpret the dependent arising of all things (pratītyasamutpāda) as the absence of being in them (śunyatā)."  This emphasis on interdependence develops to completion the early Buddhist doctrine of impermanence: there are no unconditioned elements of existence (dharmas), for all things arise and pass away according to conditions. The undeniable relativity of everything is the means by which self-existence (svabhāva) is refuted.
At the same time, Nāgārjuna redefines pratītyasamutpāda in such a way as to negate causality altogether. This is apparent even in the prefatory dedication of the Mūlamādhyamikakārikās, in the eight negations which Nāgārjuna attributes to the Buddha:
Neither perishing nor arising in time,
neither terminable nor eternal,
Neither self-identical nor variant in form,
neither coming nor going;
Such pratītyasamutpāda.. . .
Consistent with this, the first and most important chapter of the Kārikās concludes that the causal relation is inexplicable, and later chapters go further to claim that causation is like māyā. "Origination, existence, and destruction are of the nature of māyā, dreams, or a fairy castle."  The last chapters seize on this issue as one way to crystallize the difference between saṁsāra and nirvāna. The nirvāna chapter distinguishes between them by attributing causal relations only to saṁsāra: "That which, taken as causal or dependent, is the process of being born and passing on, is, taken non-causally and beyond all dependence, declared to be nirvāna."  In his commentary on the previous chapter, Candrakirti defines saṁvṛti and duḥkha in the same way: ". . .to be reciprocally dependent in existence, that is, for things to be based on
 Mūlamādhyamikakārikās (hereafter "MMK') XXIV 18, in Lucid Exposition of the Middle Way: The Essential Chapters from [he Prasannapada of andrakirti, trans. Mervyn Sprung (Boulder: Prajña Press, 1979), p. 238. This identification must be kept in mind to avoid Śaṅkara's error of misinterpreting sunyaia as non-being and making Mādhyamika into a nihilism.
 I bid., pp. 32-33, 35. For pratītyasamutpāda Sprung gives "the true way of things."
 MMK VII 34, as quoted in T. R. V. Murti, The Central Philosophy of Buddhism (London: Allen and Unwin, 1960), p. 177.
 MMK XXV 9, in Sprung, p. 255. In my opinion this is the most important verse in the Kārikās.
each other in utter reciprocity, is saṁvṛti." ". . it is precisely what arises in dependence that constitutes duḥkha, not what does not arise in dependence." 
How are we to understand this obvious contradiction? That is, how do we get from interpreting pratītyasamutpāda as "dependent origination" to "non-dependent non-origination," and, what is more, reconcile the two? Explanations which differentiate between two different types of sutras (exoteric neyārtha and esoteric nītārtha), or which refer to the two-truths theory, just raise the same question in a different form, for we need to know how such contradictory truths can both be true, i.e., how these two are related to each other.
This paper will offer one way (perhaps not the only way) of resolving this problem, which may more properly be said to be a paradox. I shall argue that complete conditionality is phenomenologically equivalent to a denial of all causal conditions. That is, a view which is so radical as to analyze things away into "their" conditions is offering an interpretation of experience which becomes indistinguishable from a view that negates causality altogether. If this is true, we have another instance where it becomes very difficult to distinguish the Mādhyamika nonduality from that of Śaṅkara's Advaita Vedānta.  The argument will be made in two steps. We shall see that there is a dialectic inherent in the Mādhyamika analysis. The first stage (discussed in Part I) is apparent: looking at the common-sense distinction between things and their cause-and-effect relationships, Nāgārjuna uses the latter to "dissolve" the former and deny that there are things. Less obvious is the second stage (Part II), which reverses the analysis: as we shall see, the lack of "thingness" in things implies a way of experiencing in which there is no awareness of cause and effect. Things and their causal relations stand and fall together, because our notion of cause-and-effect is dependent on that of things, which cause and are affected. If one collapses, so does the other. The basic problem is that any dualism between them is untenable. It is the delusive bifurcation between them that Nāgārjuna is concerned to negate, and the two ways to do this are to use each pole to "deconstruct" the other. Consistent with the general Mādhyamika project, this is criticism without affirming any philosophical position: having conflated the duality, Nāgārjuna does not offer any view about causality, because nothing remains to be related to anything else. Part III compares this conclusion with the Advaitic position regarding causality.
In order to understand the Mādhyamika critique, we must begin with a clear sense of what it is that is being criticized. This is our common-sense understanding of the world, which sees it as a collection of discrete entities (including myself) interacting causally "in" space and time. Just as space and time, if they are to function as "containers," require something understood as nonspatial and nontemporal to "contain," so the causal relation is normally used to explain the interaction between things which
 Sprung,pp. 230, 236.
 I have argued for this equivalence in two other papers. "Enlightenment in Buddhism and Advaita Vedanta: Are Nirvāna and Moksha the Same?" International Philosophical Quarterly, 22 (1982), 65-74, claims that the no-self modal-view of Buddhism is indistinguishable from the all-Self substance-view of Advaita. See my paper "The Mahayana Deconstruction of Time" (unpublished) which maintains that the nonduality between things and time amounts to making the same claim regarding temporality: if there is only time, this is phenemenologically equivalent to a nunc stans ("eternal now").
are distinct from each other. Nāgārjuna attacks more than the philosophical fancies of Indian metaphysicians, for there is a metaphysics inherent in our common-sense view. This common-sense understanding (one or the other aspect of which is absolutized in systematic metaphysics) is what makes the everyday world saṁsara for us, and it is this saṁsara that Nāgārjuna is concerned to "deconstruct." This is why one must beware of making Mādhyamika into an "ordinary language" philosophy by interpreting śunyatā merely as a "meta-system" term denying a correspondence theory of truth. By no means does the end of philosophical language-games "leave everything as it is" for Nāgārjuna, except in the sense that saṁsara has always really been nirvāṇa.
Why do we experience the world as saṁsara, if that is delusive? Why don't we experience it as it is? The traditional Buddhist answer points to craving and ignorance, but Nāgārjuna focuses on a particular type of mental attachment, that which makes all other attachment possible: prapañca. The nirvāṇa chapter of the Kārikās concludes by characterizing nirvāṇa, negatively, as "the coming to rest of all ways of taking things, the repose of named things (prapañcopaśama)The precise meaning of prapañca is, unfortunately, unclear. Sprung defines it as "the world of named things; the visible manifold."  It refers to some indeterminate "interface" between our concepts and our perceptions: that our categories of thinking are somehow responsible for our perceiving "the visible world" as "manifold." The consequence of prapañca is that I now perceive the room I am writing in, not as it really is, nondually, but as a collection of books and chair and pens and paper. . .and me, each of which is in some naive fashion taken to (1) be distinct from the others and (2) persist unchanged unless affected by something else.
This point about prapañca is important because without it one might conclude that Nāgārjuna's critique of self-existence (svabhāva) is a refutation of something that no one believes in anyway. But one does not escape his critique by defining entities in a more common-sense fashion as coming-into and passing-out-of existence. There is no tenable middle ground between self-existence independent of all conditions—an empty set, for there are no such entities—and the complete conditionality of śunyatā. The implication of Nāgārjuna's arguments against self-existence (e.g., MMK chapters I, XV) is to point out the inconsistency in our everyday way of "taking" the world: we accept that things change, and at the same time we assume that somehow they also remain the same—which is necessary if there are to be "things" at all. Other philosophers, recognizing this inconsistency, have tried to solve it by absolutizing one of these at the expense of the other; so the satkāryavāda substance-view of Samkhya emphasizes permanence at the price of not being able to account for change, and the asatkāryavāda modal-view of early Buddhism has the opposite problem of not being able to account for continuity. Nāgārjuna arranges these and the other solutions that have been proposed into a "tetralemma" which exhausts the possible philosophical alternatives; then he proceeds to refute them all. The basic difficulty is that any understanding of cause and effect which tries to relate two separate things together can be reduced to the contradiction of both asserting and denying identity. Nor can one respond to this simply by denying causality, for that is likewise contradicted by our experience. So Nāgārjuna concludes that the "relationship" between cause and effect is incomprehensible.
MMKXXV 24, in Sprung p. 264.
"Ibid., p. 273.
 The role of prapañca in our experience cannot be fully understood without relating it to our intentions. The relations among craving, conceptualizing, and causality have been discussed in "The Difference between saṁsara and Nirvāṇa," Philosophy East and West, 33 (1983), 355-365.
But the problem is not resolved simply by criticizing such positions, for the difficulty is fundamentally not abstract and philosophical but very personal: it is our lives, not just our theories, which are inconsistent in "taking" the world as a collection of discrete "self-existing" things which yet change. Although during, a philosophy seminar I may accept the complete conditionality and contingency of all experience, as soon as the seminar ends I unconsciously assume that the colleague I join for lunch is the same person whom I spoke with before the seminar—although ever-so-slightly different, due to a relatively extraneous change of mood, etc. This constitutes saṁsara because it is by reifying such "thingness" out of the flux of experience that we become attached to things. (Of course, other hypostatizations of self-existence— my wife, my car and, most of all, myself—tend to be more problematic loci for attachment, but the problem is the same in each case: we cling to things which dissolve as we try to grasp them.)
It does not suffice to answer this Humean critique of identity  with an "ordinary language" rejoinder that we should become more sensitive to the ways we use our permanence-and-change vocabulary, for the Mādhyamika position is that our usual everyday experience is deluded and this ordinary use of language is deluding. As the first prong of his attack, Nāgārjuna refutes our common-sense distinction between things and their causal relations simply by sharpening the distinction to absurdity: if things are to be "self-existent" then they must be separable from their conditions, but their existence is clearly contingent upon the conditions that bring them into being and eventually (when those conditions no longer operate) cause them to disappear. If it is objected that one cannot live without reifying such fictitious entities, at least to some extent, then the Mādhyamika response is to agree. The "lower truth", saṁvṛti, is not negated altogether—it is a truth—but it must not be taken as "the higher truth," as a correct understanding of the way things really are.
So the first stage of the Mādhyamika critique uses the complete interdependence of all things to refute their "thingness." The distinction between things and their causal relations is negated by adopting the latter as a means to "deconstruct" the former. This completes the early Buddhist attack on substance. But this absolutizing of conditions is only the first step. Now the critique dialectically reverses to use the perspective of the deconstructed thing in order to deny the reality of causal conditions. In the delusive bifurcation between things and their causal relations, the category of causality turns out to be just as dependent on things as things are on their causal conditions. Our concept of causality presupposes a set of discrete entities, whose interrelation we explain as causation. Cause-and-effect requires something to cause and something to be effected. If this is so, then a complete conditionality which is so radical that it "dissolves" all things must also dissolve itself.
In order to make this point, it is helpful to transpose the argument from the too-general category of causal conditions to the more specific one of motion-and-rest. Nāgārjuna analyzes motion and rest in chapter two of the Kārikās, immediately after his initial treatment of conditions, and it is clear that the second chapter is meant to apply the general conclusions of chapter one to a particular case. The additional advantage of shifting to motion-and-rest is that we illuminate what is otherwise a very puzzling chapter. The basic problem is that it is not clear what Nāgārjuna is actually doing in chapter two. Like Zeno, he denies the reality of motion, but this is not done
 See Hume's Enquiry Concerning Human Understanding, Section IV, Parts I and II.
to assert an unchanging Parmenidean "block-universe," for Vedantic permanence (i.e., "rest") is also denied. As a result, Nāgārjuna has been criticized for an arid play on words which "resembles the shell game" in being a logical sleight-of-hand" — that is, for basing his argument on a subtle distinction between words which are meaningless because they have no empirical referent—and for committing the fallacy of composition in arguing that what is true for the parts (in this case, traversed, traversing, and to-be-traversed) must be true for the whole.  But I think such criticisms miss the point of Nāgārjuna's arguments. Their import is that our usual way of understanding motion, which distinguishes the "mover" from "the act of moving," simply does not make sense, because the interdependence of "mover" and "motion" shows that the hypostatization of either is delusive. (Here "mover" means, not "that which causes motion", but "that thing which moves"; and "motion" is "the movement that happens to the thing") Nāgārjuna's logic in stanzas two through eleven demonstrates that once we have reified a distinction between them, it becomes impossible to relate them back together again—a quandary familiar to students of the Western mind-body problem, which is the result of another reified bifurcation. The difficulty is shown by isolating this hypostatized mover and inquiring into its status. Nāgārjuna asks: In itself, is it a mover, or not? That is, is the predicate "moves" intrinsic to this mover, or contingent? The dilemma is that neither way of understanding the situation is satisfactory. If the "mover" in and of itself already moves, then there is no need to add an "act of motion" later; the predication of such a "second motion" would be redundant. But the other alternative—that the "mover" by itself is a non-mover—does not work either because we cannot thereafter add the predicate: it is a contradiction for a non-mover to move. In neither way can we make sense out of the relation between them. It follows that the "mover" cannot have an existence of its own apart from the "moving," which means that our usual dualistic way of understanding motion is untenable. To summarize this in contemporary terms, Nāgārjuna is pointing out a flaw in the everyday language we use to describe (and hence our ways of thinking about) motion: our ascription of motion predicates to substantive objects is unintelligible.
At first encounter the above argument may be unconvincing. The options seem so extreme that we suspect there must be some middle ground between them. Of course we can't accept a double movement, but is it really a contradiction for a non-mover to move? What else could move? But, again, no such appeal to our everyday intuitions, or to the ordinary language which shapes and embodies them, is successful against the Mādhyamika critique of those intuitions, which seizes on the inconsistency that is ignored (and to some extent must be ignored) in daily life. We can elaborate upon this by applying the logic that was used earlier to deconstruct the difference between things and their causal relations. Just as (general rule) complete interdependence dissolves the thing into its rational conditions, with no residue of substance remaining, so (specific case) the completeness of movement—the fact that no part of "me" stays unmoved in the chair when "I" go to lunch—means that no self-existing and hence unchanging "thing" remains to move. As with time and with space, we think of the relation between mover and moving with the metaphor of container and contained, and in all three instances the bifurcation is delusive. In order to expose the absurdity of this, Nāgārjuna needs only to sharpen the dichotomy. Despite our intuitions, which
 "Richard Robinson, "Did Nāgārjuna Really Refute All Philosophical Views?" Philosophy East and West, 22 (1972), 325.
 Hsueh-li Cheng, "Motion and Rest in the Middle Treatise," J. of Chinese Philosophy, 1 (1980) 235 ff.
inconsistently want to postulate some "unchanging core" in order to save the mover, there is no middle ground between a self-existent and hence unmoving thing, and the complete dissolution of the thing that does the moving. Understood in this way, it becomes obvious why his arguments also work just as well against the intelligibility of rest: the bifurcation between the thing and its being at rest is just as delusive, and for precisely the same reason.
So far, we have effected only the first stage of the dialectic, both in the general analysis of causal conditions and this more specific case of motion-and-rest. We have dissolved the thing which moves / is caused and are left with a changing world of causal conditions. The second stage of the dialectic is easy to state but harder to understand: granted, if there is only movement, then there is no mover; but if there is nothing to move, then likewise there can be no movement. Implicit in our concept of change is the notion that a thing is becoming other than it was, so unless one reifies something self-existent (cf. atemporal) in order to provide continuity between these different conditioned (temporal) states, there is nothing outside the changing conditions to be changed. The concept of change (here a general term to include both conditionality and movement) needs something to "bite" on, but the first stage of the dialectic leaves nothing unconditioned to chew. If the colleague I join for lunch is not in any sense the same person I spoke with earlier, because there is no substratum of permanence to "him," then it makes no sense to say that "he" has changed. Without a contained, there can be no container; as the bifurcation dissolves, the poles conflate into a whole which, as Nāgārjuna knew, cannot be represented but remains philosophically indeterminate, since language, in order to describe at all, must dualize between subject and predicate, mover and moved, cause and effect.
But nonetheless, unless we can get a sense of what such a way of experiencing would be like, the above argument will be at best philosophically persuasive yet will seem irrelevant to daily life. What consequences does all this have for the way we actually experience the world? In particular, it is still unclear how, except by some "logical sleight of hand," all-conditionality can be phenomenologically identified with no-conditionality, as we claimed at the beginning of this paper.
Let me try to satisfy these questions with the help of a well-known Ch'an (Zen) story. The following example discusses the causal relations of a physical action, but what is said may be applied just as well to sense-perception (e.g., to the nondual sound of a pebble striking bamboo, which awakened Hsiang-yen, or to the nondual pain when Yün-mên broke his ankle) and to thought (Hui Neng: "If we allow our thoughts, past, present and future, to link up in a series, we put ourselves under restraint. On the other hand, if we never let our mind attach to anything, we shall gain liberation.").
Lin-chi was a monk in the monastery of Huang-po. Three times Lin-chi asked the Master: "What is the real meaning of Bodhidharma coming from the West?" and each time Huang-po immediately struck him. Thereupon, discouraged, he decided to leave
The direct relevance of Ch'an experience to this issue cannot be questioned. While it is true that Ch'an is not. and does not have, any philosophy, yet it is also the case that Mādhyamika, as a philosophical exposition of the Prajñāparamīta, may be said to be the philosophy that most made Ch'an possible.
and was advised to go to Master Ta-yü. Arriving at his monastery, Lin-chi told Ta-yü of his encounters with Huang-po, adding that he didn't know where he was at fault.
Master Ta-yü exclaimed; "Your master treated you entirely with grandmotherly kindness, and yet you say you don't know your fault." Hearing this, Lin-chi was suddenly awakened and said; "After all, there isn't much in Huang-po's Buddhism!" 
What did Lin-chi realize that awakened him? If (rushing in where Ch'an masters will not tread) we distort his experience into an idea in order to gloss this story, we may say that Lin-chi must have realized that Huang-po had been answering his question. The blows he received were not punishment but a demonstration of why Bodhidharma came from the West. On the common-sense level, the answer to Lin-chi's question is obvious: Bodhidharma was bringing Buddhism to China. But this is a relative, "lower truth" explanation. As a standard Ch'an question designed to initiate a dialogue, it goes without saying that what is sought is the "higher truth," and on this level there is no "why." For the enlightened person, each experience is complete in itself, the only thing in the whole universe; for each action is tathata. Nothing changes because without prapañca-reification everything is perceived afresh, for the first time. As Bodhidharma walked from India there was no thought of "why" in his head; "he" was each step. In the same way, there was no "why" to Huang-po's blows; "he" was that spontaneous, unselfconscious action. Lin-chi's sudden realization of this overflowed into his exclamation: "So, there isn't much to Buddhism after all!" (Only "just 'this'!") Upon returning to Huang-po, he revealed the depth of his understanding—more than just an intellectual insight—by not hesitating to give Huang-po a dose of his own medicine.
The paradox which makes the above story relevant to this paper is the fact that, at the same time, Bodhidharma's and Huang-po's actions are intentional. Huang-po's blow may be immediate and spontaneous, but there is also a "reason" for it; it is not a random or irrelevant gesture, but a very appropriate response to that particular question, drawn forth by that situation. If we translate this point about intention back into our more general category of causality, here we have a case of an act which is both completely caused (perfect upāya: glove fitting hand tightly, to use the Ch'an analogy) and yet is also uncaused. This paradox is a contradiction only according to our usual understanding of causality, which uses that category-of-thought to relate together the supposedly discrete objects into which prapañca carves the world. The first and most important of these hypostatized "things" is me, the subject who craves some of these objects and thus needs an understanding of cause-and-effect relationships in order to manipulate circumstances and obtain them. (It has been argued that the desire for such manipulation is the very root of our concept of causality.  ) This would apply to the story in question if Huang-po, prapañca-deluded, were to perceive Lin-chi dualistically: Lin-chi is sitting there, a person-object that needs to be enlightened, and I,
 This version of the story, from the Transmission of the Lamp, is given in Chang Chung-yuan, Original Teachings of Ch'an Buddhism (New York; Vintage, 1971), pp. 116-17.
 The idea of cause has its roots in purposive activity and is employed in the first instance when we are concerned to produce or to prevent something. To discover the cause of something is to discover what has to be attested by our activity in order to produce or to prevent that thing; but once the word 'cause' comes to be applied to natural events, the notion of altering the course of events lends to be dropped. 'Cause' is then used in a nonpractical, purely diagnostic way in cases where we have no interest in altering events or power to alter them." (P. H. Nowell-Smith, "Causality or Causation." I have a cyclostyled copy of this article but have not been able to trace its source.)
Huang-po sitting here, am the person who will try to enlighten him. Then "my" blow is reified into a deliberated effect which I hope will cause Lin-chi's awakening.
But if, as all schools of Buddhism agree, there is no self to do this causal-relating-between-things, then the above understanding of the situation must be delusive. So Huang-po must have experienced it differently, and causality too must be understood differently. It is not denied: on the contrary, without the sense of self and other prapañca- reified objects as a counterfoil, it expands to include everything, as Nagar-juna has already shown. (So the doctrine of karma can be understood simply by applying something like Newton's third law of physical motion to the mental realm as well.) From the perspective of Mādhyamika's "all-conditionality" which deconstructs all things, Huang-po's blow is part of a seamless web of conditions which can be extended, as Hwa Yen does, to encompass the entire universe. As one jewel in the infinite web of lndra, the blow reflects everything everywhere, at all times. But if every event that happens is interdependent with everything else in the whole universe, what a different way of experiencing this involves! It suggests a Spinozistic acceptance of whatever happens, as a product of the whole, but more than this it implies the irrelevance of causality as usually understood. "All-conditionality," in its complete negation of anything to be attached to, offers no practical utility, because there is no longer any object to be obtained or any self that craves it; whereas a self that wants to obtain some thing will need to isolate some discrete object or action as the cause which leads to obtaining it.
What does all this imply about the way Huang-po experienced his own action? Because he did not perceive the situation dualistically, the action was not "his." That the blow was appropriate to the situation was not due to any prior deliberation, however quick. On the contrary, the action was so appropriate precisely because it was not deliberated, just as the best responses in "dharma-combat" are unmediated by any self-conscious "hindrance in the mind." Then why did Huang-po strike, rather than shout "ho!" as Ma-tsu often did, or utter a few soft words, as Chao-chou probably would have done? This is the crucial point: He does not know and cannot know. ("Not-knowing is very profound" said Master Lo-han, precipitating Wên-i's awakening.) His spontaneous actions are traceless, "like the tracks of a bird in the sky."  They respond to a situation like a glove fits on a hand because whatever "decisions are made" (if this phrase can be used here) are not made by him. If one nondualistically is the cause (or effect, or both), rather than being a hypostatized self that dualistically uses it, then there is not the awareness that it is a cause (or effect, or both); it is experienced as whole, complete, and "traceless." In this way there turn out to be only two alternatives: either cause-and-effect relationships between discrete prapañca-objects, manipulated by the prapañca-subject, or nondual "all-conditionality" which amounts to an experience of complete unconditioned freedom. Without the interference that the self creates, Indra's all-encompassing web of causal conditions is indeed seamless. In psychological terms, the barrier between consciousness ("ego") and subconsciousness dissolves ("the bottom falls out of the bucket") and thereafter thoughts and actions are experienced as welling-up nondually from a source
 Tung-shan told his students to walk "in the bird's track," which is of course trackless, having no deliberative traces before ("Should I do this or that?") and leaving none after ("Should I have done that?"). For further discussion of the relations among action, intention, and nonduality, see "Wei-wu-wei: Nondual Action," Philosophy East and West, 35 (Jan. 1985).
unfathomably deep—or (what amounts to the same thing) from nowhere. In this sense Mahayana is not wrong to identify śunyatā with the Absolute. 
At first glance, the Advaitic account of causality is very different from the Mādhyamika conclusions and Ch'an experience which have been discussed above. Śaṅkara's position regarding causality constitutes part of his more general māyā doctrine, according to which all phenomena are the indescribable and indefinable ajñana which is superimposed (adhyāsa) upon Brahman. But if we delve beneath the surface of terminology and ask what experience this describes, it becomes difficult to find any phenomenological basis for the distinction between the Mādhyamika and Advaitic accounts.
Perhaps this similarity should not be surprising, since Śaṅkara's dialectic was clearly influenced by Nāgārjuna's. In this regard, it is relevant to compare Śaṅkara's careful critique of Vijñānavāda with his cursory dismissal of Śunyavāda as nihilistic and unworthy of repudiation.  Śaṅkara's main treatment of causality, in Brahma-Sūtra-Bhāsya II.i.14-20, is indebted to the Mādhyamika dialectic and reaches a similar conclusion, that we cannot derive the real nature of causal relations from the series of discrete cause-and-effect phenomena. As a Vedantin, Śaṅkara then leaps to the conclusion that the true cause of all effects must be Brahman, which provides the permanent substratum that persists unchanged through all experience. All effect-phenomena are merely illusory name-and-form superimpositions upon Brahman, the substance-ground. Since Brahman is the only real (svabhāva), and phenomena existing as distinct from it are illusory, this is a version of satkāryavāda: the effect pre-exists in the cause. But to distinguish this view from that of Sāṁkhya, which identifies cause and effect by granting the reality of prakṛti, Śaṅkara's theory of causality is more precisely labelled vivartavāda, since the effect (māyā) has a different kind of being from the cause (Brahman). 
Expressed in this way, the views of Nāgārjuna and Śaṅkara seem diametrically op-
 This view of Mādhyamika is important for understanding the trisvabhāva doctrine of Yogācāra Buddhism. The prapañca world of discrete forms corresponds to parikalpita-svabhāva, the "imagined nature." "All-conditionality" corresponds to parinispanna-svabhāva, the "other-dependent nature." "No-conditionality" corresponds to parinispanna-svabhāva, the absolutely-accomplished nondual nature. Read in this way, Vasubandhu's Trisvabhāvanirdeśa, for example, is completely consistent with the Mādhyamika analysis of experience. For both Mādhyamika and Yogācāra, an understanding of "all-conditionality," with its negation of the self-existence of discrete things, is the crucial "hinge" by which we turn from avidyā to prajña.
I think that the Mādhyamika view of causality—a dialectic which equates complete conditionality with no-conditionality—also implies a critique of Derrida's deconstruction. Derrida's use of the open-endedness (différance) of texts to deconstruct the self-as-writer employs only the First movement of the dialectic; the second and reverse movement (which Derrida does not make) uses the lack of a self to deconstruct the dissemination of meaning. One ends up with something more like the presence of the late Heidegger, where language is realized to be "the house of Being." The same point can be made by comparing the Mādhyamika critique of temporal relations with Derrida's critique of "logocentrism."
 It is not unlikely that Śankara discovered his own non-dual philosophy in the system of Nāgārjuna and left it unexplained. His debt to Śunyata doctrine was so great that he quietly passed over it." Lal Mani Joshi, Studies in the Buddhistic Culture of India (Delhi: Motilal Banarsidass, 1977), p. 234.
 Compare William Blake: "And every Natural Effect has a Spiritual Cause, and Not a Natural; for a Natural Cause only seems: it is a delusion. .. ." (Milton, plate 28, 44-45)
posed. They draw opposite conclusions from the illogicality and unintelligibility of the causal relationship: Nāgārjuna, as a Buddhist, denies any substratum-ground to phenomena, and leaves them empty (śunya) of any Being; Śaṅkara, as a Vedantin, postulates Brahman as an imperceptible but necessary substratum. Phenomenologically, however, these positions turn out to be equivalent. For Advaita, Brahman is the real cause of all phenomena, but, in denying the reality of all changing attributes, Śaṅkara is reduced to defining the substratum so narrowly that it ceases to have any referent. Absolutely nothing can be predicated of Nirguna Brahman, and it can be approached only through the via negatives of neti, neti. Although Śaṅkara would deny it, Brahman ends up as a completely empty ground, unchanging only because it is a Nothing from which all phenomena arise as an ever-changing and hence deceptive appearance. From the perspective of Buddhism, Vedanta reifies śunyatā into an attributeless substance, which, since it has no characteristics of its own, cannot really be said to be at all. From the perspective of Vedanta, however. Buddhism ignores the fact that such a ground is necessary, for as Parmenides pointed out nothing can arise from nothing and it is meaningless to deny all substance: something must be real. Despite this family quarrel, the descriptions converge; what is perhaps more important than the difference is that for both the emptiness of this "ground" (however otherwise understood) is also fullness and limitless richness, for it is lack of any fixed characteristics which makes possible the infinite diversity of the phenomena which arise from "it" 
What is most significant about their argument is that it is no longer a disagreement over the nature of the nondual experience. Since Brahman is qualityless and imperceptible, there is no phenomenological difference between a Mādhyamika interpretation of Huang-po's blow and an Advaitic one. In both cases, the arm-movement is experienced nondually, with no bifurcation between a self-conscious subject and "his" action. In both cases that action is mysterious māyā, inexplicable in terms of efficient causality and having no svabhāva reality of its own (nor, of course, does Huang-po, or anything else). The importance of this agreement is great. The only difference is that Mādhyamika stops here, while Advaita assumes that there must be an unchanging ground as the source of all the changing phenomena. But since this source is by definition imperceptible, the difference is reduced to a far more abstract, although not trivial, one of emphasis: concluding that phenomena are illusory māyā seems to devalue them somewhat more than if phenomena are merely śunya without any Brahman "behind" them. In other words, the difference becomes one of attitude towards the nondual experience rather than anything in the experience itself: the Advaitin, with his dualistic distinction between Brahman and māyā will be more eager to negate the phenomenal world than the Buddhist bodhisattva, for whom there are only empty forms. 
 This is reflected in the etymology of both words. Most scholars agree that Brahman comes from the root brḥ, "to burst forth, grow." "To us, it is clear. Brahman means reality, which grows, breathes or swells." Radhakrishnan, Indian Philosophy (London: Alien and Unwin, 1962J, I, 164 n. Śunyatā is from the root śu, which means "to swell" in two senses: not only "hollow or empty," but also "to be swollen" in the sense of full, like the womb of a pregnant woman. It has been unfortunate for Buddhist studies that the English translation "emptiness" captures only the first sense.
 I am grateful to the "Singapore Mādhyamika Study Group", especially Peter Della Santina, for their comments on an earlier draft of this paper. For further discussion of Mādhyamika, see "How Not to Criticize Nāgārjuna," Philosophy East and West, 34 (Oct. 1984).