Nāgārjuna (c. 200 CE) is the founder of the Madhyamaka school of Buddhist philosophy, and after the Buddha himself, easily the most influential philosopher in the Mahāyāna Buddhist tradition. Despite the great consensus on his philosophical and doctrinal importance, there is little consensus regarding the interpretation of Nāgārjuna’s work, either in the canonical Buddhist and non-Buddhist literature of India, Tibet and East Asia, or in the contemporary secondary literature of European and Asian Buddhist Studies. In virtue of his distinctive doctrine that all phenomena are empty (śūnya), nothing exists ultimately (paramārtha), and that things only exists conventionally (vyvavahāra/samvṛti), he has often been accused of defending nihilism (see Matilal 2002 and Wood 1994). Indeed, he says in Mūlamadhyamakakārikā (Fundamental Verses on the Middle Way) and in Vigrahavyāvartanī (Reply to Objections) that he defends no thesis. This claim and his willingness to deny all four limbs of certain tetralemmas (catuṣkoṭi) add fuel to this fire.
In this chapter, I will argue that this nihilistic reading of Nāgārjuna is unjustified, and that Nāgārjuna is in fact a robust realist, offering an analysis, not a refutation, of existence. On his analysis, to exist is to exist conventionally, and ultimate existence is in fact an incoherent ontological fantasy. I focus on what many take to be the sharpest case for extreme nihilism, Nāgārjuna’s negative catuṣkoṭi. The catuṣkoti (four corners or kotis), or tetralemma, is a standard figure in early Buddhist logic that portions logical space into four possibilities: truth, falsity, both truth and falsity, and neither truth nor falsity. It is deployed in both positive and negative forms. In the positive form, each of the four limbs is asserted; in the negative form, each is denied. The fact that Nāgārjuna sometimes deploys the negative catuṣkoti has been taken by his Indian and Western critics as evidence that he is an arch-nihilist. I will argue that this is far from the case, and that the negative catuṣkoti when properly understood in fact undermines, rather than provides evidence for, a nihilistic reading of Madhyamaka.
In the chapter of Mūlamadhyamakakārikā entitled Examination of the Tathāgata, Nāgārjuna asserts one of the more challenging and paradoxical of his famous tetralemmas:
We do not assert “empty.”
We do not assert “nonempty.”
We neither assert both nor neither.
They are asserted only for the purpose of designation. (Quoted in Tsongkhapa 2006, 447)
Tsongkhapa (fourteenth and fifteenth centuries CE), following Candrakīrti (sixth century CE) closely, comments as follows:
We do not say that because the Tathāgata is empty he is nonexistent, because that would be to commit the error of deprecating him. Moreover, the Tathāgata has been shown to be essenceless. Because we aspire to present the undistorted meaning, nor do we say that he is nonempty— that is, that he exists inherently.
We do not assert both of these; nor do we assert neither that he exists nor does not exist because ultimately, none of these four alternatives can be maintained. On the other hand, if we did not assert these conventionally, those to whom we speak would not understand us. So, from the standpoint of the conventional truth and for conventional purposes, we say “empty” and “non-empty,” “both empty and non-empty,” and “neither empty nor non-empty.” We say these having mentally imputed them from the perspective of those people to whom we are speaking. Therefore, we simply say that “they are asserted for the purpose of designation.” (Tsongkhapa 2006, 448)
This tetralemma might seem to be about as nihilistic as one can get, denying not only the possibility of asserting anything, but also the possibility of denying anything.1 On the other hand, Candrakīrti and Tsongkhapa, in commentary, take some of the sting out of this reading by substituting a positive tetralemma in its place (that which we can say “from the standpoint of the conventional truth and for conventional purposes”), and distinguishing between two philosophical perspectives: an ultimate perspective and a conventional perspective.
In brief,2 the conventional perspective is that of ordinary human consciousness, filtered through conceptual thought, language and our collective and individual modes of taking up with the world. The ultimate perspective is that of a fully awakened being—a buddha—from which all things appear as empty of any intrinsic identity or essence, as interdependent, and independent of their presentation by conceptual thought, or as described by social or linguistic conventions. While Candrakīrti and Tsongkhapa argue that we can’t say anything at all from the ultimate perspective in virtue of its transcendence of the discursive, it is agreed that there is quite a bit that we can say from the conventional perspective. The fact that we can say so much, and that we can make so much sense of truth from the conventional perspective undermines a nihilistic reading of Nāgārjuna’s project and of the negative tetralemma in particular. Understanding why Madhyamaka is a rejection, not an embrace, of nihilism requires us to attend to the two truths, as Candrakīrti and Tsongkhapa suggest, but also to much else, a great deal of which can be examined through the prism of this verse.
In what follows, I will show how Madhyamaka ideas about logic and about the nature of metaphysical dialectic allow us to unpack this tetralemma, to understand its relation to its positive shadow, and to thereby reveal a decidedly realistic attitude emerging from the rhetoric of emptiness. The analysis I offer here is not new, and owes much to Candrakīrti and Tsongkhapa.3 But the refraction through the prism of this particular tetralemma and its commentary offers a new perspective, and may allow us to see how different colors in the Nāgārjunian spectrum combine to illuminate emptiness as an affirmation, not a denial, of the actuality of the world.
This essay is not intended to provide an introduction to Madhyamaka thought in general, or even to the thought of Nāgārjuna in particular. I will have to presuppose a basic familiarity with this landscape. The interested reader is directed to the more general presentations I cite in this paper. I do expect, however, that even for those readers to whom Madhyamaka is terra incognita, some of this will make sense. If so, I hope that it will provide a stimulus to venture further into the world of Indian and Tibetan Madhyamaka, a philosophical world that repays its visitors with rich insights.
As noted above, in a Madhyamaka logical framework, sentences can be (just) true, (just) false, both true and false, or neither true nor false. Nāgārjuna’s concern was not the semantics of conditionals, per se, but was closely related to the concerns that led to a four-valued semantics for relevant logics. In particular, Nāgārjuna was concerned with the possibilities of presuppositional failure, as well as the possibility of a deeply paradoxical reality and the need to reason in these logically challenging environments.4
We often see in Nāgārjuna’s work the catuṣkoṭi, or four-corners, a partition of logical space with regard to a proposition into four possibilities: true, false, both or neither. Nāgārjuna’s approach to the tetralemma is distinctive in that he does not always see these as four alternatives. We see instances in Mūlamadhyamakakārikā in which all four koṭis are affirmed (such as that in chapter XVIII, in which it is said that the self exists, does not exist, both does and does not exist, and neither exists nor does not exist), and instances in which they are all denied, as in the verse on which we focus here. In this case, as we have seen, while all are denied in the root verse, the commentary explains that joint denial through an implicit joint affirmation in a different register.
The joint denial indicates the perspective of ultimate truth, the standpoint from which all phenomena are apprehended as empty. From that perspective, Nāgārjuna argues, nothing at all can be said, since language itself implicates reification, the imputation of intrinsic identity to its putative referents. There are several ideas at work here, and to develop them all would take us far afield. But quickly: In chapter XXII, the context of the negative catuṣkoti under consideration, Nāgārjuna is concerned with the Tathāgata (an epithet used for the Buddha, literally meaning thus-come or thus-gone, depending on how the compound is parsed), and so with how the world appears from the perspective of full awakening. He is asking first what is ultimately true. The answer is, “nothing.” When we assert anything, we implicate the reality of the referent of the subject term, and the qualification of the subject term by a property. But ultimately, nothing exists, and properties, from a nominalist Buddhist perspective, are all fictional. So, from the ultimate standpoint, every sentence we utter is misleading.
Furthermore, even to say that the Tathāgata is neither empty nor non-empty is ultimately impossible, since for him to be ultimately neither empty nor nonempty would be for the state of being neither empty nor non-empty to be his ultimate or intrinsic nature, and that would undercut Nāgārjuna’s insistence that emptiness is the emptiness of any intrinsic nature. This is an important issue. Earlier in the text (XV: 7, 8), Nāgārjuna reminds us that emptiness is the relinquishing of all views, and that anyone for whom emptiness becomes a view is hopeless. In his careful gloss of this verse, Candrakīrti compares someone who takes emptiness to be the intrinsic nature of things to someone who, upon hearing that a shopkeeper has nothing to sell, asks for some of that nothing. On this reading, emptiness really is the lack of any intrinsic nature, or any nature that things have ultimately, not an ultimate nature to replace others.
On the other hand, as Candrakīrti and Tsongkhapa emphasize, from the conventional perspective, we can assert all four koṭis. We can say that the Tathāgata is empty—after all, he lacks any intrinsic nature; that he is non-empty—that is, emptiness is not his intrinsic nature; that he is both empty and non-empty—that is, he lacks intrinsic nature, but does not do so intrinsically; and that he is neither empty nor non-empty—neither intrinsically empty nor actually non-empty.
Note the use of parameterization at the conventional level, contrasting with the univocal nonassertability at the ultimate level. That is, at the conventional level, each of the four koṭis is asserted in a slightly different voice, with qualifiers such as “intrinsically” or “not intrinsically” implicitly qualifying the predicate empty. At the ultimate level, on the other hand, there is no such need for these insertions. Note also that at the conventional level the catuṣkoṭi does not involve us in paradox—the four assertions, suitably parameterized, are mutually consistent—while at the ultimate level the tetralemma takes us straight to paradox, since we are forced to say what we cannot say.5 The elegance of Nāgārjuna’s analysis consists in the demonstration that these tetralemmas in fact encode identical ideas and that the consistent conventional truth is identical with the paradoxical ultimate truth. This insight indeed is Nāgārjuna’s most profound contribution to ontology, and explains why his analysis of reality as empty is realistic, not nihilistic.
As I said above, the charge of nihilism is not a new indictment for Madhyamaka. It was leveled against Madhyamaka by classical orthodox Indian philosophers. Indeed, even Buddhist philosophers who subscribed either to the realistic and reductionist Śrāvakayana schools regarded Madhyamaka as nihilistic, as did the idealist Yogācāra. In each case, their critics accused Mādhyamikas of nihilistically denying the reality of obviously real entities (fundamental dharmas to which all of reality reduces, or consciousness, respectively) in virtue of the doctrine that all phenomena are empty. We will return to Nāgārjuna’s own reply to this charge in Mūlamadhyamakakārikā below. But for now, it may be instructive to introduce Tsongkhapa’s approach to rebutting this charge. Tsongkhapa was a major Tibetan commentator (fourteenth and fifteenth centuries) on Nāgārjuna and Candrakīrti whose interpretation of Madhyamaka is developed in several of his important treatises, including Great Exposition of the Path to Enlightenment (2000), Ocean of Reasoning (2006) and Elucidation of the Purport of Madhya-maka (not available in a Western language, see Tsongkhapa 2009).6
“Emptiness,” Tsongkhapa emphasizes, requires a parameter. We must specify that of which an empty thing is empty. For example, the room in which I write is indeed empty of elephants, but it is not empty of people. When I leave, the room will be empty of people, but not of furniture, and so on. If I were to remove all of the furniture, it would still not be empty of air. To assert simply that something is empty is hence not to assert anything at all. We must add what Tsongkhapa calls “a qualifying phrase” in order to complete the assertion. He emphasizes that emptiness is a negation, indicating the absence of something— and the kind of negation relevant to emptiness in the Mādhyamika’s sense is an external, or nonimplicative, negation, not projecting an alternative property or entity of which the empty thing is full.7 In order to understand the claim that something is empty in Nāgārjuna’s sense, we must identify what Tsongkhapa calls “the object of negation (dgag bya)” or the property that is denied to inhere in the subject in question. The opponent who regards Madhyamaka as nihilistic, Tsongkhapa argues, misidentifies the relevant object of negation as mere existence. If this were the object of negation, then to say that all phenomena are empty would be to say that all phenomena lack even mere existence, and that would of course to say that everything is nonexistent, a nihilistic position if ever there was one. But, Tsongkhapa argues, this is a misreading of Nāgārjuna and his followers. The correct object of negation, he urges, is not mere existence, but rather intrinsic existence.
The Sanskrit word here is svabhāva, and translation into English is a notorious can of worms. I have often used essence to capture its meaning, but that is not ideal. Others have used substance, substantial existence, self-nature, self-being, own-being, essential existence, and so forth. A widely accepted translation now is intrinsic nature. But no English expression captures the Sanskrit perfectly. The idea is this: to have svabhāva is to exist independently, to have a property (a svabhāva) that makes the thing the thing it is, to be capable of existing as that thing independently of anything else. Tsongkhapa argues that the primal confusion that Buddhists regard as conditioning attraction and aversion—and hence as the very root of the existential suffering Buddhism aims to eradicate—consists principally in the projection of this kind of existence onto entities. The mistaken attitude includes the attribution of this kind of existence to our selves and to all objects of our experience, as well as to any fundamental constituents in which we might believe that these entities consist. Emptiness as articulated by Nāgārjuna, Tsongkhapa argues, is the absence of this kind of intrinsic existence.
To be empty of intrinsic identity, according to Candrakīrti and Tsongkhapa, is hence not to be nonexistent, but rather to exist interdependently, relationally, nonessentially, conventionally. To assert that things are empty, then, is not to assert that they are nonexistent, but to assert that nothing has any essence, any intrinsic identity or reality, that to see the world in terms of substantially existent phenomena is a fundamental metaphysical error. This insight in turn grounds Nāgārjuna’s doctrine of the two truths or two realities—conventional and ultimate truth/reality.8
Conventionally, there are people, dogs, tables, and chairs, and each of these things has countless properties, prosaic and profound. These things exist dependently upon one another, on other entities, and on their constitutive parts. Their status as entities as well as their properties further depend upon our cognitive and sensory apparatus and social, cognitive, and linguistic conventions. Since they are dependent, these things and properties do not exist ultimately. To exist conventionally is hence precisely to be empty of ultimate reality. The ultimate truth is that everything is empty; that nothing is ultimately real, not even emptiness. Emptiness too, is only conventionally real; it is only the emptiness of empty things. If emptiness were more real than that, then conventional reality would indeed be a less than full reality, since it would be a second-grade existence contrasting with a possible first-grade existence. But with no such contrast, conventional existence is the only kind of existence that is possible.
Before returning to the tetralemma with which we began, it is instructive to examine two other well-known verses of Mūlamadhyamakakārikā, verses that many take to be the very heart of the text, and indeed the heart of Madhyamaka philosophy itself. The eighteenth and nineteenth verses of chapter XXIV, the investigation of the four ennobling truths, appear as the climax of a response to an interlocutor anticipated in the opening verses of the chapter. This interlocutor is clearly a Śrāvakayana Buddhist, who Nāgārjuna imagines charging him with nihilism. In verse one, we hear the opponent complaining:
If all this is empty,
There would be neither arising nor ceasing.
And for you, it follows that
The four noble truths do not exist.
(Quoted in Tsongkhapa 2006, 472)
After a cascade of reductio consequences, the opponent concludes at verse six:
Hence you undermine the fruits
As well as the profane:
The Dharma itself
And all mundane conventions.
(Quoted in Tsongkhapa 2006, 476)
The opponent’s position here is clear. Madhyamaka is nihilism, and Nāgārjuna’s philosophical position is not merely inconsistent with Buddhist doctrine; it flies in the face of ordinary reason, making a hash of everyday life and common sense. It is in the context of this charge that we encounter verses eighteen and nineteen:
That which is dependent origination
Is explained to be emptiness.
That, being a dependent designation
Is itself the middle way.
There does not exist anything
That is not dependently arisen.
Therefore there does not exist anything
That is not empty.
(Quoted in Tsongkhapa 2006, 503, 505)
These verses repay careful reading and contemplation. In the first, Nāgārjuna identifies dependent origination and emptiness, and by implication, conventional and ultimate truth. He then asserts that this identification is itself merely conventional, and hence empty, and that it, and its emptiness, constitute the middle path between reification and nihilism he is at pains to limn. Let us consider each point in turn.
First, to be dependently originated—to exist in dependence on causes and conditions, in relation to other things, and to have an identity dependent on conceptual designation—is what it is to be empty. Note that to exist dependently is not to be nonexistent; instead, for a Mādhyamika, it is the only way to be real. When we consider things carefully, Nāgārjuna suggests, that is how everything exists. Emptiness, then, is not nonexistence.
Second, even though emptiness is the ultimate reality of things, it is no more real than any conventionally real thing. Emptiness is a dependent designation— merely a verbal formula; dependent origination is merely a dependent designation, and even the fact of the identity of emptiness and dependent origination is merely a dependent designation, a way of taking things. Hence, all are empty of intrinsic identity. Emptiness is only the emptiness of empty things, not a self-subsistent universal. It, too, is therefore merely conventionally real. The conventional reality of things is their emptiness, and hence their ultimate reality. And so, the two truths—conventional and ultimate, the world of dependent arising and the emptiness that is ultimate reality—are identical.
The identification of the two truths, Nāgārjuna asserts, is the middle way. To take emptiness to be distinct from dependent origination or to take emptiness to be ultimately real while taking everything else to be merely conventionally real, would be, as Tsongkhapa felicitously puts it, to fall into both extremes simultaneously. Emptiness would be reified as an ultimately existent phenomenon, and conventional reality would be deprecated as a second-class existence, as illusion. We would then be stuck with an inaccessible real world and an illusory world we are condemned to inhabit. Only by understanding the real world as empty of intrinsic reality, by understanding the emptiness of intrinsic reality to be interdependence, and so by understanding reality to be causally and conventionally interdependent, can we take reality, ourselves, and our analysis of being itself seriously. Emptiness is not on this view an alternative to existence, but an analysis of existence. This is the true spirit of the middle way.
So, let us return to the tetralemma with which we began and consider two ways of reading it in the context of this understanding of emptiness. Let us recall the tetralemma.
We do not assert “empty.”
We do not assert “nonempty.”
We neither assert both nor neither.
They are asserted only for the purpose of designation.
There are two ways to read the verse and its positive image, and each is illuminating. First, we can think of the verse itself straightforwardly in terms of presupposition failure. On this reading, when Nāgārjuna recuses himself from each of the four koṭis of the catuṣkoṭi, he does so to indicate that he is speaking from the ultimate perspective (or at least attempting to do so, indicating the impossibility of actually doing so). As we noted at the outset, since all language is conventional, and since it can only designate what exists conventionally, any assertion can at best be conventionally true. The last line affirms this. Conventionally, we can say any of these things; ultimately, we can say none, because the presuppositions of assertion—the reality of referents and the possibility of instantiating properties—are not satisfied from that perspective. This is the reading with which we began.
But there is a second, slightly deeper, way to take this tetralemma and the positive shadow to which Candrakīrti and Tsongkhapa direct our attention. I have argued that this is the reading that takes us closer to the heart of Madhya-maka and dispels once and for all any nihilistic reading. This reading illuminates the deep paradox Nāgārjuna finds at the heart of reality. When Nāgārjuna says “we do not assert’empty’,” he indicates that to say that phenomena are empty is fundamentally contradictory: it is to say that they have no intrinsic nature at all, but since to exist is to be empty, nothing can exist without being empty, and so emptiness is the intrinsic nature of anything that exists; hence the intrinsic nature of things is to lack intrinsic nature (see Garfield and Priest 2003). However, Nāgārjuna does not assert “non-empty” either. To do so would also be contradictory. It would be to assert that things have intrinsic nature, which, for a Mādhyamika, is incoherent. Each of the first two koṭis is thus, in some sense, nonassertible, or at least, not consistently assertible.
To assert both is a plain contradiction, as is to assert neither, since they are mutually exclusive and exhaustive. So, on this reading, the negative tetra-lemma indicates the fact that ultimate reality is deeply paradoxical. To say anything about it, to maintain any of the four alternative positions, lands us in a paradox. Silence would then appear to be the only way to maintain consistency. But even silence can only achieve consistency if it is articulate silence—the silence of Vimalakīrti, who replies to Mañjuśrī’s request to explain nonduality, despite the fact that that all language is dualistic by remaining silent. It is not the silence of Śāriputra, who earlier in the Vimalakīrti-nirdeśa-sūtra replies to the goddess’s request to explain a difficult philosophical point by remaining silent and citing inexpressibility as an excuse. But even Vimalakīrti’s silence, if it is articulate at all, lands us in a paradox. For it says both that nothing can be said, and that. Silence, that is, if it is to be articulate, is a kind of speech; and when taken as speech, it lands us back in the paradox it seeks to avoid.
On the other hand, the fourth line still indicates the positive shadow tetra-lemma. Conventionally we do assert all four of these, and each, from a conventional standpoint, is entirely unproblematic. That positive version might appear to be the contradictory tetralemma; after all, its first two kotis are the negations of one another, and the third and fourth are explicit contradictions. But, in virtue of its straightforward parameterization, it is in fact consistent. Conventionally, we can certainly assert that all phenomena are empty—after all, they are dependently arisen; we can assert that all phenomena are not empty (ultimately), both conventionally empty and ultimately non-empty, and neither ultimately empty nor ultimately non-empty. Conventional reality, the only reality we inhabit, is the domain of speech and of all truth.
But as the fourth line of the root verse indicates, and as XXIV: 18, 19 explain, the assertible positive tetralemma is just the other side of the nonassertability of any limb of the negative tetralemma. It is because emptiness is dependent origination—because the ultimate truth and the conventional truth are in the end identical—that it is true conventionally that all things are empty, that it is true that their emptiness is the ultimate truth; and that ultimately that cannot be true, simply because nothing is ultimately true.
It is because we can say nothing from the ultimate standpoint that the conventional is the only standpoint we have. The conventional is thus the only framework within which truth is possible, including this truth. It is for this reason that we need all four possibilities represented in the catuṣkoṭi. We need to be able to talk about getting it right, getting it wrong, speaking when only contradictions can reveal the truth, and when all speech fails. The catuṣkoṭi is therefore not a mere rhetorical flourish in Nāgārjuna’s hands; it is a reflection of the structure of emptiness and its relationship to reality. It is the only logic adequate to a Madhyamaka metaphysics, the only logic that can express that metaphysics in a way that reveals both its paradoxical character and its thoroughgoing realism.
The negative tetralemma is taken by most critics of Madhyamaka as the most decisive evidence for that reading. After all, it does seem to say explicitly that nothing whatever is true. We have seen not only that it provides no evidence for that reading, but that when unpacked, the negative tetralemma is in fact a profound logical and rhetorical device for exploring the positive ontological significance of the Madhyamaka doctrine that all phenomena are empty of intrinsic nature.
In reading the tetralemma this way, we have also seen that the negative tetra-lemma is equivalent to the positive tetralemma it projects. Everything that is unsayable from the ultimate perspective is assertible from the conventional perspective, and not accidentally so. It is emptiness that makes sense of conventional reality, and conventional reality that explains emptiness. A proper understanding of emptiness thus entails the identity of the two truths. But if that is so, to take the emptiness of all phenomena seriously is to take the conventional reality of all phenomena seriously. And to take reality seriously is precisely to deny nihilism. From the standpoint of Madhyamaka, to be empty is hence not to be nonexistent, but rather is the only possible way to exist.
1 Nāgārjuna certainly has been read that way, both by his classical Indian interlocutors, especially the Nayāyikas, and by such contemporary commentators as B. K. Matilal (2002) and Thomas Wood (1994).
2 See Cowherds (2010); Garfield (1995, 2001); Newland (2011); and Westerhoff (2010) for more detail.
3 I have also defended much of this both on my own in numerous essays, and with the Cowherds in our recent polygraph Moonshadows (Cowherds 2010).
4 As Richard Routley noted a few decades ago (Meyer and Routley 1972, 1973; Plumwood and Routley 1982), Nāgārjuna scooped Meyer and Dunn (1972) by a few millennia in proposing a four-valued approach to logic (or at least a four-way valuational approach, as the set of truth values remains binary—only the assignment function goes quarternary. A similar approach, and similar concerns, of course, motivates the Meyer-Dunn semantics for Relevant Logic, and the gappy and glutty worlds it requires.
5 See Garfield and Priest (2003) for more on these paradoxes.
6 The Cowherds (2010) develop this idea in detail in Moonshadows, but we can sketch the important points here in short compass.
7 Indian logic recognizes two kinds of negation, external, or nonimplicative negation, and internal, or implicative negation. The former is a sentential negation, and the latter a predicate negation. When I say, for instance that I have no brown horse (external) I do not implicate that I have a horse of a different color; when I say that my horse is not brown (internal) I do.
8 The Sanskrit satya denotes either truth or reality, and different readings may be appropriate in different contexts.