language, understanding and reality: a study of their relation in a foundational metaphysical...

31
Language, Understanding and Reality: A Study of Their Relation in a Foundational Indian Metaphysical Debate Eviatar Shulman Published online: 16 May 2012 © Springer Science+Business Media B.V. 2012 Abstract This paper engages with Johaness Bronkhorst’s recognition of a “cor- respondence principle” as an underlying assumption of Na ¯ga ¯rjuna’s thought. Bronkhorst believes that this assumption was shared by most Indian thinkers of Na ¯ga ¯rjuna’s day, and that it stimulated a broad and fascinating attempt to cope with Na ¯ga ¯rjuna’s arguments so that the principle of correspondence may be maintained in light of his forceful critique of reality. For Bronkhorst, the principle refers to the relation between the words of a sentence and the realities they are meant to convey. While I accept this basic intuition of correspondence, this paper argues that a finer understanding of the principle can be offered. In light of a set of verses from Na ¯ga ¯rjuna’s Śūnyatāsaptati (45–57), it is maintained that for Na ¯ga ¯rjuna, the deeper level of correspondence involves a structural identity he envisions between understanding and reality. Here Na ¯ga ¯rjuna claims that in order for things to exist, a conceptual definition of their nature must be available; in order for there to be a real world and reliable knowledge, a svabhāva of things must be perceived and accounted for. Svabhāva is thus reflected as a knowable essence. Thus, Na ¯ga ¯rjuna’s arguments attacks the accountability of both concepts and things, a position which leaves us with nothing more than mistaken forms of understanding as the reality of the empty. This markedly metaphysical approach is next analyzed in light of the debate Na ¯ga ¯rjuna conducts with a Nya ¯ya interlocutor in his Vigrahavyāvartanī. The correspondence principle is here used to highlight the metaphysical aspect of the debate and to point out the ontological vision of Na ¯ga ¯rjuna’s theory of empti- ness. In the analysis of the Vigrahavyāvartanī it becomes clear that the discussion revolves around a foundational metaphysical deliberation regarding the reality or unreality of svabhāva. In this dispute, Na ¯ga ¯rjuna fails to answer the most crucial point raised by his opponent—what is that he defines as empty? E. Shulman (&) Hebrew University of Jerusalem, Jerusalem, Israel e-mail: [email protected] 123 J Indian Philos (2012) 40:339–369 DOI 10.1007/s10781-012-9158-z

Upload: huji

Post on 23-Jan-2023

0 views

Category:

Documents


0 download

TRANSCRIPT

Language, Understanding and Reality: A Studyof Their Relation in a Foundational IndianMetaphysical Debate

Eviatar Shulman

Published online: 16 May 2012

© Springer Science+Business Media B.V. 2012

Abstract This paper engages with Johaness Bronkhorst’s recognition of a “cor-

respondence principle” as an underlying assumption of Nagarjuna’s thought.

Bronkhorst believes that this assumption was shared by most Indian thinkers of

Nagarjuna’s day, and that it stimulated a broad and fascinating attempt to cope with

Nagarjuna’s arguments so that the principle of correspondence may be maintained

in light of his forceful critique of reality. For Bronkhorst, the principle refers to the

relation between the words of a sentence and the realities they are meant to convey.

While I accept this basic intuition of correspondence, this paper argues that a finer

understanding of the principle can be offered. In light of a set of verses from

Nagarjuna’s Śūnyatāsaptati (45–57), it is maintained that for Nagarjuna, the deeper

level of correspondence involves a structural identity he envisions between

understanding and reality. Here Nagarjuna claims that in order for things to exist, a

conceptual definition of their nature must be available; in order for there to be a real

world and reliable knowledge, a svabhāva of things must be perceived and

accounted for. Svabhāva is thus reflected as a knowable essence. Thus, Nagarjuna’s

arguments attacks the accountability of both concepts and things, a position which

leaves us with nothing more than mistaken forms of understanding as the reality of

the empty. This markedly metaphysical approach is next analyzed in light of the

debate Nagarjuna conducts with a Nyaya interlocutor in his Vigrahavyāvartanī.The correspondence principle is here used to highlight the metaphysical aspect of

the debate and to point out the ontological vision of Nagarjuna’s theory of empti-

ness. In the analysis of the Vigrahavyāvartanī it becomes clear that the discussion

revolves around a foundational metaphysical deliberation regarding the reality or

unreality of svabhāva. In this dispute, Nagarjuna fails to answer the most crucial

point raised by his opponent—what is that he defines as empty?

E. Shulman (&)

Hebrew University of Jerusalem, Jerusalem, Israel

e-mail: [email protected]

123

J Indian Philos (2012) 40:339–369

DOI 10.1007/s10781-012-9158-z

Keywords Nagarjuna · Śūnyatāsaptati · Vigrahavyāvartanī · Indian Metaphysics ·

Nyaya · Madhyamaka · Emptiness · Svabhāva

The present contribution to the study of Indian philosophy springs from Johannes

Bronkhorst’s recent Language and Reality: On an Episode in Indian Thought. In thisbook, published originally in French in 1999, Bronkhorst exposes a principle that,

he argues, underlies much of Indian philosophical thinking for about half a

millennium, beginning with the work of Nagarjuna. Bronkhorst defines this as “the

correspondence principle”, the idea “that the words of a sentence correspond rather

exactly to the things constituting the situation described by the sentence.”1

Bronkhorst’s view of this formative period in the history of Indian thought

possesses one particularly compelling feature: it reveals the common interests that

are normally hidden behind the multiplicity of philosophical schools. In this way

Bronkhorst is able to point out a fundamental coherence that is at the base of Indian

philosophical discourse; the disparate systems prove to be part of a well integrated

and multi-faceted shared concern that stems from an essential connection they

identify between language and physical reality. The different darśanas—Bronkhorst

refers to the main systems and thinkers of the period, including Vaises˙ika (which is

of central importance to his reading), Nyaya, Gaud˙apada’s and Sankara’s Advaita-

vedanta, Sam˙khya, Jainism, Kashmiri Saivism, Bhartr

˙hari and a number of central

Buddhist thinkers and traditions—are all shown to respond to Nagarjuna’s

arguments, which purport to deconstruct reality under the assumption of a

comprehensive, one-to-one correspondence between words and things; since the

verbal structure is shown to defeat itself, the reality it describes is said to be void.

Various Indian thinkers were thus propelled by an impulse to conceptualize and

defend positions that attempt to uphold the existence of reality, while at the same

time maintaining the implicit assumption of correspondence.

Many will argue against Bronkhorst’s inclination to regard so much of Indian

philosophy as revolving around this one principle of correspondence. Others will

argue with the way he formulates the principle itself. In my mind, although there is

room for debate regarding the way Bronkhorst frames the position of specific

systems and connects them to the broader theme of his discussion, and although his

definition of the principle is not fully satisfying, I support the emphasis on

correspondence as an underlying structural theme and a true lakṣaṇa of Indian

philosophy. As shown by Gonda (1969), this notion has strong Vedic antecedents;2

it is also clearly central to many of the major Upanis˙ads, as shown by Bronkhorst in

the introduction to his book. Here he quotes passages such as Bṛhadāraṇyaka-upaniṣad 1.4.7, which teaches that “At that time, indeed, the world was undivided.

Name and form divided it [or: it was divided by name and form], such that one says:

1 P. 1. For other definitions of the correspondence principle see pp. 92, 135.2 See also Gonda (1966, pp. 105–106): “Attention may be also drawn to the fact that a correlation is

sometimes explicitly supposed to exist between the main characteristics of the ‘idea’ or ‘concept’ with

which one concentrating upon it identifies oneself, the status to be obtained in this world and the position

(lokāḥ) to be won…”. See also Brereton (1999), who sees thought as being at the base of reality according

to Ṛg-Veda 10.129.

340 E. Shulman

123

it has this name and that form. Today name and form divide this same world, such

that one says: it has this name and that form.”3 Word and thing, we learn, were a

joint creation, when primeval being was given shape; the two relate to each other by

way of their very constitution. With such deep roots in the core of Indian speculative

thought, Bronkhorst sees the correspondence principle as “not primarily a logical

position, but rather an intuition shared by thinkers of the period” (pp. 38–39).

It is important to notice that the correspondence principle can be easily seen as an

expression of a wider interest in the relation between subjective perception, or between

subjectivity in general, and reality. Bronkhorst is interestedmainly in the way the wordsof a sentence correspond to the realities they aim to convey. But if a broader

understanding of the correspondence principle is allowed for, an understanding that

would involve the relationship between human consciousness, rather than only the

words of the sentence or language, on the one hand, and reality on the other, then the

explanatory power of the principle improves dramatically. Under such a definition,

someof the hallmark theories of Indian philosophical thought, such as the creation of the

world by prakṛti in response to puruṣa’s interested gaze in classical Sam˙khya,4 or the

Buddhist understanding of karma (that infiltrated into most Indian religious systems),

would prove to be founded on such a notion of correspondence aswell.5 In fact, the issue

of correspondence can itself be seen as a particular expression of a more fundamental

insight that sees understanding, or subjectivity, as essential to reality. This is indeed one

of the more intriguing aspects of the correspondence principle—it demonstrates how

central strands of Indian philosophical thinking were structured around an understand-

ing that grants subjectivity a pivotal role in the molding of the physical world.6

3 Translation from Bronkhorst (2011, p. 3): tad dhedhaṃ tarhy avyākṛtam āsīt / tan nāmarūpābhyām evavyākriyatāsaunāmāyam idaṃrūpa iti / tad idam apy etarhi nāmarūpābhyām eva vyākriyate ‘saunāmāyamidaṃrūpa iti //4 See Sāṃkhyakārikā 21, 57–63.5 See also Shulman (2012), for the importance of this principle for Indian religion and contemplative

practices.6 Mohanty (1992, p. 33) is aware of this basic pattern in Indian philosophical thought when he says:

What then is the nature of consciousness? While all Indian philosophers agreed that consciousness,

when it has an object, performs the function of revealing, manifesting, illuminating it, they differ

widely over what consciousness itself is, and in their answers to the various standard disputational

issues. There was near unanimity, however, on one thing: consciousness in the ultimate source ofthe guarantee that anything exists, or is thus. You cannot say that something exists, and is thus,

without making use of the testimony of (your) consciousness at some level. If asked to justify your

ontological claim, you may first fall back on evidence at your disposal: perceptual evidence,

grounds for legitimizing your inference if you have made one, or reports of reliable persons. But

each of these must be presented to your consciousness. In the absence of your consciousness, you

cannot assert anything; nothing can be said by you to be or to be thus or not thus. If you are arealist in your ontology, you may still have your world, but you cannot posit, assert its existence,determine its nature. There would be ‘darkness of the world’ (jagadāndhyaprasaṅga). That

knowledge is a mode of consciousness, or at least presupposes it, was widely recognized. Note thatthis priority is only epistemic or evidential priority. It may be that once this much is granted, thethesis that consciousness is also ontologically prior will follow. But that step was not taken by all”(emphases mine).

For further discussion see Mohanty (1992, Chap. 2; 1995 [1993], Chap. 1, “Philosophy as Reflection on

Experience”).

Language, Understanding and Reality 341

123

In what follows I will not be interested in expanding the correspondence

principle to such a degree; my main aim is much more restricted. Given the

centrality of Nagarjuna’s critique of reality to Bronkhorst’s account7—his

discussion is mainly an exposition of the way different Indian systems responded

to Nagarjuna, in what probably lays to rest Hayes’ (1994, p. 299) thoughts about the

tangential role Nagarjuna played in the development of Indian philosophy8—I will

introduce a set of thirteen verses from Nagarjuna’s Śūnyatāsaptati (SS) that resonatein an interesting way with Bronkhorst’s correspondence principle.9 In fact, I believe

these verses suggest a surprising and subtle level on which the correspondence

principle operates: Nagarjuna appears to be saying that understanding is an

inevitable condition of existence. Thus, it may be better to speak not only of a

correspondence between words and things, but even of a structural identity between

words, perception or understanding, and the so-called objective reality. For

Nagarjuna, according to these verses, not only do words/sentences and things

correspond, but in order to exist, things must possess the capacity for being

understood and conceptually defined. According to Nagarjuna, things do not exist,

of course—nothing can be vividly grasped, since nothing has self-nature (svabhāva)—and hence all is unreal. Nevertheless, this is it to him what it means to exist, and it

is in this sense things can be said to be true—as forms of (mistaken) understanding.

Interestingly, the verses we will discuss from the SS also betray at least some degree

of awareness of the correspondence principle itself, for which Bronkhorst finds

little, if any, evidence in the scriptures.10

It is possible that to a Madhyamaka scholar, the statement that understanding is a

condition of existence may not seem all that innovative; it can be read as a

rephrasing of the well-known Tibetan Madhyamaka principle of things having to be

“findable under analysis” in order to be seen as real,11 or of the statement that things

are “merely imputed.”12 Nevertheless, I believe that these verses state much more

and that they highlight the metaphysical aspect of Nagarjuna’s approach; these

verses suggest that susceptibility to analysis is a part of the very existence of things,

an intuition that is not usually seen to be part of the sober and rational Madhyamaka

critique. Furthermore, this discussion implies a position according to which things

are true to the degree that they correspond with the confounded way in which they

are known.

7 Bronkhorst had previously concentrated on Nagarjuna’s employment of the Correspondence Principle

in Bronkhorst (1997).8 Bronkhorst refers to Hayes’ article directly on page 44.9 Bronkhorst relies in his book only on the MMK, since the question of the authorship of the other works

attributed to Nagarjuna is “far from resolved” (p. 39). I have presented my arguments in favor of

assessing Nagarjuna’s philosophy in light of the four philosophical treatises whose attribution to his name

is relatively reliable—the MMK, VV, SS, and the Yuktiśaṣṭikākārikā—in Shulman (2007 [2009], Sect. I).10 Bronkhorst (2011, pp. 134–135).11 E.g. rJe Tshong Khapa (2006, pp. 39–41), as well as Hopkins (1996 [1983], pp. 406–407) and Garfield

(1995, pp. 90–91).12 E.g. Hopkins (1996 [1983], pp. 437–438).

342 E. Shulman

123

Once the importance of the correspondence principle for Nagarjuna is

appreciated and its definition refined, I will show how this principle helps us

understand the main issues that are at stake in Nagarjuna’s Vigrahavyāvartanī (VV).We will see that the main question debated in this text is a metaphysical one

regarding the ontological reality of svabhāva; said another way—the debate about

svabhāva is primarily metaphysical. Nagarjuna and his opponent in this compelling

text are divided mainly in their idiosyncratic understandings of what type of reality

is suggested by the correspondence principle, whether it is one that conforms to a

realist or to an illusionist world view. The dispute in the VV thus proves to be a

prime example of this foundational Indian metaphysical debate.

According to Bronkhorst, the challenge Nagarjuna poses to Indian thinkers can

be summarized as such: given the implicit assumption of the correspondence

principle, Nagarjuna argues that since he shows that the words of a sentence cannot

represent an intelligible reality, there is no such a reality. The paradigmatic case in

which Nagarjuna introduces this enigma is in the discussion of motion in Chap. 2 of

the MMK. The heart of the problem is in the phrase gamyamānasya gamanaṃ—the

movement of the mover13—which is central to MMK 2.3–2.6: If the terms

‘movement’ and ‘mover’ are taken to imply true realities in the sense that Nagarjuna

demands a reality to be, as something that possesses its nature in and of itself (as

possessing svabhāva), this phrase must be seen as speaking of two movements, one

for the movement and one for the mover, which must in turn be conducted by two

different movers.14

For Bronkhorst, the implicit assumption of the correspondence principle helps us

understand some of the more puzzling statements of the MMK, a claim I believe is

13 In Shulman (2010) I argued for the reading of gamyamāna as “the mover,” although it is often taken,

under the influence of Candrakırti, to refer to the part of the path that is being traversed in the present. The

precise meaning of gamyamāna is of no importance to the questions discussed here, since the problem

addresses the connection between the gamyamāna, however it is defined, and gamanam, movement.

Bronkhorst quotes MMK 2.5–2.6 (p. 41), but refers also to the statement gantā gacchati from MMK

2.9–2.11 (see Bronkhorst 1997, p. 35).14 The heart of the argument is presented in MMK 2.3–2.6:

2.3 How will it be reasonable for there to be movement of the mover, when there can certainly be

no mover devoid of movement?

gamyamānasya gamanaṃ kathām nāmopapatsyate / gamyamānaṃ vigamanaṃ yadānaivopapadyate //2.4 For whom there is movement of the mover, there will follow a mover without movement. The

mover, though, is the one being moved (/upon).

gamyamānasya gamanaṃ yasya tasya prasajyate / ṛte gater gamyamānaṃ gamyamānaṃ higamyate //2.5 When there is movement of the mover, two movements must be involved: the one by which he

is a mover, and the one which is (his) movement.

gamyamānasya gamane prasaktaṃ gamana-dvayaṃ / yena tad gamyamānaṃ ca yac cātragamanaṃ punaḥ //2.6 When there follow two movements, two movers must follow. Movement without a mover

obviously doesn’t make sense.

dvau gantārau prasajyete prasakte gamana-dvaye / gantāraṃ hi tiraskṛtya gamanaṃnopapadyate //

The translations are my own. MMK 2.5–2.6 are quoted and translated by Bronkhorst on p. 41 of his book.

Language, Understanding and Reality 343

123

justified. In fact, if the issue of the way words of sentences connect to each other

need not be understood as the focus of the correspondence principle, but rather as

only an expression of it, we can see that the idea that there is a correspondence

between analysis and reality is precisely what grants Nagarjuna’s argument its

exceptional force. The fact that Nagarjuna assumes that his conceptual refutation of

bhāva (“existence” or “things”) and of svabhāva (“self-nature”) corresponds to the

state of things invests his argument with a unique deconstructive capacity.15 This is

true both if his argument is construed as making a metaphysical and ontological

claim, or if he is making an epistemological and a skeptical one. In both cases, the

assumption is that things correspond to conceptual categories; the debate is only

whether what is refuted is the existence of things or the categories that attempt to

define them.

Bronkhorst’s claim regarding the implicit assumption of the correspondence

principle (presented initially in Bronkhorst 1997), has been challenged by Oetke

(2004). Oetke supports Bronkhorst’s search for “hidden premises” or “hermeneu-

tical principles” in Nagarjuna’s texts.16 Still, he suggests as assumptions that are at

the base of Nagarjuna’s logic the “idea of conceiving conceptual relationships as

something pertaining to the inherent nature of the entities falling under the concepts

concerned”, or “the principle of understanding dependence-relationships which are,

as a matter of fact, of conceptual nature after the model of causal dependence.”17 I

have questioned Oetke’s criticism of Bronkhorst’s views on this matter elsewhere

and will avoid entering a lengthy discussion of this issue here. Suffice to say that I

do not believe that Oetke’s approach manages to cast a serious doubt regarding

Bronkhorst’s argument in favor of seeing the correspondence principle as lying at

the base of some of Nagarjuna’s most important arguments; in fact, I believe

Oetke’s and Bronkhorst’s views have much more in common than Oetke appears to

admit.18 The present discussion will show that Bronkhorst’s correspondence

15 See the quotes from Oetke (2004) in the following paragraph, which express this idea clearly.16 Oetke (2004, p. 84).17 Oetke (2004, p. 92). Oetke offers this new principle in place of the one quoted by Bronkhorst (1997,

p. 30) and taken from Oetke (1990) that: For all x and all y: if x is the condition of y / if x is the condition

of the existence of y, then y must be something that exists during the existence of x (or: that does not exist

exclusively later than x). Note that there is, in fact, quite a dramatic shift in Oetke’s position, and that this

new view has much in common with Bronkhorst’s positions (see in the following note).18 In Shulman (2010, p. 403, n. 74) I spoke of Oetke’s reference to a “noun-related” and a “verb-related”

movement in MMK 2 as being in tune with Bronkhorst’s correspondence principle. I grant that this

statement may not have done justice to the subtlety of the views developed in Oetke’s paper.

Nevertheless, I still maintain that much of what Oetke says fits well with Bronkhorst’s formulation. In

fact, when Oetke states in his analysis of MMK 2.22–2.23 that “Nagarjuna’s argument here as elsewhere

crucially depends on a transformation of relationships holding on the sentential or ‘propositional’ level to

relationships holding on the level of entities denotable by singular terms” (p. 86), this seems to me, if I

understand Oetke correctly, to be a helpful explanation of Bronkhorst’s correspondence principle; the

problem is in fact that sentences are composed of terms that pretend to denote singular entities. I also do

not see why Bronkhorst’s definition cannot account for Nagarjuna’s attempt as a “logician” to refer “to all

pertinent logical possibilities” (p. 89); the analysis of all possibilities is obviously central to any account

of “Nagarjuna’s logic”, including the one developed by Bronkhorst. Oetke’s arguments gain force when

he states that only a strong version of the correspondence principle, which would include the notion

‘X exists at t’ whenNagarjuna says something of the like that ‘X originates at t’, would possibly vindicate the

correspondence principle (p. 90). Here I would suggest that this is indeed a necessary condition the

344 E. Shulman

123

principle is well suited to make sense of the verses from the SS under discussion,

even-though it will call for a certain adaptation of the principle.19 We will thus gain

some back-wind in our effort to decipher the positions that are at the core of

Nagarjuna’s logic.

Returning to the question of double movement articulated in MMK 2—are there

two movements when the mover moves?—this problem can easily be seen to

involve the relation between the words of a sentence and the reality they are meant

to express. This is the relation I understand Bronkhorst to have in mind when he

speaks of the correspondence between the words of a sentence and the reality they

represent. In this case, the heart of the problem is in the relation between a subject

and a predicate, or in a broader definition between the subject and any property he is

thought to possess or action he may carry out. The difficulty to specify the nature of

the relation between these two fundamental elements not only of human language,

but of human cognition, the subject and his property or action, casts a serious doubt

on the liability of both words and things.20

A verse that is of special importance to Bronkhorst’s argument regarding the

correspondence principle in Nagarjuna is MMK 7.17:

If there existed anywhere something unarisen, it could arise.

Given that no such thing exists, what arises?

yadi kaścid anutpanno bhāvaḥ saṃvidyate kvacit /utpadyeta sa kiṃ tasmin bhāve utpadyate ‘sati // 21

This verse exemplifies the problem created by Nagarjuna’s arguments for the view

of reality that relies on the correspondence principle. The correspondence principle,

as I understand Bronkhorst to state it, demands both a fundamental distinction and

an absolute one-to-one relationship between words and things; there must be a gap

Footnote 18 continued

correspondence principle must rely on, but one that I assume it naturally does. This is precisely what

gives Nagarjuna’s argument the metaphysical aspect Oetke argues for elsewhere (e.g. Oetke 2011). What

becomes more difficult to answer is the claim made by Oetke in light of VV 23 that for both Nagarjuna

and us it is naıve to assume “that sentences of the form ‘A pratiṣedhayati B’ must entail the existence ofan entity referred to by ‘B’” (p. 94). Indeed, Nagarjuna argues against things that he believes do not really

exist. This is truly a thorny problem, which in mymind is one of the most difficult problems in Nagarjuna—

what is the status of the initial objects or concepts that are shown to be empty; the Madhyamika needs

something to refute. But here, once again, the correspondence principle proves to be helpful since at the

bottom line, in Nagarjuna, it denies the reality of the correspondence—there is only one realm of inquiry,

that of conceptual analysis. Hence one needs not assume anything but the conceptual analysis itself when

realizing that a reality beyond the analysis in undermined. (This itself leads to another yet more pressing

problem in Nagarjuna that is the reliance on the given of the conceptual analysis itself. See below, p. 37.)

Finally, I believe that Oetke’s revised definition of the principle he identifies at the heart of Nagarjuna’s

thinking and quoted in the paragraph above, which sees “conceptual relationships as something pertaining

to the inherent nature of the entities falling under the concepts concerned” comes very close to being an

explanation of a central aspect of the correspondence principle as defined the Bronkhorst and as it will be

developed in this paper.19 See the discussion of SS 45–46 in “Śūnyatāsaptati 45–57” section.20 Chapters 4–6 of the MMK offer cogent and relatively straightforward expressions of these problems.21 Translation from Bronkhorst (2011, p. 39).

Language, Understanding and Reality 345

123

between them, and they must reflect each other perfectly. It is precisely such an

assumption that is at the root of MMK 7.17—one who wishes to uphold the veracity

of arising must specify what it is that arises—an arisen or an un-arisen thing. Both,

of course, make little sense, and hence, nothing can arise. In order to say that

something arises, only something that hasn’t yet arisen could be considered for the

task, and hence the somewhat humorous “if there were something non-arisen

anywhere, it could arise.” There is, of course, no such thing, and therefore the

simple statement “X arises” cannot correspond to any reality. Assuming the

correspondence principle, Nagarjuna can thus deny the truth of arising.22

Much of the cogency of Bronkhorst’s presentation derives from his ability to

show how deeply Indian philosophy engaged with this problem regarding arising,

defined so poignantly by Nagarjuna. For Bronkhorst, it is clear that this discussion

revolves around the correspondence principle: “this example shows that the arising

of things presents a problem for those who accept the correspondence principle.”23

In what may be the most contestable part of his argument, he reads any discussion of

“the arising of things”, such as different versions of satkāryavāda24 or

asatkāryavāda,25 the Nyaya distinction between ākṛti (“form”) and dravya(“individual” or “substance”),26 or the Vaises

˙ika’s introduction of the category of

abhāva,27 as engagements with Nagarjuna’s arguments against causation. All these

theories express, according to Bronkhorst, an attempt to defend the truth of

phenomenal reality in light of Nagarjuna’s arguments, while upholding the

correspondence principle. Although Bronkhorst’s discussion convincingly shows

that these and other systems of thought took issue with Nagarjuna’s arguments

against arising, it is less clear that this philosophical enterprise necessarily had

anything to do with the correspondence principle.28

22 Regarding this verse, Oetke’s (2004, pp. 89–91) criticism of Bronkhorst is particularly inapplicable.

Oetke believes that “it should be a truism that anything which originates is also unoriginated at some time,

and accordingly the theorem that if anything originates it is at some time to be found as something which

is not originated does not require any special principle in order to appear plausible.” This statement, quite

uncharacteristic of Oetke’s usual tough-mindedness, is precisely the type of lax thinking Nagarjuna

militates against. The point is precisely that an analytic concept must be supplied to make sense of reality

and that one should not rely on such intuitive “truisms”. Rather, the full correspondence between word

and thing demands that one who intends to argue in favor of arising must specify exactly what it is that

arises. Bronkhorst’s book convincingly shows that this challenge was indeed taken up by a great variety

of Indian thinkers.23 p. 40.24 In both Sam

˙khya and certain strands of Advaita. See pp. 50–67.

25 Here Bronkhorst refers mainly to early Vaises˙ika. See pp. 79–88.

26 See esp. pp. 92–93.27 See pp. 117–120.28 It is worth noting that texts from the late Vedic and early Upanis

˙adic period already express an interest

in the question of causation and the nature of arising. The question of how the world first arose—whether

from existence or non-existence, surely troubled Upanis˙adic thinkers (see, for example, Chāndogya-

Upaniṣad 6.2). Creation hymns from book ten of the Ṛg-veda (see esp. 10.129, 10.121), also display a

deep distrust for any available theory of creation. It could be argued that this interest in arising predates

Nagarjuna, and that he is only a stepping stone into the Indian inquiry into the problem. It should also be

kept in mind that Nagarjuna was not alone in forming and articulating the argument against arising, as it

finds expression in many early Mahayana sutras and especially in those that are related to the

346 E. Shulman

123

There are thus two main problems with Bronkhorst’s definition of the

correspondence principle: First, the relation between the correspondence principle

and the question of the arising of things is not as strong as Bronkhorst appears to

believe. The arguments against Nagarjuna’s critique of the notion of arising,

arguments that were produced by a broad range of Indian thinkers, can also be easily

read as arguments against his ontology, without necessarily being related to the

correspondence principle. It is true that the correspondence principle gives

Nagarjuna’s argument an extra edge, but the argument is challenging enough as it

is. In any case, whether the problem that troubled Indian philosophers of the day

was that of correspondence or not, Bronkhorst’s account convincingly shows that all

these thinkers were dealing with similar questions that fit together into one coherent

pattern.

The second problem is more important for the present concerns. This involves

Bronkhorst’s highlighting of the words of a sentence as the item on the mental side

of the correspondence equation. Here we are best to note that although some of

Nagarjuna’s arguments hinge on problems that could be defined as involving the

relation between the different words of a sentence—such as the relation between the

mover and his movement in MMK 2.3–2.6—others do not. An example of the latter

type would be MMK 7.5. Responding to a claim by an opponent who argues that the

arising of arising is equal to arising itself, Nagarjuna says:

If, as you say, the arising of arising is (only) the arising of the basic arising; (so

long as it is still) uncreated by the basic (arising), how do you think it creates it?

utpādotpāda utpādo mūlotpādasya te yadi /maulenājanitas taṃ te sa kathaṃ janayiṣyati //

The argument here appears to be that something must cause the arising of arising, and

that if the basic arising hasn’t caused it, the arising of arising cannot happen.29

Whatever Nagarjuna’s point precisely was, for our context we need only note that this

verse clearly reflects on a problem not with the words of a sentence, but with the verynature of words themselves; the term arising is critiqued not by showing that it cannot

function in a sentence, but by saying that there is a conceptual fallacy in its definition—

arising cannot arise. The question is what gives rise to arising, and Nagarjuna is

claiming that no such thing can exist. The problem is with the concept of arising itself,

and not with the way arising relates to other concepts it could join in a sentence.30

I believe that many of Nagarjuna’s verses in the MMKwould lend themselves to such

Footnote 28 continued

Prajnaparamita genre. Nevertheless, he is surely the first thinker we know of to supply a systematic

presentation of the argument.

29 Eli Franco (personal communication), points to the debate here with the Sarvastivada theory of arising,

which sees utpādotpāda as an anulakṣaṇa of the lakṣaṇa utpāda.30 Notice that in Bronkhorst (1997, p. 31), it is maintained that Nagarjuna was clearly aware of the

distinction between problems concerning the liability of words and of sentences. There Bronkhorst

Argues that, for Nagarjuna: “the phenomenal world…is completely determined by words. It is also,

ultimately, unreal. Nagarjuna inherited this belief, and, it appears, extended it. He may have been the first

to draw sentences into the picture, besides individual words…”

Language, Understanding and Reality 347

123

an understanding, in line with prevalent interpretations of Nagarjuna’s thought in

modern scholarship.31

We will now look more deeply at Nagarjuna’s employment of the correspon-

dence principle, in light of the relevant discussion in the SS.

Śūnyatāsaptati 45–57

The set of verses from the SS I wish to analyze here question the accountability of

perception. These verses are interested in questions that echo Bronkhorst’s

correspondence principle, albeit they assume a finer definition of the principle than

the one Bronkhorst argues for. We join Nagarjuna in verse 45, after an eleven verse

reflection on Karma (verses 33–43), which in general can be seen as a summary of

MMK 17 (the chapter of the MMK that investigates karma).32 As noted above, the

doctrine of karma itself bears a natural affinity to the underlying logic of the

correspondence principle, since it is concerned with the subjective determination of

reality. As in the concluding verses of MMK 17 (31–33), the verses that deal with

karma in the SS end with a statement that karma is similar to an illusion procured by

an imaginary being. Then, in verse 44, Nagarjuna reminds his reader that the

Buddhas’ intention is hard to penetrate since he spoke on different occasions in the

language of existence, non-existence or both.33 This statement on the elusive, non-

determinate nature of ontological reality paves the way for a discussion of the

implausibility of perception; obviously, with no reality, the notion of perception is

raises suspicion. Here Nagarjuna wishes to point out that the intuitive understanding

of the perceptual process is incoherent. He begins with an inquiry into the relation

between matter (the elements, mahābhūta, ‘byung ba chen po) and perceptual form

(rūpa, gzugs):

45. If form is of the nature of the elements, that form does not arise from the

elements: It does not arise from itself—is that not so? Nor does it arise from

another, since it is not (other than them).

46. Fourness does not exist in one, and oneness does not exist in four. How

could form be established from a cause that is the four great elements?

gal te gzugs ni rang ‘byung bzhin / gzugs de ‘byung las ‘byung ma yin / ranglas ‘byung min ma yin nam / gzhan las kyang min de med phyir //

31 Another example would be the classic 1.1: na svato na parato na dvābhyāṃ nāpy ahetutaḥ / utpannājātu vidyante bhāvāḥ kva cana ke cana // (Not from themselves, not from another, not from both or

without a cause, are arisen things ever found, anywhere). Once again, the problem is not with the words of

a sentence but with analytical concepts such as arising, causation and “things.”32 Much of the SS can be seen as summary of the MMK or a paraphrase of its contents. Candrakırti

expresses a view akin to this in the opening of his commentary to the Yuktiṣaṣṭikā. See Scherrer-Schaub

(1991, p. 21 [Tibetan text], pp. 107–108 [translation]).33 SS 44. “(The scriptures) speak of ‘existence’, ‘non-existence’, and even ‘existence and non-existence.’

The speech by special intention of the Buddhas is difficult to penetrate.”

yod ces pa yod med ces yod / yod dang med ces de yang yod / sangs rgyas rnams kyi dgongspa yis / gsungs pa rtogs par sla ma yin //

348 E. Shulman

123

gcig la bzhi nyid yod min cing / bzhi la’ang gcig nyid yod min pas / gzugsni ‘byung ba chen po bzhi / rgyur byas nas grub ji ltar yod // 34

Perceptual form (rūpa, gzugs) is an unreliable category, since its relation with

matter (the four [great] elements), of which it is supposed to inform us and which it

supposedly represents, cannot be deciphered. Form cannot be equal to or different

than matter—if it was “matter”, it would not be “form”; but if it is not matter, than it

cannot be considered a faithful enough representation of it and the relation between

the two remains enigmatic. Specifically, Nagarjuna argues in verse 45, perceptual

form cannot arise from matter, if the two are thought to be identical (he appears not

to feel compelled to give serious attention to the possibility that perceptual form and

matter are absolutely distinct).35 Form cannot arise from the elements, he argues,

since if it were equal to the elements, it could not arise from itself, since such arising

would be pointless. But being equal to the elements, it could not arise from another

either. Verse 46 adds that there is a problem in defining the relation between the

“oneness” of perceptual form and the “fourness” of the great elements.

For the present purposes, the arguments themselves are less important than

the method they employ: We must note the importance of some version of the

correspondence principle for making sense of these two verses. The logic of the

argument hinges very clearly on an initial, implicit assumption that words must be

equal to things. Since this is so, and since the relation between perceptual form and

matter cannot be made sense of conceptually, there must be no realities that

correspond to the terms “perceptual form” and “matter/the great elements.” It makes

little sense to deny the concepts and sustain the belief in the reality they convey;

with suspicions raised against the concept, its corresponding reality loses its

credence as well—the point is not that we don’t understand perception, but rather

34 The verses in question have survived only in their Tibetan translations. For the text of the SS I rely on

the critical edition I prepared in Shulman (2009, Chap. 5), which I hope to make available in English in

the near future. In this edition I generally rely on the verses of the SS as found in the version of the text

without the svavṛtti that is attributed to Nagarjuna. For reasons I cannot discuss here, I find this version

more reliable, even though Lindtner (1986) and Tola and Dragonetti (1995) prefer the text embedded in

the svavṛtti. Komito (1987), who supplies the most thorough and reliable discussion of these issues [see

Chap. 3; but see also Lindtner (1986, p. 263) and Tola and Dragonneti (1995, p. 62)] works with the text

that is embedded in Candrakırti’s commentary, which is very close to the one of the verses without the

commentary. In my edition I at times accept emendations offered by Lindtner (1982, 1986). In any case,

the verses under discussion are not amongst those in which any significant changes in the meaning of the

verses between the two versions are recorded; rather, most of the divergences involve the order and

positioning of certain words or sentences. There are also minor differences between the texts preserved in

the different Tibetan Canons. I will note here only changes that would affect the discussion of the issues

that concern the present paper.35 It is a good question why Nagarjuna only deals directly with the problem of perceptual form and

matter being one, but ignores the obvious possibility that they are other; otherness is only discussed in the

last pāda with respect to the impossibility of form arising from matter under the conditions that they are

one. There could be a missing verse here, but more probably this decision reveals the way Nagarjuna

looks at the problem: Perceptual form and matter must be one, according to Nagarjuna’s logic, if “form”

is to supply reliable information about physical reality. But then form looses its psychological or mental

character and is exposed as a problematic category hovering between the physical and the mental, not

fully comfortable in either. Since the identity with matter is the only way perceptual form could be

relevant for what he would consider a strong enough theory of perception (even though it is deeply

problematic for other reasons), this is the option Nagarjuna targets.

Language, Understanding and Reality 349

123

that because that we don’t understand perception, there is no perception. A

schematic framing of the argument would say:

Axiom: Words = things

Argument: There is no coherent relation between the notions

of “perceptual form” and “matter”, its alleged cause

Conclusion: There is no reality to “perceptual form”

Without the assumption of the correspondence principle, the argument is almost

meaningless. The force of the claim hinges on the correspondence, or better—the

identity—between concepts and things. Notice especially the reliance on the

conceptual categories of “oneness” and “fourness” in verse 46, which emphasizes

the centrality of conceptual analysis to Nagarjuna’s account of perception and

reality; reality must correspond not only to words but to any conceptual tool used to

inquire into its workings. This itself already suggests that the correspondence

principle Nagarjuna had in mind, whether as a logical position or as an implicit

assumption, has to do with the degree to which reality conforms to conceptual

analysis; things must be susceptible to rational scrutiny.

Of primary importance is the conclusion drawn based on the correspondence

principle: When no theoretical definition of a concept is available, the reality it

conveys is undermined. The intimate connection between understanding and reality

is emphasized, and we learn that when something cannot be explained, when there is

no available conceptual design for it cannot be. Bronkhorst’s formulation of the

correspondence principle is immensely helpful in penetrating the arguments here,

which make little sense without it. Indeed, Nagarjuna now stretches his argument

deep into the roots of the correspondence principle:

47. If it is asked—from what sign (do you deduce that) since it (form) is not

grasped, it doesn’t exist?

Because it arises from causes and conditions, that sign cannot be,36 and is

unreasonable.

36 I avoid discussing this statement here, which is quite common in Nagarjuna, that relational existence is

equal to non-existence (without affirming a nihilistic non-existence); there is no such thing which relates.

I have discussed this issue in Shulman (2007 [2009], Sect. III), and it appears with exceptional clarity in

verses such as MMK 10.8–10.10, SS 64, Yuktiṣaṣṭikā 43–45, and Ratnāvalī 1.88. To the arguments

presented there—that one would need something independent in order to have it relate to something else

—we could add the recurrent claim in Nagarjuna (e.g. MMK 15.4–15.6) that things not only do not

possess svabhāva—self-nature, but that they do not possess parabhāva—“other-nature” as well. Stated

otherwise, the denial of both these terms amounts to the negation of both independent and dependent

existence. This connection between the terms svabhāva and especially parabhāva and relational existenceis made evident by VV 53–54 (although the meter and wording suggests that these verses may be corrupt

to some degree, their overall statement is quite clear). Discussing the question of the self-nature of

positive states (kuśala), Nagarjuna says:

53. If this positive (kuśala) self-nature of positive states arose in dependence, it would be an other-

nature of positive things. How could it be a self-nature?

yadi ca pratītya kuśalaḥ svabhāva utpadyate sa kuśalānām / dharmāṇāṃ parabhāvaḥ svabhāvaevaṃ kathaṃ bhavati //

350 E. Shulman

123

shin tu mi ’dzin phyir de med / rtags las zhe na rtags de’ang med / rgyu dangrkyen las skyes pa’i phyir / rtags med par yang mi rigs so //

This may be one of Nagarjuna’s most significant verses, stating explicitly that

“since it (form) is not grasped, it doesn’t exist”; the ability to understand something

is thought to be a condition of its existence; what cannot be understood does not

exist. This verse says not only that no reality could correspond to the words of a

sentence, or even to words or concepts in general. Rather, it argues that in order to

be considered real, things must possess the property of being understood.37

This last verse is significant in pointing to Nagarjuna’s awareness of at least some

version of the correspondence principle. But the correspondence he appears to have

in mind is different from the one Bronkhorst attributes to him, which involves the

technical correlative positioning of words of sentences vis-a-vis their objects. Here,

the correspondence is between the conceptual framing of a term, of its convincing

functioning within a theoretical framework, on the one hand, and between the

reality the terms and theory aim to convey on the other. Concepts and things

correspond to a degree that things must be able to be comprehended analytically in

order to be real. Understanding is a very part of existence.

Nagarjuna will now proceed with further arguments against the validity of

perception, which will later lead to broader arguments regarding the value of

knowledge:

48. If that mind grasps a form, it would grasp its own nature. Because it arises

from causes and conditions, it doesn’t.38 How could it perceive an unreal

form?

Footnote 36 continued

54. And if this self-nature of positive things arises without depending on something, there would

be no living of the holy-life.atha na pratītya kiṃcit svabhāva utpadyate sa kuśalānām / dharmāṇām evaṃ syād vāso nabrahmacaryasya //

Verse 53 is explicit in saying that if svabhāva arose in dependence it would be a parabhāva. In other

words—something that arises in dependence can only be a nature of another.

It is true that this negation of existence in dependence must be reconciled with Nagarjuna’s forceful

statement in favor of the truth of dependent-origination (pratītyasamutpāda), such as MMK 24.18, VV 22

and the vṛtti to VV 54 itself. Yet the common scholarly practice to ignore the many statements of

Nagarjuna against relational existence in light of verses such as these cannot be justified any longer. As I

see it, the initial level of empirical observation does in fact teach us that things are dependent. But, when

these same observations are subjected to the dialectic of emptiness, our initial perception is corrected and

we understand that there was nothing dependent in the first place. Pratītya-samutpāda can only be a

conventional truth.

37 Note that we can understand the shin tu mi ‘dzin (na saṃgṛhyate?)—(since) it is not grasped—in two

ways. Either it can refer to the act of understanding perception, as I read it here, or it can refer to the fact

that form itself is not perceived. Even if this second reading of the Tibetan is preferred, it must be

acknowledged that the reasoning behind the claim that form is not perceived is that it cannot be grasped in

the conceptual analysis of the working of perception conducted in these verses.38 We could also translate yod min pas as “it does not exist”, a translation that would work well with the

logic of the previous verse.

Language, Understanding and Reality 351

123

gal te blo des gzugs ’dzin na / rang gi rang bzhin la ’dzin ’gyur / rkyen lasskyes pas yod min pas / yang dag gzugs med ji ltar ’dzin //

This verse is styled according to an argument familiar from MMK 3.2,39 which

encapsulates much of Nagarjuna’s views regarding perception. The heart of the

position is that in order to relate to each other, the constituting elements of the

perceptual process must be able to maintain their identity and functions

irrespectively of the other elements that are part of the process. Thus, for instance,

sight must see in and of itself in order to be considered “sight”, and then to be able

to see form. If sight does not see itself, it cannot see form, a function that also

becomes superfluous since sight has already fulfilled its mission of seeing by seeing

itself. No less importantly, form arises from causes and conditions, and thus cannot

possess the one attribute Nagarjuna believes something must possess in order to be

considered truly real—it must have svabhāva, true nature or identity. We discover

that part of what svabhāva entails is that it must be knowable. Something with

svabhāva should be identifiable; a svabhāva can be known. Here we begin to sense

the metaphysical framework Nagarjuna is working in, one which is so problematic

to both the metaphysical and the epistemological realist—Nagarjuna demands

things to have a knowable essence in order to participate in a realist system.

Yet more problems regarding the functioning of perception are defined in the

following verses. Let us quickly scan through them. First, the temporal relation

between the mind and its objects is considered:

49. When it is said that simultaneousness of the mind and an arisen form is not

grasped—how will it (the mind) perceive a past or a future form?

ji skad bshad gzugs skyes pa’i blo’i / skad cig skad cig gis mi ’dzin / ’das dangma ’ongs pa gzugs kyang / de yis ji ltar rtogs par ’gyur //

This familiar argument—the mind and its object can be neither simultaneous nor

successive in time—is followed by another verse, which elaborates on the

problematic relation between oneness and diversity expressed in verse 46:

50. Since color and shape are never apart, they are not perceived40 apart from

each other. They are counted in one single form.

gang tshe nam yang kha dog dang / dbyibs dag tha dad nyid med pas / de dagtha dad ’dzin yod min / gzugs de gcig tu’ang grags pa yin41 //

39 MMK 3.2 Seeing does not see itself. What does not see itself—How will it see another?

svam ātmānam darśanaṃ hi tat tam eva na paśyati / na paśyati yad ātmānaṃ kathaṃ drakṣyati tatparān //

For an extensive discussion of this intriguing verse, see Shulman (2010, esp. pp. 384–390).40 Note that here, as opposed to verse 47 (see note 37 above) there is no ambiguity in the translation of

‘dzin (probably gṛhyate).41 I read yin against all textual versions I have consulted, including Lindtner (1986, p. 110), which all

record min. I do not see how the reading min could make any logical sense. The svavṛtti’s text for the lasttwo pādas (that read tha dad gcig tu ‘dzin pa med / de gnyis gzugs su grags phyir ro //) fit well with my

reading.

352 E. Shulman

123

Next appears an argument that is based on the Buddhist psychological identification

of a specific consciousness for each sense. The verse offers valuable insights:

51. Eye-consciousness is not in the eye; it is not in the object and is not

in-between the two. Since it depends on the eye and on form, it is a mistaken

conceptualization.

mig blo mig la yod min te / gzugs la yod min bar na med / gzugs dang mig labrten nas de / yongs su rtog pa log pa yin //

Here we find further evidence for the problems Nagarjuna finds with relational

existence—whatever depends has no nature of its own and is therefore defined as a

mistaken conceptualization (rtog pa log pa, mithyā vikalpa): What depends is

unreal. Any perceptual event is clearly dependent and must, therefore, be a mistaken

conceptualization. But the perception is not any more mistaken than the objects or

events it aims to communicate—both are dependent, and both are mistaken

conceptualizations.

This statement offers an interesting expression of the correspondence principle,

here seen as a structural, underlying identity between understanding and reality. The

verse bridges the gap between conceptualization and reality: What is dependent is a

mistaken conceptualization, a statement that is true both about the “thing-in-itself”

and about the concepts that define it, both about ontological reality and about the

theory that describes it. In order not to be “mistaken conceptualizations” things must

have a truly identifiable, knowable nature, which they possess of themselves,

without relying on anything else. There is a full correlation between how things

exist and the way they are conceived; both are equally erroneous constructions.

The next four verses continue with the arguments against the truth of perception.

We saw that verse 48 echoed MMK 3.2, which defines Nagarjuna’s most central

views on the fallacies of perception. MMK 3.2 is now nearly quoted in the next

verse, followed by a formal statement regarding the problems in defining the nature

of the eye and its objects:

52. If the eye does not see itself—how can it see form? Therefore, the eye and

form are selfless, and so are the other sense modalities.

53. The eye is empty of own-nature (svabhāva). It is also empty of other-

nature (parabhāva). Form too is empty in the same way, and so are the rest of

the sense modalities.

gal te mig bdag mi42 mthong na / des gzugs mthong bar ji ltar ‘gyur / de phyirmig dang gzugs bdag med / skye mched lhag ma’ang de bzhin no //

mig ni rang bdag nyid kyis stong / de ni gzhan bdag gis kyang stong / gzugskyang de bzhin stong pa ste / skye mched lhag ma’ang de bzhin no //

Since the other senses have been mentioned, Nagarjuna continues to an argument

regarding the relation between the senses, familiar from MMK 9.8,9:

42 Following Lindtner (1986, p. 110).

Language, Understanding and Reality 353

123

54. When, one (sense) is in a moment of contact (with its object), the others

are empty. Neither does the empty relate to the non-empty, nor does the non-

empty relate to the empty.

gang tshe gcig reg lhan cig ‘gyur / de tshe gzhan rnams stong pa nyid / stongpa’ang mi stong mi bsten la / mi stong pa yang stong43 mi brten //

This is a rather difficult verse,44 and as it stands a less convincing one as well.

Nagarjuna seems to think that if one sense meets its object, than the other senses

cannot do so as well. Why this should be true I am uncertain, but the argument could

make better sense of it is constructed as saying that if a sense organ is fully occupied

with its own object, then it cannot relate to the other senses. Or, otherwise, the point

may be more radical, saying that if an object exists for one sense, it cannot exist for

the others.45 In any case, the verse appears to be based once again on the problem of

connecting between oneness and multiplicity.

The next verse is clearer, offering a reflection on the same type of problems

defined in verses 48 and 52, following MMK 3.2. When there is no self-identity for

the elements of the perceptual process, they cannot connect to each other:

55. They have no non-abiding essence, and therefore the three (sense, object

and consciousness) do not come together. Since there is no contact by way of

essence, there is no feeling (vedanā).

ngo bo mi gnas yod min pas / gsum ‘dus pa yod ma yin no / de bdag46 nyidkyis47 reg med pas / de tshe tshor ba yod ma yin //

This verse begins to move toward a generalization of the conclusions reached in

this section of the SS. Nagarjuna argues that the three elements of the perceptual

process—sense, object and consciousness—cannot connect, since all are devoid of a

permanent nature that would maintain its identity and thus allow it to connect to the

other elements of the process. The precise problem here is that the three elements do

not have an impermanent nature, a “non-abiding essence” (ngo bo mi gnas), whichwould offer a way to define its identity in light of the dynamics of change. The

problem essentially concerns the difficulty to reconcile identity with impermanence,

since for impermanence itself to be accountable for, we would need a permanent

thing that could be characterized as impermanent.48 Relying on this statement

43 Following Lindtner (1986, p. 112).44 There are also significant divergences between the different versions of the text regarding the last two

pādas. Most significantly, for the end of the verse, the svavṛtti gives stong pa min, while the version of thekārikā alone gives brten (D: rten) mi brten. Also, near the end of the third pāda, D reads ston for bsten.45 The idealistic overtones are not alien to the SS. Indeed, verses 59–61, which follow the discussion of

perception we are analyzing here, assign a strong creative power to conceptuality. See Shulman (2007

[2009], pp. 164–166).46 N,P: dag.47 N,P,Lin: kyi.48 Nagarjuna will refer to this problem explicitly in verse 58:

If everything is impermanent, and impermanence is also impermanent, how will there be

permanent or impermanent things?

354 E. Shulman

123

regarding the senses, Nagarjuna further argues that the processes that follow

(perceptual) contact according to the 12 links may not take place as well.49

Verses 56 and 57 sum up Nagarjuna’s analysis of perception by applying the

results of his inquiry to knowledge in general. Verse 56 will speak of consciousness

(rnam par shes pa, vijñāna), while verse 57 will move on to speak of knowledge

(rnam shes, jñāna):

56. Since consciousness appears depending on the inner and outer sense

modalities, consciousness is similar to a mirage and an illusion, empty.

57. To a knowledge that appears depending on an object of knowledge, there

is no object of knowledge. Since there are no knowledge and object of

knowledge, there is no knower.

nang dang phyi yi skye mched la / brten nas rnam par shes pa ’byung / de ltabas na rnam shes ni50 / smig rgyu sgyu ma bzhin du stong //

rnam shes shes bya la brten nas / ‘byung la shes bya yod ma yin / shes bya shespa med pa’i phyir / de phyir shes pa po nyid med //

Verse 56 makes a dramatic statement about consciousness—all that consciousness

conveys is an illusion. The reasoning supplied, once again, is that consciousness is

dependant, and therefore, like anything else, is as empty as a mirage. I read this verse

not merely as providing a tight definition of the unreliability of the concept of

consciousness. Rather, every moment of consciousness is portrayed as illusory and

empty.51 This statement turns into a broader definition regarding the very prospect of

knowledge in the following verse, where Nagarjuna argues that the dependence

between knowledge and knowable nullifies both.

This last verse comes close to offering a conscious reflection on the

correspondence principle, applying it self-referentially to the understanding of

knowledge; knowledge cannot be known, and hence cannot be true. With no viable

definition of knowledge, whatever it conveys must be false. Nagarjuna’s employ-

ment of the correspondence principle, now defined as an structural relation between

knowledge and reality, insists not only that there is no objective truth, but also that

there is no reliable subjective one either. These two cannot be distinguished: There

is no reality; there is, however, mistaken understanding. The mistake, notice, is not

in the failure of knowledge to faithfully reflect truth, but in the attempt to present

truth in the first place.

Footnote 48 continued

thams cad mi rtag yang na ni / mi rtag pa yang rtag pa med / dngos po rtag dang mi rtag nyid /

‘gyur na de lta ga la yod //

For further discussion, see Shulman (2008).

49 For more on this theme, see SS 11–12 and 62–63.50 Lindtner (1986, p. 112) suggests med instead of ni.51 If we accept Lindtner’s (1986, p. 112) emendation from ni to med at the end of the third pada (see the

previous note), taken from the svavṛtti text, the verse makes a yet more forceful statement that fits well

with the reading developed here: “therefore, consciousness does not exist.”

Language, Understanding and Reality 355

123

The correspondence principle has helped us see how radical Nagarjuna’s critique

of knowledge actually is. Since knowledge cannot be theoretically conceptualized,

all knowledge events are seen as illusory and empty. We realize that at the deep core

of Nagarjuna’s empty vision, there is a necessary, crucial connection, which

constantly breaks, between two elements—knowledge and svabhāva. The connec-

tion breaks because there is no svabhāva, and thus nothing can be truly known and

the aspiration for knowledge is reflected as no more than an act of grasping. If there

were a svabhāva, it could be known, and the construction of our worlds of reality

and of consciousness would be legitimized. Without it, however, all is empty. The

ultimate failure of this correspondence to frame the structure of a real world raises

essential metaphysical problems, which we now turn to investigate in light of

Nagarjuna’s Vigrahavyāvartanī.

The Vigrahavyavartanı and Nāgārjuna’s Implicit Assumptions

It is interesting to read the arguments put forward by Nagarjuna in his

Vigrahavyāvartanī (VV) in light of the understanding of the correspondence

principle and of its significance for Nagarjuna reached in the discussion of SS

45–57.52 In the VV Nagarjuna defends himself against arguments put forward by an

adversary, usually thought to voice the views of the Nyaya school. Although the

understanding of the opponent as a Naiyayika, at least as the Nyaya is reflected in

the Nyāya-sūtras, has recently been questioned,53 it is clear that his positions

endorse a realist view that is perfectly in tune with the metaphysical inclinations of

Nyaya philosophy. As we shall see, Bronkhorst’s discussion of the correspondence

principle is helpful in reconstructing what this debate is really about; we find here a

clash between two of the most basic Indian metaphysical views—we could call

them Realism and Anti-realism,54 or Realism and Emptiness. The debate centers on

the question regarding the existence of svabhāva. Both positions embrace the

correspondence principle, but when this acceptance is coupled with the affirmation

(for the Naiyayika) or the denial (for the Madhyamika) of svabhāva, diametrically

contradictory metaphysical positions are entailed.

52 Since the VV has received ample attention in modern scholarship, I will focus only on what is of

primary concern for the present discussion. For the text and translation, see Bhattacarya et al. (1978). For

recent analyses of its contexts, see Westerhoff (2009, Chap. 9) and the detailed response in Oetke (2011,

pp. 288–325).53 Recently, Franco (2004) and Oetke (2011, pp. 288–293) have shed some doubt on this generally

accepted attribution. See also Johnston and Kunst (1978, pp. 39–40). It is also interesting to note the

suggestion raised in Tola and Dragonetti (1998, p. 154) that the first 20 verses of the VV were originally

an independent text for which Nagarjuna composed a refutation.54 I use the term Anti-realism in the Madhyamaka context in a broader sense the one employed by

Siderits (1988). For Siderits, Anti-realism in Madhyamaka is mainly epistemological, as it goes against

the correspondence theory of knowledge. Here Anti-realism is both metaphysical and epistemological,

while the epistemological meanings are derived from the metaphysical ones.

356 E. Shulman

123

The problem that opens the VV is an accusation leveled at Nagarjuna by his

opponent that he maintains a controversial position (or statement, vacana) that

denies svabhāva and upholds emptiness, while the same position itself possesses

svabhāva and is non-empty; if Nagarjuna’s words are to refute anything, the

opponent argues, they must be real in themselves.55 Nagarjuna’s response consists

mainly in stating that his statement, like anything else, lacks svabhāva, thus

nullifying the main thrust of his opponent’s criticism.56 He claims that his words are

similar to the illusions created by an illusory person, refuting the existence of the

illusory person he himself created.57 Or, he says, his words are like a statement that

Devadatta is not in the house when this is in fact the true state of things—such a

statement would not remove Devadatta from the house but only indicate his

55 VV 1. If there is no self-nature of all things whatsoever, your statement lacks self-nature and is

incapable of refuting self-nature.

VV 2. Moreover, given that this pronouncement possesses self-nature, your previous claim (pratijñā,that all things lack svabhāva) is abandoned; you must state the reason of the distinction for this

discordance.

sarveṣāṃ bhāvānāṃ sarvatra na vidyate svabhāvaś cet / tvad vacanam asvabhāvaṃ nanirvartayituṃ svabhāvam alam //atha sasvabhāvam etad vākyaṃ pūrvā hatā pratijñā te / vaiṣamikatvaṃ tasmin viśeṣahetuś cavaktavyaḥ //

56 The kernel of Nagarjuna’s position in the VV is articulated in the opening of his response to his

adversary in verse 21, quoted here together with its vṛtti:

If my statement does not exist in the assemblage of cause and conditions, or apart from them, then

the emptiness of things is established, because of the lack of self-nature.

If my speech does not exist in the cause of the great elements, together or separately, and is not in

the conditions related to the activity of the chest, throat, lips, tongue, teeth, palate, nose, head, nor

in the two of them together, and is not independently of the assemblage of cause and conditions,

then it is devoid of self-nature, and due to being devoid of self-nature, it is empty, and surely

emptiness of this speech of mine in established, due to lacking self-nature. And just as this speech

of mine is empty due to its being devoid of self-nature, so too are all things empty due to being

devoid of self-nature.

hetupratyayasāmagryāṃ ca pṛṭhak cāpi madvaco na yadi / nanu śūnyatvaṃ siddhaṃ bhāvānāmasvabhāvatvat //

yadi madvaco hetau nāsti mahābhūteṣu saṃprayukteṣu viprayukteṣu vā pratyayeṣu nāstyuraḥkaṇṭhauṣṭhajihvādantamūlatālunāsikāmūrdhaprabhṛtiṣu yatneṣu nobhayasāmagryām asti het-upratyayasāmagrīvinirmuktaṃ pṛthag eva ca nāsti tasmān niḥsvabhāvam niḥsvabhāvatvācchūnyam nanu śūnyatvaṃ siddhaṃ niḥsvabhāvatvād asya madīyavacasaḥ / yathā caitanmadvacanaṃ niḥsvabhāvatvāc chūnyaṃ tathā sarvabhāvā api niḥsvabhāvatvāc chūnyā iti /…

For the text of the VV I rely on the edition in Bhattacarya et al. (1978).57 VV 23. This refutation (of the svabhāva of things conducted by Nagarjuna) is like a conjured man who

would refute another conjured man, or an illusory man who would refute (another illusory man) who was

formed through his (the first man’s) own powers of illusion.

nirmitako nirmitakaṃ māyāpuruṣaḥ svamāyayā ṣṛṣṭam / pratiṣedhyeta yadvat pratiṣedho ‘yamtathāiva syāt //

The translation here follows the vṛtti. A straightforward translation of the verse could also give:

“This refutation (of the svabhāva of things conducted by Nagarjuna) is like an illusory conjured man

who would refute another conjured man, formed through his own powers of illusion.”

Language, Understanding and Reality 357

123

absence.58 For Nagarjuna, it is crucial to point to the understanding that his words

should not be thought of as a reliable point of reference in the empty world.

Ontologically, they are just like anything else—fully empty. Therefore his position

is consistent all the way through—there is nothing real whatsoever, including this

statement itself, and Nagarjuna can persist in maintaining, in the controversial verse

29, that he has no thesis (pratijñā);59 given the ubiquitous lack of svabhāva, he canmake no argument when there is nothing to argue about.60 Images of illusions and

such phantasmagoria are central to the picture of the world he portrays, and his

argument is nothing but one more such illusion.

B. K. Matilal has been a leading exponent of the reading of the VV as concerned

with the (im)possibility of metaphysical assertion, as an attempt made by Nagarjuna

to specify the nature of his empty position regarding emptiness. Matilal sees

Nagarjuna as maintaining a “non-committal attitude in ontology”,61 and as directing

his arguments against the relevance of any metaphysical theory of reality—“It is

neither proper nor strictly justifiable to regard any particular metaphysical system as

absolutely valid.”62 He speaks of Nagarjuna as seeing the nature of the phenomenal

58 See the vṛtti for verse 64:

Regarding what you claim—“the refutation of something non-existent is established even without

words, so what does your statement ‘all things are devoid of self-nature’ do?”—we say:

This statement ‘all things are devoid of self-nature’ does not make all things devoid of self-nature.

Rather, when there is (already) no self-nature, it makes known that things are devoid of self-nature.

This is like if somebody would say, when Devadatta is not in the house “Devadatta is in the

house”; regarding this someone would reply “no, he is not.” This statement does not create the

non-existence of Devadatta (in the house), but only makes known the non-being of Devadatta in

the house. In the same way, this statement “there is no self-nature of things” does not make the

lack of self-nature of all things, but rather makes known the non-existence of self-nature regarding

all things…

yac ca bhavān bravīti ṛte ‘pi vacanād asataḥ pratiṣedhaḥ prasiddhaḥ tatra kiṃ niḥsvabhāvāhsarvabhāvāh ity etat tvadvacanaṃ karotīti, atra brūmaḥ / niḥsvabhāvāḥ sarvabhāvā ity etat khaluvacanaṃ na niḥsvabhāvān eva sarvabhāvān karoti / kiṃ tv asati svabhāve bhāvā niḥsvabhāvā itijñāpayati / tad yathā kaścid brūyād avidyamānagṛhe devadatte ‘sti gṛhe devadatta iti / tatraianaṃkaścit pratibrūyān nāstīti / na tadvacanaṃ devadattasyāsadbhāvaṃ karoti kiṃ tu jñāpayatikevalam asaṃbhavaṃ gṛhe devadattasya / tadvan nāsti svabhāvo bhāvānām ity etad vacanaṃ nabhāvānāṃ niḥsvabhāvatvaṃ karoti kiṃtu sarvabhāveṣu svabhāvasyābhāvaṃ jñāpayati / …

59 If I had a certain thesis, I would have that very fault. Yet I have no thesis, and hence I surely have no

fault.

yadi kācana pratijñā syān me tata eṣa me bhaved doṣaḥ / nāsti ca mama pratijñā tasmān naivāstime doṣaḥ //

60 Should this statement itself, however, count as a “view” or a “thesis”? This question has naturally

concerned many Madhyamaka scholars (e.g. Ruegg 2000; Oetke 2003). Personally, I believe that there is

no denying that on some conventional level this must amount to a philosophical position. This reading of

VV 29, which emphasizes the “conventional truth” of the Madhyamaka position, is supported also by the

reference to such a level of truth in the previous verse (28 cd): “(And) we do not speak without having

accepted the conventional” (samvyavahāraṃ ca vayaṃ nānabhyupagamya kathayāmaḥ).61 Matilal (1971, p. 147; 2008 [1985], p. 317).62 Matilal (2008 [1985], p. 318).

358 E. Shulman

123

world as indeterminate; all predication in relation to it is inapplicable.63 Matilal thus

regards Nagarjuna as maintaining not a skeptical philosophical position, but more of

a basic skeptical attitude.64 For Matilal, the true challenge in understanding the VV

relates to the attempt to figure out how Nagarjuna can at the very same time make

assertions and maintain that he makes no assertions.65

Yet there is a deeper question at stake in the debate between Nagarjuna and his

realist opponent than the one regarding the status of Nagarjuna’s statements. The

reading of the VV as concerned with the nature of philosophical assertion neglects

important aspects of the discussion and ignores a more foundational problem that is

at the heart of the matter. The status of Nagarjuna’s pronouncements is only a

particular case of a far more fundamental question—this is a struggle with grave

metaphysical consequences over the reality or unreality of svabhāva.66 The

opponent is attempting to make conceptual space for svabhāva so that he will be

able to continue maintaining not only that knowledge is truthful, but that the world

is real as well. He must do so since Nagarjuna’s denial of svabhāva is nothing less

than an ontological theory. It is true that Nagarjuna’s ontology leaves many question

marks—indeed his ontology is a question mark—what is there with no svabhāva,but while there still remains something similar to an illusion? Nevertheless, it is

clear that for both Nagarjuna and his opponent, so long as there is no svabhāva,there is no real world either, and no reliable knowledge is available. With no

objective reality and no true knowledge, there of course is no room for furtherauthoritative metaphysical assertion; although something certainly is, it can never

be real, and it is at best a convention or a possibility. The denial of metaphysical

assertion, however, is only the aftermath of the destruction of all realistic ontology

and epistemology.

The positions of Nagarjuna and his opponent agree on two points: First, that there

must be a full correspondence between word and thing. Second, that given this

correspondence, a successful refutation of svabhāva denies any possibility for the

objective existence of reality. For this reason it is imperative for the opponent, who

is familiar with Nagarjuna’s dialectic that deduces the lack of svabhāva from

relational existence, to re-introduce svabhāva so that his world will be saved. The

opponent is not merely toying with Nagarjuna regarding the nature of his argument;

this is not merely a philosophical game. Rather, the adversary is attempting to

uphold the reality of the objective world, and thus to allow for the possibility of

knowledge. These are not questions about the nature of philosophy, by about the

very nature of human life. Quite surprisingly, though, Nagarjuna ignores the crux of

his opponent’s critique and avoids dealing with the main problems he raises.

63 Matilal (1971, pp. 155–159). Notice, however, the problematic consequences this position regarding

the indeterminate nature of reality has for a Madhyamaka system; some reality must remain, it can only

not be conceived of. This indeed leads to the inevitable introduction of a mystical reality the adept

Madhyamika has some privileged access to, a form of absolute truth that is fundamentally resilient in face

of the Madhyamaka dialectic. Indeed, Matilal inclines to such a view of a mystical truth beyond the world

of conventions in a number of his writings (e.g. 1971, pp. 166–167).64 Matilal (1982, p. 63).65 Matilal (1990, pp. 153–155).66 Here I follow and expand on points made in Oetke (2011, pp. 309–325).

Language, Understanding and Reality 359

123

Committed to his ontological paradigm of emptiness, he does not allow the

opponent’s considerations to enter his system, since his method of analysis cannot

accommodate them. We will soon see why.

In order to appreciate the urgency of the opponent’s attack, we shall concentrate

on verses 9 and 11–13:

9. (And) if there is no self-nature of things, then even the very term “lack of

self-nature” would not be; for there is no term that has no object.

yadi ca na bhavet svabhāvo dharmāṇāṃ niḥsvabhāva ity eva /nāmāpi bhaven naivaṃ nāma hi nirvastukaṃ nāsti //

The main point the opponent makes is that the lack of svabhāva must be shown in

respect to something. This something would imply a svabhāva, as verses 11–12

will make clear. In this verse the opponent reveals the basic logic that underlies

his position, which enforces the correspondence principle to the extreme: “for

there is no term that has no object” (nāma hi nirvastukaṃ nāsti). The opponent

voices what we would take to be the normal expression of the correspondence

principle—that words and things are wholly commensurate. His take of this

principle demands that there be both word and thing and that the correlation

between them to be self-evident; this is the true realism of his position. This clear-

cut relation between word and thing further demands, as the Naiyayika sees it,

that if Nagarjuna wishes to refute anything, there must be something to refute. Theinevitable conclusion suggested by this line of thought is that what is to be refuted

must exist.

It is of pivotal importance to understand what precisely troubles the opponent.

The opponent is fervently trying to show that there must be svabhāva and that this

svabhāva reflects existent things.67 He desires to uphold his realist ontology—and

not only his realist epistemology—in face of Nagarjuna’s argument for emptiness.

Hence he must show that there is svabhāva, that something objective is there, andthat this is demanded even of a theory of emptiness. We see that the first verses of

the text, which argue that Nagarjuna’s statements must possess svabhāva if they are

to refute anything,68 were only a first item in a series of attempts to re-establish

the validity of svabhāva. The initial attempt targets the subjective side of the

correspondence, Nagarjuna’s statements; the next one moves to discuss the

objective side of the equation, the refuted objects.

The ontological meaning of the opponent’s arguments becomes evident in verses

11–12:

11. This refutation must be of an existent (thing), as in “there is no pot in the

house.” Therefore, this refutation of yours is seen to be of something that

possesses svabhāva.

67 Note the opponent’s willingness to contemplate even the idea that there is a svabhāva that exists apart

from things (dharmair vinā) in verse 10.68 See note 55 above for verses 1–2.

360 E. Shulman

123

12. Moreover, if this svabhāva does not exist, what exactly is refuted by your

statement? For the refutation of something non-existent is established without

saying anything.

sata eva pratiṣedho nāsti ghaṭo geha ity ayaṃ yasmāt /dṛṣṭaḥ pratiṣedho ‘yaṃ sataḥ svabhāvasya te tasmāt //

atha nāsti sa svabhāvaḥ kiṃ nu pratiṣidhyate tvayānena /vacanenarte vacanāt pratiṣedhaḥ sidhyate hy asataḥ //

The opponent is a dedicated realist; the world he envisages is furnished with real

things one can know. It is only because that there are such things that one can argue

against them. Here he claims that something must exist in order to be refuted—if

there is no svabhāva to the refuted thing, the refutation cannot work. When

something is refuted, some-thing is in fact refuted. If there was no-thing there, there

would be no need to refute it. Otherwise, the opponent argues, Nagarjuna is refuting

only illusions:

13. For just like a mirage that is mistakenly perceived by foolish people, so

would be the mistaken perception of your non-existent thing that is being

refuted.

bālānām iva mithyā mṛgatṛṣṇāyāṃ yathājalagrāhaḥ /evaṃ mithyāgrāhaḥ syāt te pratiṣedhyato hi asataḥ //

The opponent is thoroughly devoted to his intuition regarding the correspondence

principle: There are things and there are words, and a word must relate to a thing.

With this in mind, he is afraid that if Nagarjuna’s refutation of svabhāva succeeds,

there will be nothing on the side of the object to bear the burden of objective

existence and all will turn into chaotic illusion. He therefore ardently battles this

metaphysical option that would completely demolish the real world he believes in.

The awareness of the correspondence principle allows us to realize what exactly is

at stake—it is the truthful relation between (subjective) word and (existent,

objective) thing.

We must grant that the opponent has made a good point—Nagarjuna does need to

specify what it is that lacks svabhāva; how can he make any argument if there is

nothing there? It is quite legitimate to demand of Nagarjuna an answer to this

problem, which should trouble him even under an anti-realist or an idealist

metaphysic. But somewhat disappointingly—we could have benefited from

Nagarjuna’s answer to what may be the most pressing problem in Madhyamaka—

Nagarjuna dodges the question; his response to these verses does not meet his

opponent’s query, but rather continues to tread a path on which his opponent cannot

join him:

57. If someone would say that a term is truly existent (sadbhūta), you could

surely answer “it possesses svabhāva.” We however do not say this.

58. This “non-existent name”, how can it be of an existent or a non-existent

thing? For in both cases, whether it is existent or non-existent, your position is

abandoned.

Language, Understanding and Reality 361

123

yaḥ sadbhūtaṃ nāmātra brūyāt sasvabhāva ity evam / bhavatā prativaktavyonāma brūmaś ca na vayaṃ tat //

nāmāsad iti yad idaṃ tat kiṃ nu sato bhavaty utāpy asataḥ / yadi hi sato yadyasato dvidhāpi te hīyate vādaḥ //

Nagarjuna’s reply focuses only on the nature of the term referring to the empty

object. He does not specify what it is that the terms refer to, what it is he claims to

be empty; he will give not even a minimal clue to how he is able to make his

argument or why it is that the need for it arises. He does supply an answer, of

course, a convincing one at least to a degree, stating that the term the opponent is

grasping at, which he hopes will point to the existence of a svabhāva that it denotes,

is itself non-existent. Nagarjuna focuses on the non-existence of the term which is to

represent the existence or the lack of svabhāva, feeling no apparent need to answer

the deeper point that was raised by his opponent.

More troubling still is Nagarjuna’s next argument:69

61. If the refutation is of an existent, then this emptiness is established. For

you, sir, refute the lack of self-nature of things.

62. Or, if you refute emptiness, and there is no emptiness, then your statement

“a refutation is of something existent” is abandoned.

sata eva pratiṣedho yadi śūnyatvaṃ nanu prasiddham idam / pratiṣedhayate hibhāvān bhāvānāṃ niḥsvabhāvatvam //

pratiṣedhayase ‘tha tvaṃ śūnyatvaṃ tac ca nāsti śūnyatvaṃ / pratiṣedhaḥ sataiti te nanv eṣa vihīyate vādaḥ //

The opponent tries to avoid the consequence of niḥsvabhāva by stating that the

refutation is of something existent. But by this he supposedly affirms the refutation,

as well as its object—the lack of self-nature, and therefore supports the position of

emptiness (verse 61). Or, if the opponent wishes to refute a non-existent emptiness,

his position that a refutation is of something existent is abandoned (62). With this

Nagarjuna believes that he has checked his opponent’s attack, even though he has

not answered the central claim he made.

Nagarjuna’s response in verses 61–62 is logically feasible, if not very

convincing, and it responds, with dry technical arguments, to the specific statements

made by his adversary. Nonetheless, his answer is disappointing, since he has not

seriously addressed the question regarding what he refutes. His refutation must start

somewhere, and even if this is only a “conventional truth”,70 there is still some truth

there he must account for. Rather than facing the question, Nagarjuna bypasses it

while playfully occupying our attention with his logical fiddle. He then swiftly

69 For brevity’s sake I skip verses 59 and 60—I side with Oetke (2011, pp. 322–323) that verse 59 does

not add important content to verses 57 and 58; Verse 60 answers the relatively tangential position raised

by verse 10 that there is a svabhāva apart from things.70 The VV employs the concept of vyavahāra. See the quote from verse 28 in note 60 above, as well as

VV 55.

362 E. Shulman

123

moves head on to his metaphysical denial svabhāva, repeating that there is nothing

to refute and that all is illusion-like.71

For some reason, Nagarjuna is untouched by his opponent’s main argument in

VV 9, 11–12 (with verse 13 he actually doesn’t have much trouble). He feels that he

can conclude the text without offering a response to the pressing question of what is

it that he refutes. Should we not expect more of his customary so acute and rigorous

thinking? Here, once again, the correspondence principle comes to our aid in our

effort to understand Nagarjuna’s maeuvers. Both Nagarjuna and the Naiyayika

accept the correspondence principle, but it directs them in diametrically opposite

philosophical vectors. What for the Naiyatika points toward existence, to Nagarjuna

speaks of illusions and conceptual fallacies. The opponent understands the

correspondence principle to imply a distinction and a complete match between

word and thing, according to a realist system. Nagarjuna understands the principle

to imply the arch-opposite idea: It defines an absolute relation between a

misleading, mistaken word and a non-existent, illusory thing. For him, there are,

essentially, no words and things; both are fantastical. Thus, in order to conceive of

the refuted thing, from which we detect a sure scent of existence, Nagarjuna has to

step outside of his system and employ and understanding of the correspondence

principle he finds absurd; he would have to posit something that is non-empty. But

this he cannot do, since his system is all-pervading, and for him all abides by the

logic of the empty. What was empty in the first place? Nagarjuna can only answer

this question from within his analysis, which proves all to be empty, and to which no

independent thing ever appears, and which sees all terms as mistaken conceptu-

alizations. At the same moment the “refuted thing” is conceptualized, it is consumed

by Nagarjuna’s logic, which denies its svabhāva; nothing can be conceived of that

transcends the empty. He is thus content to supply an artificial and dismissive

argument against the terms his opponent uses to structure his claim, without offering

a true rebuttal of the main theme of his argument.

71VV 63. I refute nothing and there is nothing to refute. Therefore, you are projecting when

you say “you refute.”

pratiṣedhayāmi nāhaṃ kiṃcit pratiṣedhyam asti na ca kiṃcit / tasmāt pratiṣedhayasīty adhilayaeṣa tvayā kriyate //

64. And regarding your statement that “the establishment of a statement that refuteswhat does not

exist is done without words”, the claim “not existent” (only) makes known, it does not destroy it

(the refuted thing). (See also the text and translation of the commentary in note 58 above).

yac cāharte vacanād asataḥ pratiṣhedavacanasiddhir iti / atra jñāpayate vāg asad iti tan napratinihanti //65. Regarding the example of the mirage that was introduced by you, there is a great

deliberation. Here too listen to the correct deduction of how this example makes sense.

mṛgatṛṣdṛṣṭānte yaḥ punar uktas tvayā mahāṃścaraḥ / tatrāpi nirṇayaṃ śṛṇu yathā sa dṛṣṭāntaupapannaḥ //

66. If this perception would exist by way of svabhāva, it would not exist in dependence.

A perception that exists in dependence, well that, verily, is emptiness.

sa yadi svabhāvataḥ syād grāho na syāt pratītya saṃbhūtaḥ / yaś ca pratītya bhavati grāhonanu śūnyatā saiva //67. If a perception existed by way of svabhāva—who could refute this perception? This rule

applies to all other things as well. Therefore there is no fault.

yadi ca svabhāvataḥ syād grāhaḥ kas taṃ nivartayed grāham / śeṣeṣv apy eṣa vidhis tasmād eṣo‘nupālambhaḥ //

Language, Understanding and Reality 363

123

The crucial point to realize here is that Nagarjuna’s answer focuses on the nature

of philosophical assertion, not because this is the main issue at hand. Rather, he

speaks of the “terms” (nāma) and of the nature of refutation (pratiṣedha) since this isall there is to speak about. Under his ontology, there are not objects that philosophical

theory can refer to. All has succumbed to “mistaken conceptualizations.”

Both Nagarjuna and the Nyaya advocate are at ease when they remain entrenched

in their own metaphysical positions: Neither can or will make any concessions even

when their adversary is at his best. They can debate with each other, since they both

make the implicit assumption of the all-pervasive correspondence principle; both go

by nāma hi nirvastukaṃ nāsti. (“For there is no term that has no [corresponding]

object.”) But they never really notice what this means for the other—for the

Naiyayika, Realism, for Nagarjuna emptiness of both word and thing—and can only

debate according to their own premises; their metaphysical inclinations do not allow

them to listen to each other, especially when the point becomes most meaningful.

When the essential questions surface, they turn deaf.72

It is important to connect these understandings regarding the metaphysical aspect

of the refutation of svabhāva in the VV to the well-known denial of pramāṇaNagarjuna effects in this text. As a central part of his defense strategy, Nagarjuna

introduces a set of arguments that target the concept of pramāṇa, the “means of

valid cognition” that figure prominently in the epistemologies of the classical Indian

schools of thought. Here Nagarjuna conducts a discussion quite similar to the one

we encountered in SS 45–57; the arguments are different, but the conclusions are

very similar—there is no reliable knowledge. The reasoning is ultimately very

similar as well—since there is no way to define the workings of knowledge, every

form of knowledge is suspect. Seen in this light, there is no intrinsic value to the

discussion of pramāṇas here, it is only another way to speak of the unreliability of

knowledge. The pramāṇas corresponds to the eye or eye-consciousness in verses

such as SS 51–52, while the prameyas correspond to the “form” of SS 46–48. None

of these—the faculty of vision, perceptual form, the pramāṇas and prameyas—can

be identified by the analysis and can therefore not be real.

It thus seems clear that there is no real epistemology in the VV, and that, as Oetke

(2011, p. 324) claims, the heart of the text expresses a metaphysical position.73

Indeed, this part of the text can be read as an anti-epistemological treatise, which

72 See Taber (1998, p. 237), who speaks of Nagarjuna’s reader not as being convinced by his rigorous logic,

but being invited to perform a “paradigm shift” with which he will “leap beyond ordinary experience.”73 Oetke heavily criticizes Westerhoff’s (2009) reading of Nagarjuna’s arguments in the VV as a

“semantic pronouncement.” Although I share with Oetke the emphasis on the metaphysical aspect of the

VV, I believe he is overly critical of Westerhoff (2009), whose definition of Nagarjuna’s “semantic”

position possesses many metaphysical aspects. This is very clear when Westerhoff says, in the passage

quoted by Oetke on page 288 of his article, that the “standard picture” he thinks Nagarjuna argues against

“assumes that the world of referents is endowed with a mind independent structure and that our language

manages to latch onto the world not just by force of convention, but by the existence of some objectively

existent structural similarity between language and the world” (quoted from Westerhoff 2009, p. 17). For

another example of the strong metaphysical inclination of Westerhoff’s Madhyamaka reading, see

Westerhoff (2011, p. 189).

364 E. Shulman

123

argues straightforwardly that knowledge is unfeasible. Although this statement must

be nuanced so as to be able to include at-least the partial viability of conventional

practical language,74 the main idea conveyed by this portion of the VV is that there

is in fact nothing to know. As the opening verse of this section states:

30. If I could apprehend something by way of the objects of perception and of

the other (pramāṇas), I could affirm or deny it. Since there is no such thing,

I cannot be found fault with.

yadi kiṃcid upalabheyaṃ pravartayeyaṃ nirvartayeyaṃ vā / pratyakṣādibhirarthais tadabhāvān me ‘nupalāmbhaḥ //

This position appears to be at odds with later traditional Madhayamka attempts to

reconcile the accountability of pramāṇas with the ontological theory of emptiness.75

But then, quite naturally, a problem akin to the one raised by the adversary in the

VV re-surfaces—if there is nothing real, then what are we talking about? If the true

theory of emptiness is accepted, one in which there is in fact no svabhāva and all

things are empty—what is it that is empty? This problem becomes especially

pertinent when we resist the “Madhyamaka-wholesale” view according to which

emptiness is equal only to relational existence. Rather, as the VV makes clear time

and again, relational existence is only the first step toward the understanding that

with the identification of the lack of svabhāva, no true existence remains (while this

does not affirm non-existence).76 Therefore—if there is not even emptiness,

let alone something non-empty that could be subjected to the logic of emptiness,77

how is that we are speaking of emptiness?

It is here that our analysis comes full circle and where we grasp the full

significance of the correspondence principle, defined now as the structural

connection between understanding and reality. Recall that Nagarjuna’s arguments

against the pramāṇas are truly compelling only when a full correspondence between

analysis and reality is assumed. It is only if pramāṇas are meant to function in line

with the way they are defined that the arguments against them retain their full

appeal. A proponent of pramāṇa theory could easily have argued, for instance, that

the pramāṇas are known from inference (anumāna)—they can be deduced from the

fact that we live in the world; without pramāṇās we would be walking into walls

and could not eat an apple. Without pramāṇas we could also not know there was a

74 See note 60 above.75 This approach is at least as old as Candrakırti and became the hallmark of Tibetan dge-lug-pa

Madhyamaka. For a cogent presentation of this position see Thakchoe (2011).76 See note 36 above, and the quotes from the VV in Shulman (2007 [2009], p. 151).77 See, for example, MMK 13.7: If there were anything non-empty, there could be something empty too.

And there is no non-empty thing—how will there be something empty?

yady aśūnyaṃ bhavet kiṃ cit syāc chūnyam api kiṃ cana / na kiṃ cid asty aśūnyaṃ ca kutaḥ śūnyaṃbhaviṣyati //

Language, Understanding and Reality 365

123

notion of pramāṇas or of emptiness in the first place, let alone strive for an

agreement regarding what these concepts mean.78

But for Nagarjuna, all these arguments miss the mark, since analyzability, or

more simply—understanding, is a preliminary condition of existence. If there is a

real, it must be discernable; existence is accessed through understanding. Yet when

analysis proves there are no things, and the reliability of knowledge is lost, all that is

left is the analysis itself, and there is nothing to be apprehended by pramāṇas. Onecannot speak of anything that is not already subjected to a theoretical pattern, to a

scheme of understanding. Whatever is subject to such a scheme, though, is empty by

definition. Thus, for this brand of Madhyamaka, philosophical analysis comes as

close as one can get to an ontological substratum—the thinking about reality is the

closest thing to reality (or to un-reality). And when analysis cuts through its objects

of scrutiny to discover that both they and the terms with which the analysis is

conducted are empty, nothing remains but subjective, illusory perception; naked,

mistaken understanding is reality itself. Knowledge, just as ontology, is nothing but

shapes in the mass of illusion. When we are clear that all is unintelligible, and given

that shear nothingness is unintelligible as well, we are left in a hazy labyrinth of

fancy, which encompasses both words and things. All surrenders to emptiness. This

emptiness in not only a negative reality, but also a positive one; it is the truthful

appearance of the unreal.

There thus does appear to be a given for Nagarjuna—it is the very reliability of

his own analysis. It is justified, for him, since he assumes that all existence relates to

knowledge, and only since knowledge is dumbfounded does reality prove to be

empty. Yet why is it that Nagarjuna trusts his own analysis? Why does he accept

emptiness even though his analysis refutes itself? It is because this analysis itself is

for him a structural element of the very substance of the universe.

Does Nagarjuna’s thinking possess such a strong degree of metaphysical

commitment? Does his system not involve a calming of the attempt at understand-

ing? These are relevant and important questions, which surely reflect shades of the

logic of the empty. Nevertheless, as the Nyaya adversary of the VV accepts, the

negation of svabhāva is an ontological watershed, which allows nothing true to

remain. Here, one can suspend knowledge and refuse metaphysical reflection. But

these forms of philosophical relaxation are accessed only once everything real—

both knowledge and its objects—have become fully empty.

I would like to conclude with a suggestion I believe Nagarjuna may have thought

to address to his modern students, who at this point may incline to accuse him of

being driven by metaphysical impulses they find difficult to stomach. Why, it

may be asked, must both analysis and reality succumb to this problem of

78 Arguments based on inference were surely employed in Indian philosophical debate. See for instance

in Ratie (2007, Sect. I), who discusses the inference used by the Vijnanavada regarding the existence of

other minds. These argument also echo the ones voiced by the pūrvapakṣin of the VV we read above. See

also verse 5:

If you refute things after first apprehending them through perception, (according to your system)

there is no perception through which things may be apprehended!

pratyakṣeṇa hi tāvad yady upalambhya vinivartayasi bhāvān / tan nāsti pratyakṣaṃ bhāvā yenaupalabhyante //

366 E. Shulman

123

unintelligibility? Why is emptiness so all pervasive? Why, we may persist, can we

not be satisfied with the suspension of all knowledge and continue to maintain that

objective reality remains beyond the grasp of the human mind?79

Here Nagarjuna may ask whether some notion akin to the correspondence

principle is not a necessary pre-condition of all analytical inquiry. Can we in fact

ever analyze only a word or a concept? Are there such innocent, “naked”, concepts?

Are these concepts not themselves intentional, in the sense that they derive their

meaning from an interaction between subjectivity and the lived, the supposedly

external, world? And, also, can “things” be fully detached from their conceptual

framing? Given the inevitability of subjectivity for every experience of world,

coupled with the objective orientation of every moment of knowledge—is it not

perhaps more feasible that when we scrutinize our understanding, we are subjecting

reality itself to our inquiry?80

Whatever the answers to these questions may be, if there can ever be such

answers, I believe that they were part of what Nagarjuna had in mind when he was

doing philosophy. For him, words and concepts that are detached from “reality”

could never have been all he was analyzing. In his system, there can be neither

nothingness, nor an unspecified reality. Furthermore, important parts of classical

Indian and Buddhist thought can be seen as interested in articulating the logic of

these and other similar positions, according to which there is no clear boundary

between perception and the world. These traditions studied the forms of reality

dictated by such intuitions. This contemplation can be seen, in effect, as a sustained

interest in the notion of karma; or, perhaps, as a continued inquiry into the ancient

Indian vision, most familiar from the early Upanis˙ads, of a world propelled by

conscious reflection.81 This shaky reality, always deriving its being from a complex

and problematic subjectivity, continued to attract the attention of Indian

79 Huntington (2003, 2007) is a modern representative of this approach. Historically, The Prasangika-

Madhyamaka tradition has come to be associated with this type of skeptical approach that involves a

suspension of all claims to truth. Patsap Nyimadrag appears to have been one of the earliest and most

influential thinkers to advance this approach (for an initial presentation see Dreyfus 2011). Yet even

before Patshap, there appears to have been a central debate within the tradition regarding the correct

understanding of the Madhyamaka truth—whether it should be understood to be of metaphysical or of

epistemological import. In an important contribution to the study of the historical development of

Madhyamaka, Almogi (2010) has outlined the basic contours of the debate between the Mayo-pamādvayavāda (sgyu ma lta bu gnyis su med pa smra ba, “the strand of Madhyamaka which maintains

that [phenomena] are one, inasmuch as they are like illusions”, called also the Mayopamavāda) and the

Sarvadharmāpratiṣṭhānavāda (rab to mi gnas pa smra ba, “the strand which maintains that all

phenomena have no substratum whatsoever”). While the first of these schools tended toward a

metaphysically oriented view that emphasized the concept of illusion, the second inclined to the

“no-thesis” approach that is more popular in modern Madhyamaka studies. The main difference,

I believe, between the Sarvadharmāpratiṣṭhānavāda and their modern counterparts is that the first, at least

according to their rivals, advocate a nihilistic position regarding the absolute; when truth is realized, all

the dharmas that had no substratum are no more.80 According to my understanding, Vasubandhu presents arguments similar to these in his Vimśatikā,especially in verses 6 and 7. See Shulman (2011). See also the quote from and references to the work of

Mohanty (1992, 1995) in note 6 above.81 Many Upanis

˙ads articulate a vision of creation according to which the initial reflection of a conscious

element leads to the externalization and differentiation of the world of diversity. See for example

Bṛhadāraṇyakaiupaniṣad 1.4.1–1.4.4 and Chāndogya-upaniṣad 6.2–6.4.

Language, Understanding and Reality 367

123

philosophers through the ages. Specifically, in the present case, Nagarjuna’s thought

can be seen as a reflection on the texture of life in a world that is never more, and

never less, real than the way it is thought.

Acknowledgments I wish to thank Eli Franco, Yohanan Grinshpon and Roy Tzohar for their mosthelpful comments on an earlier draft of this paper. All errors that remain are my own.

References

Almogi, O. (2010). Mayopamadvayavada versus Sarvadharmapratis˙t˙hanavada: A late Indian subclassi-

fication of Madhyamaka and its reception in Tibet. Journal of the International College forPostgraduate Buddhist Studies, XIV, 135–212.

Bhattacarya, K., Johnston, E. H., & Kunst, A. (1978). The dialectical method of Nāgārjuna:Vigrahavyāvartanī. Delhi: Motilal Banarasidass.

Brereton, J. P. (1999). Edifying puzzlement: R˙gveda 10.129 and the uses of enigma. Journal of the

American Oriental Society, 119.2, 248–260.Bronkhorst, J. (1997). Nagarjuna’s logic. In P. Kieffer-Pulz & J. Hartman (Eds.), Bauddh-

avidyāsudhākāraḥ: Studies in honour of Heinz Bechert on the occasion of his 65th birthday.Swistall: Swistall-Odendorf.

Bronkhorst, J. (2011). Language and reality: On an episode in Indian thought. Leiden: Brill.Dreyfus, G. (2011). Can a Madhyamaka be a skeptic? The case of Patsab Nyimadrak. In The Cowherds

(Ed.), Moonshadows: Conventional truth in Buddhist philosophy (pp. 89–113). New York: Oxford

University Press.

Franco, E. (2004). A note on Nagarjuna and the Naiyayikas. In S. Hino & T. Wada (Eds.), Threemountains and seven rivers: Prof. Musashi Tachikawa’s felicitation volume. Delhi: Motilal

Banarsidass.

Garfield, J. L., (1995). The fundamental wisdom of the middle way: Nāgārjuna’s Mūlamadhyamakakārikā.New York and Oxford: Oxford University Press.

Gonda, J. (1966). Loka: World and heaven in the Veda. Amsterdam: N.V. Noord-Hollandsche Uitgevers

Maatschappij.

Gonda, J. (1969). Eye and gaze in the Veda. Amsterdam: North-Holland Publishing Company.

Hayes, R. F. (1994). Nagarjuna’s appeal. JIP, 22, 299–378.Hopkins, J. (1996 [1983]). Meditation on emptiness. Boston: Wisdom Publications.

Huntington, Jr., Ch. W. (2003). Was Candrakırti a Prasangika? In G. B. J. Dreyfus & S. L. McClintock

(Eds.), The Svātantrika-Prāsaṅgika distinction: What difference does a difference make. Boston:Wisdom Publications.

Huntington, Jr., Ch. W. (2007). The nature of the Madhyamaka trick. JIP, 35, 103–131.rJe Tshong Khapa. (2006). Ocean of reasoning: A great commentary on Nāgārjuna’s Mūlama-

dhyamakakārikā (G. N. Samten & J. L. Garfield, Trans.). New York: Oxford University Press.

Johnston, E. H., & Kunst, A. (1978). In K. Bhattacarya, E. H. Johnston, & A. Kunst (Eds.), TheVigrahavyāvartanī of Nāgārjuna with the author’s commentary.

Komito, D. R. (1987). Nāgārjuna’s “Seventy stanzas”: A Buddhist psychology of emptiness. Ithaca: SnowLion.

Lindtner, Chr. (1982). Nāgārjuniana: Studies in the writings and philosophy of Nāgārjuna. Delhi: Motilal

Banarsidass.

Lindtner, Chr. (1986). Master of wisdom: Writings of the Buddhist master Nāgārjuna. Oakland: Dharma

Press.

Matilal, B. K. (1971). Epistemology, logic and grammar in Indian philosophical analysis. The Hague:

Mouton.

Matilal, B. K. (1982). Logical and ethical issues of religious beliefs. Calcutta: Calcutta University Press.

Matilal, B. K. (1990). The word and the world: India’s contribution to the study of language. Delhi:Oxford University Press.

Matilal, B. K. (2008 [1985]). Logic, language and reality. Delhi: Motilal Banarsidass.

Mohanty, J. N. (1992). Reason and tradition in Indian thought: An essay on the nature of Indianphilosophical thinking. Oxford: Clarendon Press.

368 E. Shulman

123

Mohanty, J. N. (1995). In P. Bilimoria (Ed.), Essays on Indian philosophy traditional and modern. Delhi:Oxford University Press.

Oetke, C. (1990). On some non-formal aspect of the proofs of the Madhyamakakarikas. In Ruegg, David

S., and Schmithausen, Lambert, (Eds.), Panels of the VIIth World Sanskrit Conference, Vol.II:Earliest Buddhism and Madhyamaka. Leiden, New York, København, Koln: E. J. Brill.

Oetke, C. (2003). Some remarks on thesis and philosophical positions in early Madhyamaka. JIP, 31,449–478.

Oetke, C. (2004). On ‘Nagarjuna’s logic’. In H. R. Bodewitz & M. Hara (Eds.), Studia PhilologicaBuddhica, monograph series XVII, Gedenkschrift J. W. de Jong. Tokyo: The International Institute ofBuddhist Studies of the International College for Advance Buddhist Studies.

Oetke, C. (2011). Two investigations on Madhyamakakārikās and the Vigrajavyāvartanī. JIP, 39,245–325.

Ratie, I. (2007). Otherness in the Pratyabhijna philosophy. JIP, 35, 313–370.Ruegg, D. S. (2000). Three studies in the history of Indian and Tibetan Madhyamaka philosophy: Studies

in Indian and Tibetan Madhyamaka thought part 1. Vienna: Arbeitskreis fur Tibetische und

Buddhistische Studien Universitat Wien.

Scherrer-Schaub, C. A. (1991). Yuktis˙as˙t˙ikavr

˙tti: Commentaire à la Soixantaine sur le Raisonment ou Du

Vrai Enseignment de la Causalité par le maître Candrakīrti. Bruxelles: Institut Belge des Hautes

Etudes Chinoises (Melanges Chinois et Bouddhiques).

Shulman, E. (2007 [2009]). Creative ignorance: Nagarjuna on the ontological significance of

consciousness. Journal of the International Association of Buddhist Studies, 30(1–2), 139–173.Shulman, E. (2008). Nagarjuna on impermanence, the Buddha on illusion. In D. Shulman & S. Weil

(Eds.), Karmic passages: Israeli scholarship on India. Delhi: Oxford University Press.

Shulman, E. (2009). The fullness of emptiness: Nāgārjuna’s thought in light of the Yukti-ṣaṣṭikā-kārikā andthe Śūnyatā-saptati. PhD dissertation, The Hebrew University of Jerusalem.

Shulman, E. (2010). The commitments of a Madhyamaka trickster: Innovation in Candrakırti’s Prasanna-padā. JIP, 38, 379–417.

Shulman, E. (2011). Vasubandhu on truth and subjectivity. Journal of Indian Philosophy and Religion.Shulman, D. (2012). More than real: A history of the imagination in South India. Cambridge, MA:

Harvard University Press.

Siderits, M. (1988). Nagarjuna as anti-realist. Journal of Indian Philosophy, 16, 311–325.Taber, J. (1998). On Nagarjuna’s so-called fallacies: A comparative approach. Indo-Iranian Journal, 41,

213–244.

Thakchoe, S. (2011). Prasangika epistemology in context. In The Cowherds (Ed.), Moonshadows:Conventional truth in Buddhist philosophy (pp. 39–55). New York: Oxford University Press.

Tola, F., & Dragonetti, C. (1995). On voidness: A study in Buddhist nihilism. Delhi: Motilal Banarsidass.

Tola, F., & Dragonetti, C. (1998). Against the attribution of the Vigrahavyavartanı to Nagarjuna. WZKS,XXXXII, 151–166.

Westerhoff, J. (2009). Nāgārjuna’s Madhyamaka: A philosophical introduction. New York: Oxford

University Press.

Westerhoff, J. (2011). The merely conventional existence of the world. In The Cowherds (Ed.),

Moonshadows: Conventional truth in Buddhist philosophy (pp. 189–212). New York: Oxford

University Press.

Language, Understanding and Reality 369

123