By a misperception of emptiness,

A person of little intelligence is destroyed.

Like a snake wrongly held

Or like a spell wrongly executed (MMK 24.11)

My propositions serve as elucidations in this way: he who understands me eventually recognises them as nonsensical, when he has used them — as steps — to climb up over them. (He must, so to speak, throw away the ladder after he has climbed up it). He must overcome these propositions, and then he will see the world aright (TLP 6.54).


Given the emphasis placed on the perceived philosophical alignment between Nagarjuna’s Madhyamaka and Wittgenstein, we might think there would be some consensus about what this comparison involves. Instead, not only do comparisons disagree about how we should interpret Wittgenstein’s philosophy but also about which Wittgenstein too. Therefore, it is not clear whether we should see Nagarjuna’s MadhyamakaFootnote 1 as comparable with either the ‘early’ Wittgenstein of the Tractatus Logico-Philosophicus (Waldo 1975, 1978; Matilal 2002; McEvilley 1982; Anderson 1985; Garfield 1995, 1996, 2002; Tillemans 1999; Garfield and Priest 2003; Priest 2009, 2015) or the ‘later’ Wittgenstein of the Philosophical Investigations (Streng 1967; Hudson 1973; Gudmunsen 1974, 1977; Waldo 1975, 1978; Katz 1981; Huntington Jr 1989, 1995; Vukomanovic 2004; Tillemans 2003, 2016, 2017).

What makes this such a curiosity is that nearly all previous comparative studies adopt what is known as a ‘standard’ reading of Wittgenstein’s philosophical development. This is to say, as a progression from realism in the Tractatus to anti-realism in the Philosophical Investigations (Crary 2000, 7-8). Therefore, Madhyamaka seemingly admits comparison with philosophies which are radically discontinuous: this is a problem which has seemingly all but gone unnoticed by scholars of Madhyamaka and comparative philosophy alike. However, far from counting against either alleged similarity, recent developments in the scholarship of Wittgenstein admit the possibility of alignment between Madhyamaka and both Wittgenstein’s early and later philosophy.

Via the use of a ‘resolute’ reading of Wittgenstein’s Tractatus as an interpretive model, we may identify a philosophical affinity with Madhyamaka whilst also rejecting the idea that the development of Wittgenstein’s thought is radically discontinuous (Crary and Read 2000).Footnote 2 As a result, we are then in a position to dissolve the aforementioned interpretive problem, providing an account of why both the ‘early’ and ‘later’ Wittgenstein may be compared with Madhyamaka. Therefore, the aim of this paper is to provide a comparative study concerning the affinity between Nagarjuna and Wittgenstein’s Tractatus which is also fundamentally aligned with Wittgenstein’s later work.

However, my aim is not simply to provide a reading of Madhyamaka which maximises comparison with Wittgenstein. Rather, it is to account for the appearance of philosophical impulses in Nagarjuna’s philosophy which allow such contrasting readings to occur in the first place, thereby also suggesting a new understanding of the polarised history of the Madhyamaka interpretation. To see Madhyamaka as either a form of realism or as a rejection of realismFootnote 3 (as in the ‘standard’ reading of the Tractatus or Philosophical Investigations respectively) is, I suggest, not simply an interpretive mistake. Instead, we should understand each in terms of philosophical impulses that both Nagarjuna and Wittgenstein intend us, as James Conant puts it, to ‘feel the force of’ but without becoming ‘in the thrall of’ (Conant 2002, 377). Seen this way, such temptations are a manifestation of a crucial stage of Nagarjuna’s ‘therapeutic’ method, one which, if we transverse correctly, can liberate us from the problems of both impulses.

As I will explore below, it is the deliberate employment of this tension as a therapeutic device which has also led to what I propose is a parallel set of interpretive trends shared between the Tractatus and Nagarjuna’s philosophy. My suggestion then is towards a radical reinterpretation of Madhyamaka, one which both provides an account for and avoids existing contrary metaphysical and anti-metaphysical interpretations.

Similarity of Form in Nagarjuna and Wittgenstein’s Tractatus

In the past 40 years, numerous scholars have variously compared Nagarjuna and Wittgenstein’s Tractatus. This might seem unexpected, given that the early Wittgenstein and Madhyamaka could hardly appear more dissimilar.Footnote 4 Given their remoteness, why then have such comparisons continued to be drawn? My suggestion is that this can be accounted for in terms of a shared similarity in form or structure. What I mean by this is that both the Tractatus and Nagarjuna’s Madhyamaka seem to simultaneously offer (1) a thesis/doctrine, (2) a refutation of all theses/doctrines, and (3) a disavowal of both.

If we first consider the Tractatus, I think we can understand it in terms of two broad impulses. On one hand, Wittgenstein seems to state a metaphysical doctrine.Footnote 5 We are told that the world is made up of ‘facts, not of things’ (TLP 1),Footnote 6 being ‘the totality of existing states of affairs’ (TLP 2.04). In turn, these ‘states of affairs’ consist of a ‘combination of objects’ (TLP 2.01) which are logically simple (TLP 2.02). Wittgenstein also treats us to host of other claims, principally concerning what is common between language and the world, namely, ‘logical form, i.e. the form of reality’ (TLP 2.18). However, Wittgenstein also appears to reject metaphysics, suggesting that someone who makes such claims has ‘failed to give a meaning to certain signs in his propositions’ (TLP 6.53). This places the reader of the Tractatus in a position where the Wittgenstein appears both to propose metaphysical doctrines and reject metaphysics, the latter emerging as a direct result of the former. To confuse matters further, Wittgenstein then states the following:

My propositions serve as elucidations in this way: he who understands me eventually recognises them as nonsensical, when he has used them — as steps — to climb up over them. (He must, so to speak, throw away the ladder after he has climbed up it). He must overcome these propositions, and then he will see the world aright (TLP 6.54).

His readers are then urged to do something apparently quite inexplicable, namely, that we ‘overcome’ the very propositions over which we have laboured to understand on the grounds that they are ultimately ‘nonsensical’. Whatever conclusions or insights thought we might have garnered from his propositions have now also been thrown into question.

I suggest that we also see a very similar tension in Nagarjuna’s work. Central to his thought is what appears to be three closely related doctrines, namely, that all things are ‘Empty’ (MMK 24.19), ‘dependently originated’ (MMK 24.16-19), and that there are ‘two truths’, the ‘conventional’ and the ‘ultimate’ (MMK 24.8-9).Footnote 7 Given their form, these certainly seem to be metaphysical doctrines, statements telling us about the nature of reality. However, once we realise, as Jan Westerhoff puts it, that ‘emptiness is emptiness of something and indicates something is not there’ (Westerhoff 2009, 12), we discover that this ‘something’ is svabhāva—or ‘essence’—that is, precisely the domain of ‘ultimately real’ entities posited in metaphysical theories. Therefore, Nagarjuna appears to be simultaneously telling us about the nature of reality whilst denying that such a domain exists. In a moment reminiscent of TLP 6.54, Nagarjuna also seems to disavow ‘Emptiness’ altogether.

The Victorious ones have said

That emptiness is the relinquishing of all views

For whomever emptiness is a view

That one has accomplished nothing.Footnote 8

Just as with Wittgenstein’s Tractatus, we find an interpretive problem which turns on the question of the philosopher’s intended attitude towards metaphysics. Nagarjuna at once appears to be (1) expounding a set of doctrines about the nature of reality, (2) apparently providing us with a means to critique all theories precisely of that kind, and (3) providing a disavowing statement throwing into question the previous two elements. Both philosophies therefore present us with the paradoxical temptation both to see them as constitutive of a metaphysical thesis and as a final rejection of metaphysics.

In order to understand the way in which existing comparisons concerning Nagarjuna and the Tractatus, I suggest it is necessary to understand the interpretation of Wittgenstein’s work first. As I will explore, we will find that each form of popular interpretation corresponds to one of the two central philosophical impulses we see exhibited in the Tractatus.

Metaphysics and the Interpretation of the Paradox of the Tractatus

Given our interpretive quandary, it is perhaps unsurprising that readings of the Tractatus tend to be split along lines of their attitude to metaphysics. Upon publication, Wittgenstein’s book was first taken as concerned with demonstrating the incoherence of metaphysics, becoming highly influential amongst the followers of Logical Positivism as a work which distinguished between meaningful and metaphysical statements. However, the problem with reading the Tractatus in this way is that it ignores that TLP 6.54 seems to signal Wittgenstein understood the instability of this kind of conclusion. That is, he was aware that his apparent refutation of metaphysics paradoxically requires metaphysical propositions.

Eventually this realisation was acknowledged in later re-evaluation of the Tractatus by such readers, leading Carnap to explicitly reject Wittgenstein’s claim (in TLP 6.54) that ‘all his propositions are quite as much without sense as metaphysical propositions are’ (Carnap 1935, 37) and Ayer recanting the reading altogether (Ayer 1985, 30-1). The problem Carnap saw (and sought to deny the force of) was that in order to read the Tractatus as rejecting metaphysics, we would require a kind of statement which, although makes no assertions about the world, allows one to make firm claims about the nature of language (e.g. ‘the’ logical syntax of language). That is, we would require something capable of giving us reason to determine if a sentence is meaningful which does not fall into metaphysics itself. Given Wittgenstein’s comments, we are beginning to see that it might be that he did not distinguish between the nonsense of metaphysics and the kinds of assertions required to ground an anti-metaphysical theory at all.

It is precisely this paradoxical quality of the Tractatus which later interpretations—sometimes referred to as ‘standard’ or ‘ineffabilist’ readingsFootnote 9—were forced to come to terms with. In placing greater emphasis on the appearance of realist ontology present in the book, they attempted to understand the nonsensicality of Wittgenstein’s propositions without this precluding them from also being true. This principally involves identifying the so-called “picture theory” of meaning (TLP 2.1-2.225). With this analogy, Wittgenstein is seen to ground the possibility of meaningful language in a logical commonality between the ‘picture’ and what is ‘pictured’. As he puts it, ‘what any picture …must have in common with reality, in order to be able to depict it …is logical form, i.e. the form of reality’ (TLP 2.18). As David Pears explains, ‘Wittgenstein believed that the possibility of saying some things in factual discourse depends on the actuality of other things which cannot be said’ (Pears 1987, 143), the analogy with pictures is thus intended ‘…to illustrate the dependence of the sayable on the unsayable’ (Ibid). Attempting to describe this logical form would be to go beyond the limits of language and attempt to say what can only be shown (TLP 4.1212), precisely the error attributed to metaphysics. However, in describing this relationship, Wittgenstein’s propositions also fall into nonsense.

On this reading, what is ‘thrown away’ in TLP 6.54 is that its propositions may be stated in meaningful expression, not their truth. When Wittgenstein says ‘what we cannot speak about we must pass over in silence’ (TLP 7); for the ‘standard’ reader, this silence is a pregnant one. Whichever way we read the Tractatus, for any interpreter, ‘the basic puzzle of Wittgenstein’s Tractatus is that it makes many metaphysical pronouncements, while at the same time it declares metaphysics to be impossible’ (Goldfarb 2011, 6). The fact that this statement applies as much to Nagarjuna as it does to the early Wittgenstein is precisely what has led to comparisons between their work.

‘Standard’ Comparative Studies of Madhyamaka and the Tractatus—a Reappraisal

What all existing comparisons between Madhyamaka and the early Wittgenstein have in common is a reliance on this ‘standard’ reading of the Tractatus, thereby presupposing we can account for their ‘similarity in form’ in terms of an ineffability thesis. We can distinguish three different ways in which this ‘standard’ comparative account arises: via (1) self-refutation, (2) via negativa, and (3) paraconsistency.

The first of these three types draws upon the idea that we can understand both Nagarjuna and the Tractatus in terms of a shared paradox of expressibility. This has been made variously, including reading Madhyamaka in terms of a self-refuting claim such as ‘language is fundamentally incoherent’ (Waldo 1975, 284) or that ‘any sort of dualistic concept…will be logically contradictory’ (Anderson 1985, 166). In these cases, the point of contact with the Tractatus is to be understood as (1) a shared conviction that language is incapable of describing reality and (2) that this is explained via a paradoxical doctrine.

The second variant of interpretation (as solely proposed by Bimal Krishna Matilal) also viewed Madhyamaka as involving the expounding of an ineffability thesis of a sort ‘not very remote from the contention of Wittgenstein’ (Matilal 2002, 70). By way of a form of via negativa argument, Matilal suggested that ‘although no positive characterisation of the Ultimate Reality is possible, a negative characterisation may be in order’ (Ibid, 67). Simultaneously, Matilal’s Nagarjuna also assigns ‘contradictory attributes or predicates to…Ultimate Reality’ (Ibid). That is, given a set of exhaustive (logically contradictory) options concerning what reality must be like, Nagarjuna negates them all. Unlike the self-refutation variant, Matilal’s account draws not upon the paradoxicality of his ‘doctrines’, but rather on Nagarjuna’s method of prasaṅga, that is, the method of reductio ad absurdum used to criticise other philosophers’ attempts to describe ‘ultimate reality’. Thus, Nagarjuna’s success in elaborating these reductio arguments is what evidences to the ineffability of their subject matter.

Finally, in a recent and influential variant, Garfield and Priest (2003) have suggested an affinity between Madhyamaka and the Tractatus in terms of the employment a form of paraconsistent logic.Footnote 10 That is, the view that Nagarjuna ‘...discovers and explores true contradictions arising at the limits of thought’ (Garfield and Priest 2003, 2). Their central claim is that there are two paradoxes in Nagarjuna’s philosophy: one, a paradox of expressibility (Ibid, 13), and, the other, what they describe as ‘Nagarjuna’s paradox’ roughly, the ‘discovery’ that all things simultaneously lack and possess an essence (namely, their lack of essence) (Ibid, 18).Via the use of a non-standard logic, Nagarjuna’s philosophical contribution is therefore seen as the affirmation of these paradoxes as truth.

As with the first variant above, the Tractatus and Nagarjuna’s philosophy are seen as similar in that they both exhibit a paradox of expressibility. However, what Garfield and Priest also suggest is that ‘Nagarjuna’s paradox’ is something like an answer to the question ‘why do paradoxes of expressibility arise at all?’ They conclude that reality defies meaningful expression because it is inherently something which may simultaneously be P and ~P; this is to say that it has a ‘contradictory nature’ (Ibid). This is why the second paradox is meant to explain why paradoxes of the first sort appear, not only in Nagarjuna’s philosophy and the Tractatus, but throughout the history of philosophy.

Despite their differences, it is clear that what all three variants of the ‘standard’ comparison with the Tractatus share is the idea that both philosophers intend to gesture towards something about which we can only be silent. However, I think that the main problem with this way of cashing out the ‘similarity of form’ between Nagarjuna and the ‘early’ Wittgenstein is that, whilst the Tractatus at least has something to gesture towards (a realist ontology), there is nothing in Nagarjuna’s philosophy which can conceivably play this role. Instead, the only thing ‘contained’ in Nagarjuna’ alleged doctrines is that there is no ‘ultimate reality’; in other words, that metaphysical doctrines do not refer to anything.Footnote 11 I suggest that it is the failure to realise this error is in particular which has led to the recent invocation of paraconsistent logic in order to understand Madhyamaka, it being the logical conclusion of forcing Nagarjuna into a fundamentally realist interpretive framework. That is, we end up trying to say both that he has metaphysical doctrines and disavows them. This is why Garfield and Priest feel they must read Nagarjuna as stating that ‘everything is empty, there is no “ultimate reality”. There are, therefore, no “ultimate truths”’ (Garfield and Priest 2003, 10) and that ‘there are “ultimate truths” …he [Nagarjuna] is telling us about the nature of reality’ (Ibid).

Garfield and Priest’s reading therefore represents the inevitable terminus of the “standard” comparative reading with the Tractatus, the invocation of paraconsistency being the only way we might imagine maintaining coherency. As they put it, the use of a non-classical logic is meant to ensure that the fact that Nagarjuna’s philosophy is ‘inconsistent does not… mean that it is incoherent’ (Ibid, 20). I suggest that comparing Madhyamaka with the ‘standard’ reading of the Tractatus at all is to employ an interpretive framework which simply cannot make sense of Nagarjuna’s rejection of svabhāva. To reject svabhāva is to deny that it makes sense to invoke the allegedly ‘ultimately real’, whether this is to describe or refute its existence.

There is however another way of understanding the Tractatus which identifies affinity with Nagarjuna’s work without attributing an ineffability thesis to either philosophy. Furthermore, a reading which will allow us not only to appreciate the radicalism of Nagarjuna’s critique of svabhāva but also allows us to explain the impulse which leads us to think we can compare Madhyamaka with (e.g.) the ‘standard’ reading of the Tractatus.

The ‘Resolute’ Reading of Wittgenstein’s Tractatus

Originally arising out of the work of Cora Diamond (1988) and James Conant (1989), the ‘resolute’ reading of Wittgenstein suggests a radical interpretive strategy in order to make sense of the Tractatus’ main interpretive paradox. The central idea is that when Wittgenstein claims that his propositions are nonsensical, we should understand this to be taken completely at face value. This is to deny that nonsense can be anything more than just that, barring the kind of ‘gesturing’ towards metaphysical truths we see in the ‘standard’ reading. The ‘resolute’ reading therefore relies on the denial that Wittgenstein distinguished between the patent nonsense of gibberish and the kind of nonsense of metaphysical theses (including the propositions of the Tractatus). This ‘austere conception’ of nonsense (Conant 2000, 176) questions the central mechanism common to both Positivist anti-metaphysical readings and ‘standard’ metaphysical readings, namely, ‘(what the Tractatus calls) “an understanding of the logic of our language”’ (Conant 2002, 376). This ‘understanding’ is what allows us to imagine that we can differentiate ‘deep’ forms of nonsense which are created when (as in the ‘standard’ reading) we ‘violate the rules of logical syntax’ (Hacker 1986, 18). This is to say nonsense is not analysable in terms of when otherwise meaningful components are combined in an incorrect fashion.

Without insights into the logic of our language, there can be no such thing as a coherent thought which is somehow unable to be ‘accommodated by the logical structure of language’ (Conant 2002, 422). Consequently, given their reliance on the possibility of such insights into ‘the logic of our language’ (TLP 4.003), this calls both ‘standard’ and Positivist readings into question. In its place, Conant and Diamond (et al.) have suggested that the paradox in the Tractatus is to be understood as a dialectical moment in which we see the instability of this conception of ‘the logic of our language’. In refusing that Wittgenstein’s intention is to expound either metaphysics or its refutation, the Tractatus is recast us urging us to overcome what is common to the imagined coherency of both. Therefore, the ultimate aim of the Tractatus is seen to be seeking a ‘…way to do philosophy that does not consist in putting forward philosophical theses’ (Conant and Diamond 2004, 84).

Rather than allowing Wittgenstein to gesture towards inexpressible truths concerning language and reality, the appearance of a paradox is taken to act as a kind of reductio ad absurdum on what is common to realist and Positivist readings alike. We are told that the Tractatus ‘uses apparently metaphysical sentences, but in a way which is disposed of by the sentences which frame the book’ (Diamond 1991, 19). In other words, anything which might appear to be a metaphysical doctrine or a general rejection of metaphysics is deliberately employed by Wittgenstein as a device to enable us to reach his true intentions. However, Wittgenstein is not playing a trick on us; the purpose of this misdirection is in order for us to take seriously the contrary philosophical impulses he exhibits in the book. That is, it would simply not be enough for him to recommend that we avoid each impulse purely at the meta-philosophical level. After all, this would simply be to invite self-refutation, positing a controversial thesis about the nature of philosophy.

Nagarjuna and the ‘Resolute’ Reading of the Tractatus

In light of the interpretive strategy above, I suggest that we may now understand the ‘similarity in form’ shared between the Tractatus and Nagarjuna’s work in a new way. On a ‘resolute’ reading of the sort Conant suggests, Wittgenstein’s Tractatus is intended as a journey; we are supposed to undergo ‘a certain experience’ (Conant 2002, 422) in which we ‘feel the force of…mutually antagonistic impulses’ (Ibid, 377) but without becoming ‘ the thrall of either’ (Ibid). Wittgenstein does not mean to impart a doctrine, but wishes us to see that both metaphysics and its final refutation share a common presupposition. In the Tractatus itself, the crucial dialectical moment comes when we see that “a sentence like “‘The world is everything that is the case’ is nonsense” is itself nonsense” (Ostrow 2001, 10). In other words, just as we were led to ultimately reject metaphysics by his propositions, this very act of rejection is a rung in the ladder also to be overcome.

I propose that this is how we should also understand the tension in Nagarjuna’s philosophy. Seen this way, Nagarjuna’s doctrines are therefore intended to embody the contrary impulses which have also (inadvertently) made themselves manifest in the interpretive tradition. This is to say that, Nagarjuna’s philosophy deliberately exhibits the dual aspect of both metaphysics and the refutation of metaphysics. Seen in this way, misinterpretations of his work are not merely mistakes but explicable entirely as temptations we are to entertain and then overcome. As James Conant says of the Tractatus, ‘it was really quite unfair of me... to criticize my friends the commentators for simply having misread these works... there is a reason internal to the structure of these works themselves that readings such as theirs recurrently resurface’ (Conant 1989, 262).Footnote 12 Just in the same way, we can also understand the history of Madhyamaka scholarship as various attempts to engage with these temptations; however, the ‘middle way’ is just what it means to feel these temptations but not be held by them.

I think this understanding offers an interesting link with some of Jan Westerhoff’s recent discussion of what he calls the ‘opponent-relative’ feature of Madhyamaka (Westerhoff 2016, 372), whereby what we take to be Nagarjuna’s philosophical position seems to shift in accordance with what we see him as rejecting. Westerhoff suggests his target is often construed as one of the following:

  • Target (1). An error theorist who ‘assumes that it lies in the world behind this world, i.e. the source of the illusory appearances’

  • Target (2). A naïve realist who ‘locates “ultimate truth” in the world we live in’ (Ibid, 373)

When Nagarjuna as geared to reject one of these two positions, by way of a kind of intellectual recoil, ‘residues from or reflections of the views it sets out to negate’ (Ibid, 372) become part of what is taken to be Nagarjuna’s view. Thus, seen as a rejection of error theory, Nagarjuna’s philosophy nonetheless attempts to account for the problems that view claims to identify. This leads us to read Nagarjuna’s ‘two truths’ in terms of an appearance/reality distinction, the former acting as containment for the perceived error, whilst saving the ‘ultimately true’ reality. When the target is taken to be naïve Realism, there is a tendency to ‘push the “object of negation” into the realm of the epiphenomenal’ (Ibid); or, in simpler terms, Nagarjuna’s target is merely the imaginings of philosophers, leaving the ‘conventional’ world entirely untouched. Thus, the ‘conventional’ is only illusory in that it seems to be ‘ultimately real’.

This interdependence should not be a surprise to us. After all, there is a sense in which neither a realist nor an anti-realist wishes to deny the world as we experience it: this would be a (self-refuting) nihilism. Consequently, neither description really wishes to deny everything about the others account; there is always something of the others account left over, as, in Westerhoff’s words, ‘residues…or reflections’ (Ibid). Given Westerhoff’s use of the ‘reflections’ metaphor above, it is interesting that we see precisely the same imagery of ‘mirroring’ in Conant’s work on the Tractatus; for him, contrary interpretations are ‘the expressions of philosophical impulses that the Tractatus seeks to engage’ (Conant 2002, 377) which represent a ‘mirror image…each feeding on and sustaining the other’ (Ibid, 376). However, I do not think this is a coincidence. Instead, I suggest the two interpretive impulses Conant identifies are concomitant with those Westerhoff describes in Nagarjuna’s philosophy, being both deliberately exhibited by each philosopher in order for us to ultimately overcome.

We can now provide an account of the source of comparisons with ‘standard’ readings of the Tractatus in terms of target (1) above,Footnote 13 that is, as in the impulse to see Nagarjuna as concerned with rejecting error theory, thereby reinstating the ‘real’ behind a veil of (erroneous) appearances. Conversely, understood in terms of target (2), to see Nagarjuna as a rejection of naïve realism effectively reinstates the ‘real’ at the level of appearances. The mistake of the realist is viewed as the superimposition of a metaphysical domain over our everyday experience of a world that is already as ‘real as it gets’. This understanding of Nagarjuna is what accounts for readings of his work in terms of an anti-metaphysical thesisFootnote 14; therefore, it is also responsible for existing comparisons with the (allegedly) anti-realist Wittgenstein of the Philosophical Investigations.

The Status of Nagarjuna’s Alleged ‘Doctrines’

As a result of a comparison with the ‘resolute’ reading of Wittgenstein’s Tractatus, we are now in a position to understand the source of existing comparisons between Nagarjuna and Wittgenstein’s ‘early’ and ‘later’ work. In addition, I suggest this provides an account of why metaphysical and anti-metaphysical interpretations have arisen throughout its Nagarjuna’s interpretive history, thereby illuminating a new way of understanding the purpose of his philosophy. Reflecting on the status of Nagarjuna’s ‘doctrines’, I suggest we may now understand the appearance of paradox in the philosophy of Madhyamaka in terms of a kind of reductio ad absurdum. By this I mean as in terms of Nagarjuna’s method of prasaṅga, where a set of apparently contradictory claims are equally negated, uncovering a common presupposition.Footnote 15 In the same fashion, of the two opposing impulses embodied in his work, Nagarjuna does not wish to reject one in favour of the other but to question what is common to both, namely, a conception of philosophy as theorising.

Consequently, there is a sense in which it is inevitable that Nagarjuna’s philosophy will strike us in the way characteristic of terms of the contrary interpretive trends above. Indeed, given that we must feel the urge to fall into his trap in order to understand him correctly, I suggest it is precisely Nagarjuna’s intention.Footnote 16 This explains why Nagarjuna employs the language and method of philosophical theorising whilst also urging us to disavow all views of that kind. The reason this is so important is that if we were to convince ourselves that there is nothing genuinely tempting about each impulse, this would only be gained from the illusory self-assurance a thesis. To make this mistake would represent an inadvertent translation of Nagarjuna’s insight into a metaphilosophical claim, something he could not coherently maintain.

But what then is the status of Nagarjuna’s alleged doctrines? One of the main objections to a reading in which Nagarjuna expounds no philosophical theories is that it represents a failure to ‘take Madhyamaka argumentation seriously’ (Priest, Siderits and Tillemans 2011, 150). Tom Tillemans characterises this as a tendency to see Nagarjuna’s arguments as ‘exclusively an application of Buddhist “skill in means” (upāyakauśalya)—that is, it only consists in techniques that de facto work toward soteriological ends with particular opponents’ (Tillemans 2017, 120). The suggestion seems to be that anything other than the affirmation of Nagarjuna’s ‘doctrines’ as asserting a particular philosophical thesis represents a dangerous slide into some form of misological deconstructionism.

In the extant literature, C.W. Huntington is apparently taken to be the archetypical exponent of such an interpretation, whereby a Madhyamika like Nagarjuna should not be seen as in the game of ‘demonstrating, proving or disproving anything’ (Huntington Jr 2007, 122). The central problem with this account is, as Garfield has quite brilliantly highlighted, ‘Huntington takes rationality and essentialism to go hand in hand. Nagarjuna sees that they must come apart’ (Garfield 2008, 25). However, I do not think that refusing Nagarjuna’s alleged ‘doctrines’ the status of theses should necessarily imply Huntington’s conclusion. My suggestion is that Nagarjuna does wish to ‘demonstrate’ something—namely, a therapeutic rejection of theorising in philosophy—however, on pain of self-refutation, this is not something that he could impart with a thesis.

To see Nagarjuna as rejecting philosophical theorising is not to assume a kind of ersatz rationality—it is just as Garfield says; rationality and ‘essentialism’ do come apart—but for Nagarjuna, the form of philosophical inquiry in terms of theses is what ‘essentialism’ is. We might say that the function of svabhāva is to act as truth bearers for philosophical theories. Therefore, having identified the link between the method of inquiry (thesis-based) and the imagined object of inquiry (‘ultimate reality’/svabhāva), we may relinquish the philosophical method which generates the need to denounce or affirm such imagined objects. In other words, part of what it means to be liberated from the ultimate is to relinquish the need to either describe or disprove anything once-and-for-all. To demand a proof of this would be simply another form of the impulse which both Nagarjuna and Wittgenstein wish us to overcome.

I think we can see this tendency in the work of some contemporary interpreters of Madhyamaka, who, in attempting to distance Nagarjuna’s philosophy from metaphysical debates, nonetheless presuppose we need just such a proof. The clearest expression of this is in recent attempts to read Madhyamaka as advocating a particular semantic theory. In proposing that Nagarjuna’s philosophy can be reduced to a claim rejecting a realist conception of semantics, this also results in claims that Madhyamaka can be linked to positions in contemporary epistemology. Therefore, Nagarjuna also emerges as an alleged proponent of a deflationist theory of truth of the sort associated with Paul Horwich (Priest, Siderits and Tillemans 2011, 132). Given that such readings are an attempt to distance Nagarjuna from avowedly metaphysical interpretations of his work (e.g. Ferraro 2013), I do not disagree with their intentions per se. However, my suggestion is that they simply do not go far enough. In outsourcing what is taken to be the required central theoretical ‘insight’ of Nagarjuna’s work, I suggest the ‘semantic reading’ has simply moved Nagarjuna’s ‘doctrine’ into another domain of philosophy.

If we follow Nagarjuna in the realisation that we are unable to provide a theoretical basis for accepting (e.g.) even avowedly anti-metaphysical conceptions of truth amongst a group of possible candidates, it becomes unclear why Nagarjuna would need to take up a position in these debates at all, whatever his position. I suggest that this would be to fundamentally misunderstand him, giving into the perceived need to ground his recommendations in proofs and translating his method into a thesis. Furthermore, I think the notion that Nagarjuna would recognise the need to account for truth at all is to also misunderstand and underestimate his radicalism. The thought that truth requires theoretical grounding makes it sound too much like a privilege to be earned, which, if absent, would lead us to lose our ability to go on as proficient language users.

I suggest that this way of understanding and relating to truth qua philosophers is something which would be as alien to Nagarjuna as it would be to Wittgenstein. If Nagarjuna’s message is that if we have no reason to accept the coherency of metaphysical theories (i.e. those pertaining to “ultimate” truths), I do not need to see why he would worry about those amongst their number which claim to undermine our ‘conventional’ truths. After all, this would be a global error theory precisely of the kind he disavows. Whether we imagine Nagarjuna must endorse a theory in semantics, epistemology, or ontology, I suggest he would see each as simply a further source of disquietude.

However, could it not be objected that this is precisely what I am guilty of too? Am I not simply translating his philosophy into a meta-philosophical theory? Stefano Gandolfo offers an interpretation closest to my reply in this regard, suggesting that Nagarjuna’s philosophy is to be seen as ‘the rejection of all philosophical and meta-philosophical debates’ (Gandolfo 2014, 1). In other words, the real upshot of Nagarjuna’s method represents not a philosophical thesis concerning what philosophy must be, but rather the overcoming of the general temptation to offer such a proofs at all. In this sense, I think Gandolfo offers a reading closest to capturing Nagarjuna’s radicalism concerning the relationship between his rejection of svabhāva and what this means for philosophical methodology.Footnote 17 That is, once the truth bearer for philosophical claims no longer exists, the domain for philosophical enquiry as traditionally construed is equally absent. In light of this, to ask what philosophy must be once-and-for-all (equally) is not a coherent question.

However, in spite of this affinity, Gandolfo also fails to appreciate that what he calls the ‘the indeterminacy of the traditionally construed philosophical project’ (Ibid, 20) can be cashed out in terms of the therapeutic metaphor. He even goes so far as to suggest that the clear passages in Candrakīrti which most overtly suggest a similar metaphor (at PP 227 where Nagarjuna’s method is likened to a purgative drug)Footnote 18 ‘might not be beneficial’ (Gandolfo 2014, 20) for understanding Nagarjuna’s critique of traditional philosophy. Gandolfo’s rationale appears to be that the therapeutic analogy always leads us to see the aim of Madhyamaka as the ‘reduction of the intellectual contribution of Nagarjuna’s message in the MMK to mere psychological tool’ (Ibid, 18). Of course, if this were true, I do not deny it would represent a problem. After all, it leaves room for us to read Nagarjuna’s disavowal of philosophical theorising as only a kind of useful lie, that is, something which, although valuable for our psychological wellbeing, is ultimately just a form of trickery or deception (as in Huntington Jr 2007).

However, not only do I not think we need to drive a wedge between the philosophy of Candrakīrti and Nagarjuna to make this particular point, we also need not accept the claim that ‘therapy’ necessarily equates to a purely psychological (philosophically disinterested) pursuit of individual wellbeing at all. This is because, given that the ‘psychotherapeutic’ metaphor in Madhyamaka scholarship is sourced from the later Wittgenstein (PI §133)Footnote 19, I see no reason why using this metaphor should lead to such problems for Candrakīrti if it does not do so for Wittgenstein.Footnote 20 This is particularly true given that it is actually only a medical metaphor we see in the former’s work. Consequently, I think that for us understand (any) philosophy as ‘therapeutic’ does not mean we must be constrained by the metaphor of psychotherapy.

‘Therapy’ as Liberation

Given we have had cause to question how the ‘therapeutic’ metaphor relates to Nagarjuna, how then should we understand it? First and foremost, I think we can find the answer in what is common to both the medical and psychotherapeutic metaphor, namely, liberation from suffering. From this starting point, I suggest we can work our way towards seeing that ‘therapy’ as one and the same as Nagarjuna’s characteristic take on Buddhist soteriology. That is, it need not be the case that, on one hand (at the ‘ultimate’ level of analysis), Nagarjuna provides a ‘therapy’ to remedy a disquieting conception of philosophy and, on the other (at the ‘conventional’ level of analysis), he advocates for our Buddhist liberation. Instead, I propose that these spheres should be seen as one and the same, there being no difference between Nagarjuna’s Buddhist salvation and his philosophical ‘therapy’.

Whilst we should not underplay the extent to which Nagarjuna qua Mahayanist intends a degree of shift away from the emphasis of earlier Buddhist ideas like the Four Noble Truths, I think there is an important sense in which Nagarjuna’s ‘therapeutic’ intentions are actually aligned with these ideas. Therefore, the real genius of Nagarjuna is the way his philosophy is able to radically rework the relationship between Buddhism and philosophy, but without being forced to reject the central tenets of Buddhism altogether. This is not to say that Nagarjuna merely tolerates the ideas of his Buddhist forebears; rather, his method offers a means to protect these ideas from reification. That is, given that the existence of svabhāva is for it to possess precisely those features which the Four Noble Truths suggest are the source of our attachment and suffering—namely, a disquieting atomistic picture of the world qua discrete, unchanging and permanent entities—‘Emptiness’ acts as a mechanism by which the Four Noble Truths are prevented from becoming what they deny. Nagarjuna does this via preventing us seeing them as a once-and-for-all rejection of the atomistic picture above.Footnote 21 Therefore, we should see that the liberation offered by Buddhist teachings and the philosophical ‘therapy’ offered by his method are deeply interconnected. I think we can most clearly see this link at the finale of the MMK itself.

This halting of cognizing everything, the halting of hypostatizing, is blissful.

No Dharma whatsoever was ever taught by the Buddha to anyone. (MMK 25.24)

This is the crucial moment where Nagarjuna’s statement that the Buddha taught ‘No Dharma whatsoever’ (Ibid.) makes explicit the danger that Buddhist doctrines might be reified with the ‘bliss’ of our liberation. Thus, Nagarjuna affirms the core tenets of Buddhism, but only after that very same set of claims and commitments have been subjected to his own method (MMK 24.1-7).Footnote 22 This also explains why Nagarjuna applies the method of tetralemma (i.e. reductio ad absurdum) to his own apparent doctrines at MMK 22.11.

‘It is empty’ is not to be said,

nor ‘It is non-empty’,

nor that it is both,

nor that it is neither;

[‘empty’] is said only for the sake of instruction.Footnote 23

As Nagarjuna clearly saw, the key concepts of Buddhism are not immune to the danger of reification into philosophical theses; however, we should see that this by no means detracts from the value of Buddhist liberation. Similarly, the claim that the Buddha’s teachings do not have the status of ‘ultimate truth’ does not devalue them; after all, for Nagarjuna, there is nothing which could coherently have this status. However, I stress that this is not simply to adopt a kind of Fictionalism, that is, something like pretending that Buddhist teachings are ‘ultimately true’ (as in Garfield 2006). Instead, the question of whether or not (e.g.) the Four Noble Truths are ‘really true’ simply does not arise. This is because the question cannot be coherently answered; indeed, it cannot even be coherently asked.

However, at this point, it might be objected that my proposed reading of Nagarjuna’s conception of liberation looks as if it is reserved exclusively for that of philosophers. After all, it might seem unlikely that ‘everyone from cowherds on up’ (Tillemans 2011, 255)Footnote 24 continuously face (e.g.) the problems of higher order metaphysics on a daily basis. This of course would be a problematic way to read Nagarjuna, given that it would hardly be compatible with the Mahayanist Bodhisattva ideal of concern for the liberation of all. My answer to this objection is to understand the temptations Nagarjuna presents us with as a manifestation of more general fixations of aspect which we are all continually presented with and may lead us into disquiet.

As Rupert Read has said of the target of Candrakīrti’s philosophy, ‘The culprit for him is not confusion about language usage, but is “reification” (satyabhimana, bden ‘dzin), i.e., grasping things as being truly thus and so. This reification is not just a philosopher’s problem. For Candrakirti the ordinary man falls into “reification” too, and in a very important sense is even mixed up about the ordinary world’ (Read 2009, 242). This is not to make the implausible claim that philosophers’ confusions are not distinct from more everyday concerns; rather, that essentialism and the rejection of essence have their roots in preoccupations with seeing the world in particular ways that many of us fall into, regardless of our philosophical training. This is why Madhyamaka (qua Mahayana) is concerned with the liberation of all: the genesis of philosophical disquietudes should not be seen as originating somehow uniquely in the language of philosophical concepts, but rather (potentially) anywhere language is in use.

I note the importance of the fact that, for Nagarjuna, the will to theorise is itself only a symptom of a more general tendency towards the ‘reification’ (satyābhimāna) Read mentions, that is, the habit of mind in which we tend towards thoughts and language use in once-and-for-all terms. Whilst I think we are now at the limits of what a comparison with the Tractatus can offer Nagarjuna in terms of illuminating his philosophical method, I propose that the later Wittgenstein’s focus might be used to supplement my work here. Whilst I do think that the same method re: sensitivity to the way certain pictures take hold of us is at work in the Tractatus, I think that the Philosophical Investigations offers clearer examples of how it might be employed for less putatively philosophical concerns. In other words, comparison with the Philosophical Investigations and Nagarjuna might also shed light on a shared recognition of our fixation with ways of seeing more general than those concerned with our attitude to metaphysics. I suggest that this could help us see how Nagarjuna’s method and philosophy might be used in more everyday circumstances, offering new insights into how the non-philosopher might relate to other central Buddhist notions like the characteristic reappraisal of the Self not made perspicuous by comparison with the Tractatus alone.Footnote 25


Nagarjuna’s disavowal of views or doctrines has been described as ‘even more difficult to interpret than the famous last sentence of Wittgenstein’s Tractatus, which is preceded by the equally famous ladder-metaphor’ (Westerhoff 2009, 184). However, I think this mutual difficulty is not merely confined to the scholarly interpreters of their work; instead, the urge to interpret each of their words in several (antagonistic) ways is part of what is required to understand them at all. The temptations with which Nagarjuna and Wittgenstein present us in the MMK and the Tractatus, respectively, should be seen as representative of broader habits of mind in microcosm. That is, they are representatives of the kinds of impulse to which we are drawn to not only in philosophical analysis but elsewhere. Therefore, the ability to feel the pull of their apparent doctrines and then overcome them is both a requirement for understanding them but also the central lesson of each philosophy.

Particularly when concerning philosophers which such complex interpretive history, one of the fundamental challenges for any emergent reading of a philosophical work is to account for its novelty. What I suggest is the great advantage of the hermeneutic strategy I describe above is that it provides a way to explain the history of contradictory interpretations relating to the Tractatus and Nagarjuna’s philosophy. On this reading, contrary readings are recast as resulting from internal factors, an expected symptom of the sheer difficulty involved when grappling with the temptations presented in each work. However, we should make no mistake, to attach ourselves to what we feel is an attractive philosophical thesis in either the Tractatus or Nagarjuna’s philosophy is for us to fail to throw away the ladder after we have climbed up it. Or, as Nagarjuna puts it:

By a misperception of emptiness,

A person of little intelligence is destroyed.

Like a snake wrongly held

Or like a spell wrongly executed (MMK 24.11).Footnote 26

In conclusion, I think we can now understand the fundamental commonality between Madhyamaka and Wittgenstein’s Tractatus in a new way, namely, in terms of a wish to liberate us from a disquieting conception of philosophy via the deliberate (simultaneous) presentation of antagonistic philosophical impulses. For Nagarjuna and Wittgenstein, what they wish us to see is the way in which the contrary impulses rely on a presupposition concerning the coherency of theorising, thereby allowing us to navigate between both. However, this middle way can be perilous, and we should be aware that with metaphysical temptation, there is always the danger that we will fall under its spell; nothing is assured once-and-for-all. Just as Wittgenstein cautions us at with his ladder metaphor at TLP 6.54 to ‘throw away’ the apparent doctrines in his book, we must also be careful to see Nagarjuna’s apparent doctrines for the snakes which they are: something dangerous if mishandled. That is, something which, if we do not heed his warning, could even work against the true aim of Buddhist liberation. However, although we should be cautious with these snakes and ladders, for Wittgenstein and Nagarjuna, their therapy nonetheless has the power to liberate.Footnote 27