28
ŚĀNTARAKS․ITA
Climbing the Ladder to the Ultimate Truth
Allison Aitken
Introduction to Śāntaraksita’s Life and Works
The scholar-monk and prolific author Śāntarakṣita (c. 725–788)1 left a lasting and significant
impact on both Indian and Tibetan Madhyamaka Buddhist philosophy. He is known for his
synthesis of Nāgārjuna’s (c. second century) Madhyamaka with elements of Dignāga (c. 480–
540 CE) and Dharmakīrti’s (c. seventh century) tradition of logic and epistemology as well
as Yogācāra idealist ontology. Śāntarakṣita’s works are characterized by an emphasis on the
indispensable role of rational analysis on the Buddhist path as well as serious and systematic
engagement with competing Buddhist and non-Buddhist schools of thought.
Śāntarakṣita is said to have been abbot of the great monastic university of Nālandā in
Magadha (present-day Bihar, India) and counted Kamalaśīla (c. 740–795) and Haribhadra
(late eighth century) among his most prominent students. Yet aside from this, few biographical details about Śāntarakṣita’s life in India remain. He did, however, play a central role in the
early transmission (snga dar) of Buddhism to Tibet, and numerous semi-legendary reports
of his activities there survive. The earliest accounts agree that, upon receiving an imperial
invitation to Tibet from King Trisong Detsen (khri srong lde btsan) (742–797?), Śāntarakṣita
oversaw the establishment of the first Tibetan Buddhist monastery at Samyé (bsam yas),2 serving as its abbot and ordaining the first Tibetan Buddhist monks into the Mūlasarvāstivāda
monastic order (c. 779), whereupon he became known in Tibet as the “Khenpo (mkhan po)/
Ācārya Bodhisattva,” or “Abbot Bodhisattva.” According to Tibetan sources, Śāntarakṣita’s
own ordination lineage proceeds as follows: Śāriputra → Rāhula → Nāgārjuna → Bhāviveka
(c. sixth century) → Śrīgupta (c. seventh century) → Jñānagarbha (early eighth century) →
Śāntarakṣita.3 This lineage also reflects philosophical affinities among these authors, with the
later figures influenced by Bhāviveka’s interpretation of Nāgārjuna’s Madhyamaka (which
would later come to be known as *Svātantrika-Madhyamaka).4
A sizeable corpus is attributed to Śāntarakṣita, spanning a range of genres and subject matters,
including Madhyamaka metaphysics, logic and epistemology, Buddhist path literature, tantra,
as well as several praises. His two most important independent treatises are the Compendium of
True Principles (Tattvasaṃgraha, hereafter Compendium = TS) and the Ornament of the Middle Way (Madhyamakālaṃkāra, hereafter Ornament = MA) together with an autocommentary
(Madhyamakālaṃkāravṛtti = MAV). Kamalaśīla authored lengthy commentaries ( pañjikā-s)
463
DOI: 10.4324/9781351030908-41
Allison Aitken
on both the Compendium (TSP) and the Ornament (MAP). In practice, Śāntarakṣita’s basic
texts (particularly the Compendium) are standardly read with the aid – and thus through the
lens – of Kamalaśīla’s commentaries, such that Śāntarakṣita’s thought and that of his principal
student are often inextricable.
Several notable commentaries are also attributed to Śāntarakṣita, including one on Jñānagarbha’s
The Distinction Between the Two Truths and its Autocommentary (Satyadvayavibhaṅga and
Satyadvayavibhaṅgavṛtti = SDV and SDVV, Satyadvayavibhaṅgapañjikā = SDVP), as well as
a commentary on Dharmakīrti’s The Logic of Debate (Vādanyāya) titled Commentary on the
Logic of Debate: Elucidation of Its Meaning (Vādanyāyaṭīkā Vipañcitārthā).5 Of all the works
attributed to Śāntarakṣita, only the Compendium, his commentary on Dharmakīrti’s The Logic
of Debate, and a tantric-cum-epistemological work, Establishing the Truth (Tattvasiddhi),6 survive in Sanskrit. Śāntarakṣita’s references to his own works yield the following chronology
of composition: Ascertainment of the Ultimate, Compendium, Commentary on the Logic of
Debate, Commentary on the Ornament. This suggests that the Commentary on the Ornament
represents his most mature thought and is his definitive work on Madhyamaka.7
As noted previously, Śāntarakṣita’s place in the history of Madhyamaka philosophy is perhaps most remarkable for his synthesis of Nāgārjuna’s Madhyamaka with elements from the
Dignāga-Dharmakīrti tradition of logic and epistemology together with Yogācāra idealist ontology. Though not the first Mādhyamika to be influenced by Dignāga and Dharmakīrti, Śāntarakṣita
looks to be the first Mādhyamika to author a commentary on one of Dharmakīrti’s works. The
influence of the Dignāga-Dharmakīrti tradition of epistemology, according to which testimony
as a source of knowledge is reduced to a form of inference, is reflected in Śāntarakṣita’s emphasis on the central role of rational analysis in the gradual progression toward a correct metaphysical view. As Śāntarakṣita repeatedly suggests, the ideal reader of his works and the ideal trainee
on the Buddhist path is a discerning person ( prekṣāvat), that is, a rational epistemic agent.8
Prior to Śāntarakṣita, Śrīgupta is noteworthy for integrating the Dignāga-Dharmakīrti
tradition of logic and epistemology into his presentation of Madhyamaka, but Śrīgupta
rejects Yogācāra ontology without qualification.9 And while Jñānagarbha subsequently
alludes to Yogācāra conceptual frameworks in his presentation of the Madhyamaka theory
of two truths (satyadvaya), viz. the conventional truth (saṃvṛtisatya) and the ultimate truth
( paramārthasatya) (see, e.g., SDVV ad SDV 30), it is Śāntarakṣita who explicitly formalizes
the incorporation of Yogācāra ontology into his presentation of Madhyamaka, though relegated to the domain of conventional truth. Śāntarakṣita accepts as conventionally true not only
the Yogācāra doctrine that apparently external objects are merely mental in nature (cittamātra)
(MA 91–93) but also the Yogācāra claim that cognition is reflexively aware (svasaṃvitti/
svasaṃvedana) (MAV ad MA 91). Following Śāntarakṣita, Kamalaśīla and Haribhadra likewise adopt the Dignāga-Dharmakīrti tradition of logic and epistemology while also taking a
conciliatory approach to Yogācāra. Below, we will return to the question of how best to understand Śāntarakṣita’s Yogācāra-Madhyamaka synthesis.
Both of Śāntarakṣita’s main treatises, the Compendium and the Ornament, exemplify his
wide-ranging and systematic critical engagement with the philosophical views of competing
Buddhist and non-Buddhist schools of thought. The Compendium – over 3,600 stanzas organized in twenty-six chapters – in some ways resembles a critical doxographical survey of the
eighth-century Indian religio-philosophical landscape, yet such a description does not adequately
reflect its dialogical structure or its in-depth engagement with these competing systems.10 The
first twenty-three chapters of the treatise examine and ultimately reject a succession of cosmogonical theories, ontological categories, semantic theories, epistemological theories, and candidate
sources of knowledge advanced by Nyāya, Vaiśeṣika, Mīmāṃsā, Sāṃkhya, Jaina, Vedānta, and
464
Śāntaraks∙ita
Lokāyata/Cārvāka traditions, as well as Buddhist traditions such as the Vātsīputrīya.11 The final
three final chapters that constitute nearly the second half of the treatise are largely aimed at
Mīmāṃsakas, first rejecting their claim that the Vedas lack a human author (apauruṣeya), next
critically examining their theory that veridical cognition is self-certified (svataḥprāmān
․ ya), and
finally concluding with an argument in support of the possibility of omniscience. Given the
breakdown of the text, the Compendium would seem primarily concerned with competing theories of the Nyāya and Mīmāṃsā, and Śāntarakṣita makes important contributions in formulating
Buddhist responses to the Naiyāyika philosopher Uddyotakara (fl. c. 600), as well as the PūrvaMīmāṃsaka Kumārila Bhaṭṭa (fl. c. 660), both of whom were influential critics of Dignāga.
The Ornament similarly rejects a range of positions from competing Buddhist and nonBuddhist systems, though it is more metaphysical in its focus than the Compendium, with
roughly two-thirds of the ninety-seven stanzas devoted to the neither-one-nor-many argument
(ekānekaviyogahetu), which sets out to demonstrate that nothing possesses an intrinsic nature
(svabhāva). While the argument takes aim at entities advanced by many of the same nonBuddhist traditions addressed in the Compendium, the Ornament is concerned foremost with
competing Buddhist views.
In presenting and rejecting the views he addresses in both the Compendium and the Ornament, Śāntarakṣita utilizes a dialectical/pedagogical device of provisionally adopting what
he deems to be successively more rational positions in order to reject less rational ones. Sara
McClintock has influentially described this method whereby Śāntarakṣita argues from progressively shifting perspectives as a “sliding scale of analysis.”12 To illustrate, in the Ornament,
Śāntarakṣita adopts the Sautrāntika representationalist theory of perception in order to reject
the Vaibhāṣika direct realist theory of perception. He then assumes the Yogācāra idealist position on which mental representations have no external referents in order to reject Sautrāntika
representationalism. Finally, he uses the Madhyamaka account of the nonexistence of fundamentally real cognition to reject Yogācāra theories of the mind and mental content.
Although the Compendium includes several allusions to the superiority of the Madhyamaka
perspective,13 it might be read as culminating in the Yogācāra perspective. By contrast, the
Ornament, in which Śāntarakṣita presents his definitive account of Madhyamaka, devotes
more critical attention to Yogācāra than to any other competing system. Yet it is in this same
text that Śāntarakṣita presents his provisional endorsement of Yogācāra idealism on the level
of conventional truth. The following sections will take up Śāntarakṣita’s contributions to the
Madhyamaka theory of two truths in the Ornament.
Ultimate Truth and the Neither-One-Nor-Many Argument
The Madhyamaka central commitment, or ultimate truth, is the emptiness of intrinsic nature
(svabhāvaśūnyatā), which might be glossed as the universal negation of ontological independence. In other words, according to Mādhyamikas, nothing lays claim to ontological selfsufficiency, which is commonly identified as a necessary condition for fundamentality and
substancehood. The Madhyamaka view thus might be described as a kind of thoroughgoing
anti-foundationalism as well as a form of substance nihilism. But if there are no ontologically
independent or fundamental entities, then whatever there is depends for both its nature and its
existence on something else.
In his Ornament, Śāntarakṣita makes a major contribution to the Madhyamaka canon of
arguments for emptiness with his presentation of the neither-one-nor-many argument. Although
he expands on his predecessor Śrīgupta’s more condensed formulation of the argument in the
Introduction to Reality, it is Śāntarakṣita’s influential Ornament that popularizes the argument
465
Allison Aitken
in both Buddhist India and Tibet. In his Illumination of the Middle Way (Madhyamakāloka),
Kamalaśīla presents the neither-one-nor-many argument among a set of five Madhyamaka
arguments for emptiness, which subsequently became known in Tibet as the “five great arguments” for emptiness (gtan tshigs chen po lnga).
Nāgārjuna articulates an early precursor to the neither-one-nor-many argument in his Precious Garland (Ratnāvalī), stating:
Something is not a unity if it has multiple loci. There is nothing that lacks multiple
loci. In the absence of any unity, neither is there a multiplicity.
(Precious Garland 1.71)
Here, Nāgārjuna argues that whatever is divisible into multiple discrete spatial or temporal
loci does not count as a true unity. And everything, he claims, is so divisible. Just as each bit
of matter – regardless of how minute – has a right side and a left side, a top and a bottom, and
so on; likewise, each moment of time – no matter how brief – has a beginning, a middle, and
an end. Otherwise, the existence of the spatially and temporally extended ordinary objects
that populate our world – like computers, kangaroos, and cognitions – would be impossible.
After all, the thought goes, how could fundamental building blocks that lack spatial/temporal
extension ever yield anything that has spatial/temporal extension? And since a plurality presupposes unities as its basic constituents, if there is nothing that is truly one, then neither is
there anything that is truly many.
Śrīgupta's expanded formulation of the neither-one-nor-many argument makes explicit the
implication that nothing has intrinsic nature, while also formalizing the argument and defending its soundness according to the Dignāga-Dharmakīrti tradition of logic and epistemology.
Closely following Śrīgupta’s Introduction to Reality (TA 1), Śāntarakṣita articulates the central
inference of the neither-one-nor-many argument in the opening stanza of the Ornament as
follows:
In reality, everything that is theorized by our own and other schools of thought is
without intrinsic nature, due to lacking an intrinsic nature that is either one or many,
like a reflection.
(MA 1)
The argument poses a destructive dilemma, which says: if anything has an intrinsic nature,
then it is either one or many. Śāntarakṣita, in effect, argues that nothing can satisfy either disjunct of the consequent and therefore, by modus tollens, that nothing can satisfy the antecedent. Upon analysis, nothing possesses an intrinsic nature.
Three features of the disjunctive property pair <one or many> are critical for the argument
to go through. First, the terms translated as “one” and “many” here (eka and aneka in Sanskrit)
are perhaps more precisely rendered as “unity” and “non-unity,” reflecting the fact that they
are a mutually exclusive and contradictory pair, conforming to the conceptual, logical, and
grammatical structure F and not-F. As Śāntarakṣita makes clear, if anything had an intrinsic
nature, then on pain of violating the law of excluded middle, it would have to either be a unity
or non-unity:
Aside from unity and not-unity, an object’s having some other classification is impossible, since it is established that these two properties are mutually exclusive.
(MA 62)
466
Śāntaraks∙ita
Second, the operator, or qualifier, “in reality” (tattvataḥ) in the statement of the central inference clarifies that the target here is a true unity and a true multiplicity. This should be contrasted with a merely conventional status, like the unity of an aggregate such as a flock of
sheep or a heap of sand. And, as indicated by Nāgārjuna, a true unity is defined as a mereological simple, that is, something that lacks proper parts, where something has proper parts just in
case it is either physically or conceptually divisible. While physical divisibility is more or less
straightforward, we can understand that x is conceptually divisible in the mereological sense
just in case there are conceptually isolatable proper parts ys that compose x, such that x is the
sum of the ys. A true multiplicity, then, is something that has proper parts, the most basic of
which are themselves true unities.
Finally, a third feature of this property pair is that, unlike most contradictories, unity and
non-unity share not only a conceptual priority relation but also a metaphysical priority relation: the existence of a non-unity presupposes the existence of some unities. As Nāgārjuna
pointed out in his Precious Garland, a plurality requires singular things as its building blocks.
Śāntarakṣita explains,
Thus, a “multiplicity” is defined as a composite of unities. If no unity exists, neither
does a multiplicity, just like if no trees exist, neither does a forest.
(MAV ad MA 61)
But if a multiplicity depends for its existence on some unities just like a forest does on some
trees, then a multiplicity is not a candidate for ontologically independent being after all. As
it turns out, then, true unity is a necessary criterion for ontological independence. And this
should not be so surprising: just like ontological independence, true unity is commonly cited
as a necessary condition for fundamentality as well as for substantial reality. The neither-onenor-many argument thus reduces to a rejection of true unities, which is to say a rejection of
mereological simples.
The Neither-One-Nor-Many Argument Against Material Simples
Śāntarakṣita’s rejection of material simples closely follows Vasubandhu’s (c. fourth–fifth century) anti-atomist argument in his Yogācāra work, the Twenty Verses (Viṃśikā 11–15). This
section of his neither-one-nor-many argument also features in the “Examination of External
Objects” (Bahirarthaparīkṣā) chapter of his Compendium, wherein he assumes the Yogācāra
perspective. Śāntarakṣita targets three kinds of views about how atoms aggregate to constitute
composites, which recur in debates of this kind in pre-modern Indian philosophy:
i. Each atom conjoins with surrounding atoms.
ii. Atoms have interceding space between them.
iii. Atoms are spatially continuous, neither conjoining with surrounding atoms nor having
interceding space between them.14
To each of these views, Śāntarakṣita, in effect, poses the following dilemma: If matter is constituted by fundamental, simple particles, then those particles either face surrounding particles
at one and the same locus or at spatially differentiable loci. If, on the one hand, fundamental
particles did not have spatially differentiable loci at which to face neighboring particles, and
were thus spatially unextended, then they could not compose an extended composite. If, on
the other hand, fundamental particles did have spatially differentiable loci at which to face
467
Allison Aitken
surrounding particles (e.g., a right side, a left side, etc.), then they would have spatially discrete parts, which means that they would be composites themselves and could not be fundamental unities after all. As Śāntarakṣita argues in both his Ornament and Compendium,
Whether atoms are (i) conjoined, (ii) located at a distance from one another, or (iii)
located continuously without interceding space, if the very same part of the central
atom in a composite which is facing one atom were also imagined to be facing another
atom, then the aggregation of atoms composing mountains, etc., would not be feasible.
(TS 1989–90 = MA 11–12)
If instead it were accepted that a different part of the central atom faced another
atom, then how indeed could an atom like that, i.e., with distinct parts facing different
atoms, be truly unitary?
(TS 1991 = MA 13)
He concludes that there is no account on which matter could be founded in simple particles.
And given the metaphysical priority of true unities to a true multiplicity, in the absence of
material simples, a material multitude is also precluded.
The Neither-One-Nor-Many Argument Against Mental Simples
While the Compendium restricts the subject of the neither-one-nor-many argument to external
objects, the Ornament grants the argument a universal scope of application. Having rejected
the true unity of any extramental entities within the first fifteen stanzas of the Ornament,
Śāntarakṣita devotes stanzas 16–60 to rejecting the true unity of the mind, addressing a variety
of Buddhist and non-Buddhist accounts of the mind and mental content, with the argument
culminating in a sixteen-stanza section targeting Yogācāra theories. Śāntarakṣita introduces
this section, remarking:
Even though the Yogācāra view has merit, we shall consider whether such mental
entities are to be accepted as real or as satisfactory only when not analyzed.
(MA 45)15
The succeeding argument turns on an analysis of the relation between the mind and mental
content qua cognition ( jñāna) and mental representations (ākāra).
Śāntarakṣita targets two families of views from the Yogācāra tradition on the ontological
status of mental representations:
i.
ii.
Representational realism (*satyākāravāda): representations are real in the same way as
cognition is taken to be.16
Representational antirealism (*alīkākāravāda): representations are unreal figments.17
It is important to keep in mind that for Yogācārins, who reject mind-independent material
objects, a representation does not actually represent any extramental entity but is simply the
intentional object of a cognition. Thus, the question of the ontological status of representations
concerns not the represented content (like a desk or a dragon) but rather the representation
itself as a feature of the mind.18 Representational realism is commonly associated with a second
claim which says that cognition is necessarily and intrinsically endowed with representations
468
Śāntaraks∙ita
(sākāra), while representational antirealism is commonly associated with the claim that cognition is not necessarily endowed with representations, and that invariably veridical enlightened
cognition lacks representations (nirākāra).19
In addressing the first view, on which cognition is intrinsically and necessarily endowed
with real representations, Śāntarakṣita takes it that, according to this theory, a cognition and
its representation are non-distinct, constituting a single subject. Supposing that representations and cognition share a strict identity relation, his reasoning here turns on a version of the
law of noncontradiction according to which contradictory properties cannot be predicated of
the same subject. Śāntarakṣita observes that a moment of cognition seems to be indisputably
simple, and yet the content of cognition looks obviously complex; in any given moment, ordinary experience presents us with a multiplicity of data – a white patch here, a blue patch there,
and so on. And, indeed, perhaps the most intuitive representational realist view (the so-called
citrādvaita, or “variegated nonduality” theory) says that unitary cognition is non-distinct from
its multifaceted representation. But, Śāntarakṣita argues, given the law of noncontradiction, if
cognition and representations are non-distinct, it follows that:
i.
ii.
since a representation is manifold, so too is cognition, or else
since cognition is truly unitary, so too is its representation.
Śāntarakṣita lays out this argument from dilemma as follows:
It is difficult to deny that: (i) due to being non-distinct from real representations, cognition must accord with the nature of representations and therefore be manifold. Or
else, (ii) due to being non-distinct from unitary cognition, representations would have
to be unitary in accord with the nature of cognition. On account of having contradictory properties, ultimately, representations and cognition would have to be distinct.
(MAV ad MA 46)
Śāntarakṣita takes up the second horn of the dilemma first, arguing that if a representation were
truly unitary in accord with cognition, then absurd consequences would follow. For instance,
since a simple representation could not be analytically divisible into phenomenal proper parts, we
would be unable to conceptually isolate different aspects of our phenomenal field (like the right
side and left side of this page). Furthermore, in the absence of phenomenal parts, it could never be
the case that one element of our experience was in motion while another was at rest (MA 47–48).
On the other hand, Śāntarakṣita reasons, the alternative that cognition is manifold in accord
with its complex representation is susceptible to the same kind of argument that he leveled
against material atomism: just as an extended material object could not be constituted by unextended material simples, a phenomenally extended representation (like the one you may have
of this page) could not be composed of phenomenally unextended building blocks (MA 49).20
And given the metaphysical priority of unity to multiplicity, if there are no simple phenomenal
parts, neither can there be a true multiplicity of them. The parts of cognition, then, could not
exist in numerical parity with representational parts, since there can be no determinate number
of them to which cognition might correspond. He thus concludes that cognition and a real
representation could be neither truly one nor truly many.
Śāntarakṣita next turns to the representational antirealist view on which cognition is not
actually endowed with real representations, which only seem to appear to cognition due to an
error (MA 52). This view, Śāntarakṣita argues, is incapable of accounting for ordinary experience, for how could we perceive anything if no percept exists (MA 53–54)? Indeed, he insists
469
Allison Aitken
that there can be no cognition at all in the absence of an intentional object, since cognition is
intentional by its very nature; to cognize is to have a cognition of something (MA 55). Moreover, Śāntarakṣita argues, an unreal representation could stand in no relation whatsoever with
cognition, whether that be an identity relation or a causal relation (MA 57). If a representation
stood in an identity relation with cognition, then given the law of noncontradiction, either:
i.
ii.
since cognition is real, the representation too would be real, or else
since the representation is unreal, cognition too would be unreal (MAV ad MA 57).
Neither alternative is admissible for the representational antirealist. Furthermore, an unreal
representation is no more capable of standing in a causal relation with cognition than an identity relation, since if a representation were caused, then it would be real, but if it had no cause,
then there could be no explanation for its appearing with spatiotemporal determinacy or consistency (MA 58). Unable to get the semblance of a defeasible account of cognition up and
running on this view, Śāntarakṣita does not even bother to apply the neither-one-nor-many
analysis to the representational antirealist position.
Having dismissed both the representational realist and antirealist views as untenable,
Śāntarakṣita takes himself to have demonstrated that there is no account in which a mental
entity could exist as a true unity or a true multitude, and thus concludes that nothing – whether
material or mental – lays claim to ontologically independent being. Instead, the only kind
of unity and being that exist are conventional and dependent. To flesh out precisely what
Śāntarakṣita means by this, let’s turn to the second of the two truths.
Conventional Truth and Yogācāra Ontology
Upon concluding his neither-one-nor-many argument for the ultimate truth, Śāntarakṣita presents his account of conventional truth to clarify that the rejection of ontological independence
does not entail an unmitigated nihilism. The term satya translated here as “truth” has a semantic range that is also inclusive of “reality,” and Śāntarakṣita’s account of the conventional is
a theory of truth as well as an ontological theory. Yet Mādhyamikas, like Śāntarakṣita, affirm
only an ultimate truth and not an ultimate reality qua ontological status. In fact, the ultimate
truth as the emptiness of intrinsic nature might be interpreted as the claim that nothing is ultimately real.21 There is thus only one ontological status: conventional reality.
According to Śāntarakṣita, whatever is conventionally real (i) has the capacity for causal efficacy (arthakriyāśakti/arthakriyāsamartha), (ii) is dependently originated ( pratītyasamutpanna),
and (iii) satisfies our ordinary notions of unity and being only when not subjected to analysis into
its final nature (avicāraraman
․ īya/avicāramanohara) (MA 64). Conventional truths, then, are
pragmatically efficacious claims that concern conventionally real things and which may be verified by our epistemic instruments of perception and inference. With this account, Śāntarakṣita
once again follows his predecessor Śrīgupta, who presents the earliest extant formulation of this
threefold criterion,22 which was subsequently adopted by Jñānagarbha, Kamalaśīla, Haribhadra,
the later Bhāviveka (c. eighth century), Atiśa (982–1054), and others.23 The first criterion for
conventional reality – having the capacity for causal efficacy – is a repurposing of Dharmakīrti’s
criterion for ultimately reality.24 Though an apparent subversion of Dharmakīrti’s intent, this
criterion represents yet another Dharmakīrtian influence on this branch of the Madhyamaka
tradition. We will return to the third criterion in treating the role of analysis in Śāntarakṣita’s
account of conventional truth below, but it is with the second criterion, being dependently originated, that Śāntarakṣita incorporates Yogācāra ontology into his system.
470
Śāntaraks∙ita
The claim that whatever is conventionally real comes into being in dependence on other
things goes back to Nāgārjuna.25 The ontological dependence relation implicated here is inclusive of mereological dependence, mind-dependence, as well as causal dependence, such that
every conventionally real thing: (i) comes into being in dependence upon its parts and those
parts upon their own parts, ad indefinitum, (ii) is individuated as a conventional unity in
dependence upon mental designation, and (iii) is a product of causes and conditions, each
of which is in turn a product of its own causes and conditions, ad indefinitum. But, marking
a significant departure from prior Madhyamaka accounts, Śāntarakṣita identifies all things
involved in causal relations as mental in nature, thereby aligning his presentation of conventional reality with Yogācāra ontology and its central commitment that everything consists in
cognition alone (vijñaptimātra):26
Whatever exists as cause and effect is, in fact, merely cognition ( jñāna). Whatever is
established by cognition itself exists in cognition.
(MA 91)
In order to understand how Śāntarakṣita understands Yogācāra ontology to map onto conventional reality, it is necessary to first pin down what precisely he means by affirming the
Yogācāra commitment that everything is merely mental (cittamātra). Here are some possible
interpretations:
i.
ii.
A phenomenological claim on which the only things relevant to our experience are mental
A kind of skepticism which says that we cannot know whether or not there exist any extramental entities
iii. An epistemological idealism which says that all objects of knowledge are determined by,
or dependent on, the mind and the structure of thought
iv. An immaterialism, or metaphysical idealism, on which there are no material things, and
the only kinds of things that exist are mental
The strongest claim, (iv) immaterialism, is an eliminative idealism insofar as it effectively
eliminates, or precludes the existence of, extramental things. The former three options are
varieties of non-eliminative idealism insofar as they grant some kind of primacy to the mental
but leave open the possibility that extramental things exist.
Some have argued for a version of the (i) phenomenological reading of Śāntarakṣita’s
Yogācāra by pointing out that he seems to identify conventionally real things with appearances and thus exclusively with what lies within the domain of experience.27 Indeed, in a rather
customary Madhyamaka move following an argument for emptiness, Śāntarakṣita insists that
the conclusion of his neither-one-nor-many argument does not entail the denial of appearances (MA 78ab). To do so would be tantamount to an implausible thoroughgoing nihilism.
But Śāntarakṣita’s nondenial of appearances should not be read as an anti-metaphysical, phenomenological turn; nor does it serve to restrict of the scope of knowledge to the domain of
appearances, indicative of skepticism along the lines of view (ii). After all, he rejects as irrational and untenable both direct realist and representationalist theories of perception, which
suppose mind-independent external objects to be the direct and indirect objects of perception,
respectively. And he does not simply argue that such objects are unknowable or irrelevant to
our experience. Rather, he insists that the existence of external objects founded in atoms is
incoherent and thus metaphysically impossible. And just as Śāntarakṣita’s arguments rejecting
a substantial self and a creator god, for example, are not intended to leave the back door open to
471
Allison Aitken
the existence of such entities lying beyond the scope of our experience, ordinary cognition, or
the reach of human knowledge, presumably the same is true of his rejection of material objects.
That Śāntarakṣita’s Yogācāra is best read not as (iii) an epistemological idealism (in which all
objects of knowledge are determined by, or dependent on, the mind) but as an eliminative (iv)
metaphysical idealism is supported by his commentary on MA 91, where he states in no uncertain
terms that external material objects do not exist and what we ordinarily take to be external objects
simply are one’s own mind. This agrees with the Compendium chapter on the “Examination of
External Objects,” wherein Śāntarakṣita assumes the Yogācāra perspective. Here, he lays out two
main lines of reasoning against external objects, the first an epistemological argument against the
possibility of having knowledge of the existence of external objects founded in atoms and the
second an argument against the metaphysical possibility of such objects.28 He transitions from
the epistemological to the metaphysical argument with reference to the discerning epistemic
agent ( prekṣāvat), once again gesturing to the prominent place of rational analysis in his system:
Be as it may that atoms are not established by any source of knowledge, there may
nonetheless be doubt. But how should a discerning person come to have certainty
about their nonexistence?
(TS 1988)
Śāntarakṣita aims to pull the rug out from under the external world by rejecting the existence
of its purported foundations: material simples.29 That Śāntarakṣita takes the Yogācāra mindonly thesis to negate the existence (and not merely the epistemic accessibility) of real external
objects is reiterated in his Yogācāra formulation of the neither-one-nor-many argument:
Thus, it is appropriate for discerning individuals to ascertain that atoms are nonexistent, due to being empty of an intrinsic nature that is either one or many, like a
lotus in the sky.
(TS 1996)
According to Śāntarakṣita’s Yogācāra, atoms – the purported building blocks of material
objects – are no more real than lotuses growing in midair.
Śāntaraksita’s Madhyamaka-Yogācāra Synthesis: An Instrumentalist
Approach to the Ultimate
With Śāntarakṣita’s characterization of Yogācāra ontology in place, let us turn now to the
question of how precisely to understand his Yogācāra-Madhyamaka synthesis. That is, does
Śāntarakṣita propose a genuine and coherent unification of these two systems, or does he
have an instrumentalist story in mind? Given his identification of the conventional truth with
Yogācāra ontology, this question is tied up with how best to understand the relation between
the two truths, that is, between metaphysical idealism as the conventional truth and the universal negation of ontological independence as the ultimate truth. Are these two truths compatible
or contradictory?
One Madhyamaka story about the relation between the two truths sees them as contradictory inasmuch as (i) conventional truths are true according to the non-veridical beliefs and
linguistic-conceptual norms of ordinary folks whose understanding is obscured by metaphysical ignorance, while (ii) the ultimate truth reflects the veridical cognition of an ideal epistemic
agent. But this is not Śāntarakṣita’s story. The radically unintuitive claim that all things are
472
Śāntaraks∙ita
merely mental in nature is plainly difficult to square with the commonly accepted view of
the world.30 The conventional truth of metaphysical idealism can hardly be “read off ” our
linguistic-conceptual norms. For Śāntarakṣita, the general consensus is not a guide to what is
conventionally true or real. Instead, he insists that a discerning person should seek a correct
understanding of conventional reality by employing analysis; rationality is king as a guide to
both the conventional and the ultimate.
Another Madhyamaka story about the two truths sees them as perfectly consistent and even
synonymous. On this view, the ultimate truth as the universal negation of ontological independence is the obverse (viz. negative equivalent) of the conventional truth as the universal
affirmation of dependent origination. Yet Śāntarakṣita does not see the two truths as wholly
compatible either. This is evident from a comparative analysis of his Yogācāra and Madhyamaka versions of the neither-one-nor-many argument in the Compendium and Ornament. Not
only are the subjects of the two arguments different (the Yogācāra iteration takes up atoms and
the Madhyamaka argument concerns all things), but the predicate is also different, yielding
distinct inference warranting entailment relations (vyāpti):31
Yogācāra entailment relation: Whatever is neither-one-nor-many does not exist.
Madhyamaka entailment relation: Whatever is neither-one-nor-many lacks an intrinsic nature.
How can Śāntarakṣita consistently maintain that this same neither-one-nor-many reason establishes the non-existence of atoms in his Yogācāra iteration of the argument and the absence
of an intrinsic nature of all things in his Madhyamaka formulation? If, on the one hand, the
Yogācāra entailment relation holds, then his Madhyamaka neither-one-nor-many argument –
which applies this same reason to an unrestricted domain – commits him to a thoroughgoing nihilism. If, on the other hand, the Yogācāra entailment relation does not obtain, then
Śāntarakṣita advances a fallacious argument in support of his account of conventional truth.
Śāntarakṣita must surely reject the first alternative; nihilism is not an option. He thus looks
committed to the second horn of the dilemma, that his argument in support of his view of
conventional truth is fallacious. Yet this dilemma stands only if Śāntarakṣita intends to hold
the Yogācāra and Madhyamaka perspectives simultaneously.32 But, as he points out, once one
has realized the Madhyamaka ultimate truth that all things are equally devoid of ontological independence, one necessarily relinquishes the Yogācāra ontology that grants a privileged
position to the mental:
Those whose intellectual capacity is not slight and particularly those who are highly
industrious will, upon analyzing whether the mind has a unitary or manifold nature,
ultimately perceive no such entity. Thus, in reality the mind-only view is not accepted.
(MAV ad MA 92)
But why bother reasoning our way to metaphysical idealism if it is not ultimately true? According to Śāntarakṣita’s sliding scale of analysis, there are better and worse conventional truths,
with more rational stories supplanting the less rational. But there is no definitive conventionally true story. In setting up his characterization of cause and effect as mental in nature in MA
91, Śāntarakṣita states:
Whoever accepts the conventional reality of those things that stand in causal relations should analyze what those accepted conventional things are in order to respond
473
Allison Aitken
to fallacious arguments: Are they merely in the nature of the mind and mental constituents or is their nature also extramental?
He goes on to explain that among Mādhyamikas, there are two opinions on the matter. The
first camp, exemplified by Bhāviveka, accepts conventionally real material and mental entities
alike. He introduces the second view – that cause and effect are merely mental – as simply “the
opinion of others.” Significantly, he does not reject the first opinion. He does, however, explain
the pedagogical utility of the latter. Śāntarakṣita thus sets out an instrumentalist account of
the relation between the two truths on which Yogācāra metaphysical idealism is provisionally
accepted as conventionally true as an expedient means to arrive at an understanding of the
Madhyamaka ultimate truth.
Śāntarakṣita’s conventional truth is not determinate but contextual, as indicated by the third
criterion for conventional truth that it does not withstand analysis. Importantly, Śāntarakṣita
does not follow Candrakīrti (Introduction to the Middle Way, Madhyamakāvatāra 6.35) or
Jñānagarbha (SDV 21) in articulating this criterion as the normative claim that one ought
not to analyze conventional truths. This would be incompatible with Śāntarakṣita’s progressive path, which the discerning person traverses precisely by analyzing conventional truths.
Instead, following Śrīgupta, Śāntarakṣita simply claims that conventional truths satisfy when
not analyzed; a conventionally real thing satisfies our notions of reality, independence, and
unity when its ultimate nature is not subjected to analysis. We may arrive at progressively
more rational conventional truths through analysis, but there is no final conventionally true
theory; metaphysical inquiry into the nature of things has no termination point.
Still, if there is no determinate conventional truth, what makes one conventionally true theory
more rational than another? Śāntarakṣita implies that since nothing withstands analysis, a theory
that posits fewer ontologically independent entities is more rational than one that posits more insofar as it is closer to the ultimate truth. In both his Compendium and Ornament, Śāntarakṣita uses
analysis to gradually eliminate ontological categories, with the Yogācāra sparse ontology of “mindonly” being, as it were, the last man standing. But at the end of the day, this category too does not
withstand analysis. The primary utility of Yogācāra isn’t in what it affirms but in what it denies.
From an idealist, rather than a dualist view of conventional reality, it simply takes fewer
steps to arrive at an understanding of the ultimate truth, which undermines the fundamentality
of anything – whether material or mental. Śāntarakṣita introduces his Yogācāra-Madhyamaka
synthesis, stating:
Based on the perspective of the mind-only system (cittamātra), one should understand that there are no real external objects. Based on this Madhyamaka system, one
should understand that the mind too is utterly selfless.
(MA 92)
Those who hold the reins of rationality while riding the chariot of these two systems
will thereby achieve the state of a genuine proponent of the Mahāyāna.
(MA 93)
Śāntarakṣita recommends approaching the ultimate truth via the Yogācāra view, but it is just
that – an approach to the ultimate, not a definitive or unrevisable claim about the final nature
of things. Thus, on this picture, one cannot definitively claim that the two truths are either
compatible or contradictory. The conventional truth is a moving target on shifting sands. That
Yogācāra idealism is just one among other instructively efficacious conventional stories by
474
Śāntaraks∙ita
way of which one might arrive at the ultimate is supported by the following remark from
Kamalaśīla’s Illumination of the Middle Way:
Thus, one who is unable to instantaneously realize the fact that all things without
exception lack an intrinsic nature should, by temporarily relying on the mind-only
system, proceed in stages, beginning with understanding that external objects lack an
intrinsic nature.
(D 3887, 157a)
In this same instrumentalist spirit, Śāntarakṣita (MAV ad MA 70) cites Bhāviveka’s famed
metaphor of conventional truth as a ladder to the ultimate truth:
Without the ladder of conventional truth, it would not be possible for the learned to
ascend to the pinnacle of the palace of reality.
(Verses on the Heart of the Middle Way, Madhyamakahṛdayakārikā 3.12)
In Śāntarakṣita’s presentation of the two truths, the Yogācāra idealist ontology – though not the
exclusive, determinate, or preeminent conventional truth – is nonetheless a highly efficacious
penultimate steppingstone to understanding the Madhyamaka ultimate truth.
Conclusion
Śāntarakṣita remained a Madhyamaka authority to be reckoned with well into the final period
of Buddhism in India. His direct students, Kamalaśīla and Haribhadra, adopted and developed his Madhyamaka-Yogācāra synthesis, and he was recognized as a principal source on
the Madhyamaka neither-one-nor-many argument throughout the succeeding centuries by
authors such as Dharmamitra (fl. ca. 800), Jitāri (late tenth century), Bodhibhadra (fl. c. 1000),
Prajñākaramati (ca. 950–1030), Atiśa, Prajñāmokṣa (ca. eleventh century), Abhayākaragupta
(late eleventh–early twelfth century), and so on. Later prominent Yogācāra philosophers including Ratnākaraśānti (ca. eleventh century) and Jñānaśrīmitra (ca. eleventh century) also cited
and responded to Śāntarakṣita’s neither-one-nor-many argument, indicating that they considered his attack on Yogācāra theories of the mind and mental content a serious enough threat to
necessitate critical engagement several hundred years later. In Tibet, Śāntarakṣita’s Ornament
together with Jñānagarbha’s Differentiation of the Two Truths and Kamalaśīla’s Illumination
of the Middle Way came to be known as the major works of the so-called “three Mādhyamikas
of the East” (dbu ma shar gsum), with commentaries composed on the Ornament by such
philosophically and temporally diverse luminaries as Chapa Chökyi Senggé (Phya pa Chos
kyi seng ge, 1109–1169), Tsongkhapa Lobsang Drakpa (1357–1419), and Jamgön Ju Mipham
Gyatso (1846–1912). Śāntarakṣita’s emphasis on the role of rationality and a progressive path,
as well as his synthesis of Madhyamaka with the Dignāga-Dharmakīrti tradition of logic and
epistemology, left a lasting and definite impact on Indian and Tibetan Buddhist philosophy.33
Notes
1 See Frauwallner 1961, 141–43.
2 See, for example, the Records of the Ba Clan (sba/dba’ bzhed) for one of the earliest sources (Wangdu
and Diemberger 2000). For a much later account from a compilation of sources, see Butön Rinchen
Drup’s (bu ston rin chen grub, 1290–1364) History of Buddhism (chos ’byung gsung rab rin po che’i
475
Allison Aitken
3
4
5
6
7
8
9
10
11
12
13
14
gter mdzod), translated in Obermiller (1932, 187–92), and Gö Lotsawa Zhönnu Pel’s (’gos lo tsā ba
gzhon nu dpal, 1392–1481) Blue Annals (deb gter sngon po), translated in Roerich (1949, 41–44). For
a recent compilation of sources on the life and work of Śāntarakṣita, see Eltschinger 2019.
Portraits of the members of this lineage were painted on the walls of Samyé; see, for instance, Obermiller 1932, 190; Roerich 1949, 34.
Tibetan doxographies commonly present Śāntarakṣita as an exemplar of the so-called *YogācāraSvātantrika-Madhyamaka school of thought and Bhāviveka as the representative of *SautrāntikaSvātantrika-Madhyamaka, while Buddhapālita and Candrakīrti are standardly cited as paradigmatic
proponents of *Prāsaṅgika-Madhyamaka. “Yogācāra” in Śāntarakṣita’s doxographical designation indicates his synthesis of Yogācāra ontology into his account of conventional truth, while the
*Svātantrika label signifies (in part) his style of argumentation which, following Bhāviveka, deploys
independent inferential arguments (svatantrānumāna) to establish his theses rather than exclusively
utilizing reductio ad absurdum arguments ( prasaṅga) to undermine the theses of his opponents. On
the Tibetan doxographical assignments of Śāntarakṣita, see Seyfort Ruegg 1981, 87–100; Blumenthal
2004, 41–470, and for a critical analysis of this doxographical assignment, see McClintock 2003. For
discussions on the historical development of the *Svātantrika and *Prāsaṅgika categories, see Dreyfus and McClintock 2003 and Seyfort Ruegg 2006.
Śāntarakṣita’s authorship of this work has been called into question by Tsongkhapa Lobsang Drakpa
(Tsong kha pa Blo bzang grags pa, 1357–1419) and Tāranātha (1575–1634) owing to the fact that
Kamalaśīla appears to argue against the author of the SDVP in the TSP; see Eckel 1987, 27–31; Blumenthal 2004, 29; Seyfort Ruegg 1981, 68 n. 224.
The attribution of Tattvasiddhi to Śāntarakṣita has been called into question by Steinkellner (1999, 356–57).
In addition to the texts already mentioned, the other texts attributed to Śāntarakṣita in the Tengyur
are: two short praises, the Praise of the Eight Tathāgatas (*Aṣṭatathāgatastotra) and Praise of the
Bhagavan: Song of Śrī Vajradhara (*Śrīvajradharasaṃgītibhagavatstotra) together with autocommentary (ṭīkā); two tantric works, Five Great Instructions on Kurukulla Arisen from Hevajra
(Hevajrodbhavakurukullāyāḥ Pañcamahopadeśa) and Ritual for the Extensive Recitation of the Previous Aspirations of the Seven Tathāgatas Collected from Sūtras (Saptatathāgatapūrvapraṇidhānaviśeṣavistārakalpavacanavidhi-sūtrāntasaṃkṣepa); and a commentary (vṛtti) on Candragomin’s
(seventh century) Twenty Verses on the Bodhisattva Vows (Bodhisattvasaṃvaravṃśaka). In his Commentary on the Ornament, Śāntarakṣita appears to allude to another text he authored titled Ascertainment of the Ultimate (*Paramārthaviniścaya) (Ichigō 1985, 330). As Eltschinger (2019, 384–85)
points out, Śāntarakṣita also seems to allude to this work in TS 2083.
See McClintock 2010, 58–62, 2013; Tillemans 2011, 153–54, 2016, 143–44 on Kamalaśīla’s elaboration on this concept, particularly in the context of discerning the correct understanding of conventional truth; on the term, prekṣāvat, see also Eltschinger 2007, 137–50, 2014, 195 n. 17, 219–34.
On Śrīgupta’s Introduction to Reality (Tattvāvatāra = TA) and accompanying autocommentary (-vṛtti
= TAV), see Aitken (Forthcoming).
For a short synopsis of the Compendium, see Seyfort Ruegg 1981, 89–90, and for a compilation of secondary scholarship on this work, see Steinkellner and Much 1995, 56–63; Eltschinger 2019, 385–86.
Śāntarakṣita presents a summary of the topics of the text in the first six stanzas of the Compendium,
with the main points of the first twenty-three chapters glossed as modifying dependent origination
and those of the final three chapters modifying the Buddha who taught dependent origination. On the
two-part structure of the Compendium, see McClintock 2010, 97–98.
On the application of this framework for describing Śāntarakṣita’s method, see McClintock 2003,
2010, 85–91; on this same concept applied to the work of Dharmakīrti, see Dreyfus 1997 and Dunne
2004, 53–79. Blumenthal (2004, 43, 44, 46) articulates this same approach of Śāntarakṣita’s in terms
of a “graded ascent of philosophical views,” “multiple levels or stages of provisionality,” and an
“ascent through provisional views.”
McClintock (2003, 68–76) points to TS 1916–17 as instances where Śāntarakṣita acknowledges the superiority of Madhyamaka. Śāntarakṣita additionally nods to the Madhyamaka tradition in the framing of the text,
mirroring Nāgārjuna’s opening to his Fundamental Verses on the Middle Way (Mūlamadhyamakakārikā)
in laying out the purpose and structure of the Compendium in its opening stanzas.
(i) Vaiśeṣikas defend the conjoined view, (ii) Vaibhāṣikas such as Saṅghabhadra (c. fourth–fifth
century) defend the interceding space view, and (iii) Vasubandhu defends the spatially continuous
view in the Commentary on the Treasury of Abhidharma (Abhidharmakośabhāṣya) ad 1.43d2, where
476
Śāntaraks∙ita
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
he attributes the position to the Kaśmīri Vaibhāṣikas such as Vasumitra (c. second century CE);
Śubhagupta (c. 720–780) also defends this view in his Proof of External Objects (Bāhyārthasiddhi)
50, 52, 53, 56, although Śāntarakṣita cites Śubhagupta in connection to the second view.
In his autocommentary, Śāntarakṣita goes on to praise Yogācāra for its theory being confirmed and known
by a source of knowledge ( pramāṇa), for serving as a corrective for mistaken views, for its rejection of
material simples, and for its agreement with scriptural sources (MAV ad MA 45; Ichigō 1985, 124).
MA 46–51; see Dharmakīrti’s Explanation of the Sources of Knowledge (Pramāṇavārttika) 3.209–22
and Śrīgupta’s TA 4.
MA 52–60; see TA 5–6, TAV transitional stanzas (antaraśloka) 1–4.
In Cartesian terms, the question here concerns the formal reality of thoughts themselves rather than
the objective reality of whatever might be represented in thought.
*Satyākāravāda (rnam pa bden par smra ba) and *Alīkākāravāda (rnam pa brdzun par smra ba) (literally “theory/proponent of real representations” and “theory/proponent of unreal representations”)
are not attested in extant Indic doxographies, where we instead find the Sākāravāda-Nirākāravāda
distinction. While these labels were imposed onto diverse sets of thinkers in contriving subschools
of Yogācāra, they are nevertheless useful for clarifying the structure of Śāntarakṣita’s argument
and the dialectical lay of the land as he understood it. One should be careful to distinguish the use
of the terms sākāra vs. nirākāra in the Yogācāra context from the use of this same pair of terms to
designate, respectively, representationalist vs. direct realist accounts of ordinary perception among
realists about external objects.
To the contrary, Berkeley (Principles in Works vol. 2, 98) and Hume (Treatise 1.2.4), for instance,
both argue for theories of a minima sensibilia, a kind of phenomenal atomism on which a perception
is reducible to indivisible, unextended simples.
This account can be traced back to Nāgārjuna, who argues that nothing lays claim to the Abhidharma
ontological category of ultimate reality ( paramārthasat), or substantially reality (dravyasat).
TA 11; see Aitken (Forthcoming) and (2021a) for an interpretation of Śrīgupta’s version of this threefold criterion. For a comparison of Śrīgupta and Śāntarakṣita on the two truths, see Aitken (2021b).
See, for example, Kamalaśīla’s MAP ad MA 64, Haribhadra’s Illuminating the Ornament of Realization (Abhisamayālaṃkārālokā) (Wogihara 1932–35, 594.18–25), Bhāviveka’s Compendium on the
Meaning of the Middle Way (Madhyamakārthasaṃgraha) 9–11 and Jeweled Lamp for the Middle
Way (Madhyamakaratnapradīpa) 1.4, and Atiśa’s Introduction to the Two Truths (Satyadvayāvatāra)
4. Jñānagarbha also sets out versions of these three criteria (SDV 8, 12, and 21).
Explanation of the Sources of Knowledge 3.3ab and Essence of Reasoning (Nyāyabindu) 1.15.
Śāntarakṣita cites Nāgārjuna’s Fundamental Verses on the Middle Way 24.18 in support of this
criterion.
Importantly, Śāntarakṣita does not regard this as an innovation, citing Nāgārjuna’s Sixty Stanzas on
Reasoning (Yuktiṣaṣṭika) 21 and 34 (among other sources) in support of this presentation.
For example, Garfield (2016) argues that Jamgön Ju Mipham Gyatso (’Jam mgon ’Ju Mi pham rgya
mtsho, 1846–1912) reads Śāntarakṣita’s Yogācāra phenomenologically.
See Saccone’s 2018 edition, translation, and analysis of this chapter. It is this second, metaphysical
argument that follows Vasubandhu’s previously discussed anti-atomist argument from his Twenty
Verses 11–15.
It is worth pointing out that Śāntarakṣita’s rejection of atoms does not in principle preclude the possibility of conventionally real external objects that are not founded in atoms, though he does not
explicitly consider such a scenario, presumably owing to the fact that all external world realists in his
intellectual milieu were atomists.
To the contrary, Berkeley famously makes a case for subjective idealism as part of his project to
restore commonsense.
This difference is also reflected in the different examples cited in the two inferences (in which the
entailment relation is instantiated), with the Madhyamaka example being a reflection, which lacks an
intrinsic nature, and the Yogācāra example being a lotus growing in the sky, which is a nonexistent
thing.
Since atoms are stipulated by their proponents as partless, and therefore true unities by definition (TS
1992ab), the rejection of the unity of atoms would necessarily preclude their existence. The problem
here lies in the sweeping nihilistic implications of the generalized Yogācāra entailment relation when
applied to the Madhyamaka all-inclusive subject.
477
Allison Aitken
33 Although Śāntarakṣita’s conciliatory approach to Yogācāra was rejected by many later Tibetan
Mādhyamikas in favor of Candrakīrti’s account of conventional truth, his synthesis of Madhyamaka
with Dignāga and Dharmakīrti’s logic and epistemology remains influential in Tibetan Madhyamaka
cutting across traditions.
Bibliography
Aitken, Allison. Forthcoming. Introduction to Reality: Śrīgupta’s Tattvāvatāravṛtti. Harvard Oriental
Series. Cambridge, MA: Harvard University Press.
———. 2021a. “No Unity, No Problem: Madhyamaka Metaphysical Indefinitism.” Philosophers’
Imprint 21 (31): 1–24.
———. 2021b. “The Truth About Śrīgupta’s Two Truths: Longchen Rabjampa’s ‘Lower Svātantrikas’ and
the Making of a New Philosophical School.” Journal of South Asian Intellectual History 3 (2): 185–225.
Blumenthal, James. 2004. The Ornament of the Middle Way: A Study of the Madhyamaka Thought of
Śāntarakṣita. Ithaca: Snow Lion Publications.
———. 2009a. “Dynamic and Syncretic Dimensions to Śāntarakṣita’s Presentation of the Two Truths.”
Asian Philosophy 19 (1): 51–62.
———. 2009b. “The ‘Neither-One-Nor-Many’ Argument of Śāntarakṣita: A Classical Buddhist Argument on the Ontological Status of Phenomena.” In Buddhist Philosophy: Essential Writings, edited by
William Edelglass and Jay L. Garfield, 46–60. New York: Oxford University Press.
———. 2015. “Two Topics Concerning Consciousness in Śāntarakṣita’s Yogācāra-Madhyamaka Syncretism.” In Madhyamaka and Yogācāra: Allies or Rivals?, edited by Jay L. Garfield and Jan Westerhoff,
242–51. New York: Oxford University Press.
Dreyfus, Georges. 1997. Recognizing Reality: Dharmakīrti’s Philosophy and Its Tibetan Interpretations.
Albany: State University of New York Press.
Dreyfus, Georges, and Sara McClintock. 2003. The Svātantrika-Prāsaṅgika Distinction: What Difference Does a Difference Make? Boston, MA: Wisdom Publications.
Dunne, John. 2004. Foundations of Dharmakīrti’s Philosophy. Boston: Wisdom Publications.
Eckel, Malcolm David. 1987. Jñānagarbha on the Two Truths: An Eighth Century Handbook of Madhyamaka Philosophy. Albany, NY: State University of New York Press.
Eltschinger, Vincent. 2007. “On 7th and 8th Century Buddhist Accounts of Human Action, Practical
Rationality and Soteriology.” In Pramāṇakīrtiḥ: Papers Dedicated to Ernst Steinkellner on the Occasion of His 70th Birthday, Part 1, edited by Birgit Kellner, Horst Lasic, M. T. Much, and Helmut
Tauscher, 135–62. Wien: Arbeitskreis für Tibetische und Buddhistische Studien Universität Wien.
———. 2019. “Śāntarakṣita.” In Brill’s Encyclopedia of Buddhism, Volume II: Lives, edited by Jonathan
A. Silk, Richard Bowring, Vincent Eltschinger, and Michael Radich, 383–90. Leiden and Boston: Brill.
Frauwallner, Erich. 1961. “Landmarks in the History of Indian Logic.” Wiener Zeitschrift für die Kunde
Süd- und Ostasiens 5: 125–48.
Garfield, Jay L. 2016. “Cittamātra as Conventional Truth from Śāntarakṣita to Mipham.” Journal of Buddhist Philosophy 2: 263–80.
Ichigō, Masamichi, ed. 1985. Madhyamakālaṃkāra of Śāntarakṣita with His Own Commentary or Vṛtti
and with the Subcommentary or Pañjikā of Kamalaśīla. Kyoto: Buneido.
———. 1989. “Śāntarakṣita’s Madhyamakālaṃkāra.” In Studies in the Literature of the Great Vehicle: Three Mahāyāna Buddhist Texts, edited by Luis O. Goméz and Jonathan A. Silk, 141–240. Ann
Arbor: Center for South and Southeast Asian Studies, University of Michigan.
———. 2016. “Śāntarakṣita on Satyadvaya.” Journal of Indian and Tibetan Studies 20: 166–94.
Jha, Ganganatha. 1937–1939. The Tattvasaṅgraha of Śāntarakṣita with the Commentary of Kamalaśīla.
2 vols. Baroda: Oriental Institute.
Kapstein, Matthew. 2001. Reason’s Traces: Identity and Interpretation in Indian and Tibetan Buddhist
Thought. Boston: Wisdom Publications.
———. 2009. “Śāntarakṣita’s Tattvasaṃgraha. A Buddhist Critique of the Nyāya View of the Self.” In
Buddhist Philosophy. Essential Readings, edited by William Edelglass and Jay L. Garfield, 320–33.
Oxford: Oxford University Press.
Keira, Ryusei. 2004. Mādhyamika and Epistemology: A Study of Kamalaśīla’s Method for Proving the
Voidness of All Dharmas: Introduction, Annotated Translations and Tibetan Texts of Selected Sections of the Second Chapter of the Madhyamakāloka. Wiener Studien zur Tibetologie und Buddhismuskunde, Heft 59. Wien: Arbeitskreis für Tibetische und Buddhistische Studien Universität Wien.
478
Śāntaraks∙ita
McClintock, Sara L. 2003. “The Role of the ‘Given’ in the Classification of Śāntarakṣita and Kamalaśīla
as Svātantrika-Mādhyamikas.” In The Svātantrika-Prāsaṅgika Distinction: What Difference Does a
Difference Make? edited by Georges B. J. Dreyfus and Sara L. McClintock, 125–71. Boston: Wisdom
Publications.
———. 2010. Omniscience and the Rhetoric of Reason: Śāntarakṣita and Kamalaśīla on Rationality,
Argumentation, and Religious Authority. Boston: Wisdom Publications.
———. 2013. “Kamalaśīla and Śāntarakṣita on Scripture and Reason – The Limits and Extent of ‘Practical
Rationality’ in the Tattvasaṃgraha and Pañjikā.” In Scriptural Authority, Reason, and Action: Proceedings
of a Panel at the 14th World Sanskrit Conference, Kyoto, September 1st-5th 2009, edited by Vincent Eltschinger and Helmut Krasser, 209–38. Wien: Verlag der Österreichischen Akademie der Wissenschaften.
Obermiller, Eugene. 1932. The History of Buddhism in India and Tibet by Bu Ston: Part II. Heidelberg:
Harrassowitz.
Roerich, George. 1949. The Blue Annals, Part I. Calcutta: Royal Asiatic Society of Bengal.
Saccone, Margherita Serena. 2018. On the Nature of Things: A Buddhist Debate on Cognitions and Their
Object. Vienna: Arbeitskreis für tibetische und buddhistische Studien Universität Wien.
Sāṅkṛtyāyana, Rāhula. 1935/1936. “Dharmakīrti’s Vādanyāya. With the Commentary of Śāntarakṣita.”
Journal of the Bihar and Orissa Research Society: 21–22.
Seyfort Ruegg, David. 1981. The Literature of the Madhyamaka School of Philosophy in India. Wiesbaden: Otto Harrassowitz.
———. 2006. “The Svātantrika-Prāsaṅgika Distinction in the History of Madhyamaka Thought.” IndoIranian Journal 49 (3/4): 319–46.
Sferra, Francesco. 2004. “Śāntarakṣita: Vātsīputrīyaparikalpitātmaparīkṣa, Disamina del se immaginato
dai seguaci di Vātsīputra (Tattvasaṃgraha, 336–49 con il commento [Pañjikā] di Kamalaśīla).” In
La Rivelazione del Buddha, vol. II, edited by Raniero Gnoli, 1155–74. Milan: Arnoldo Mondadori.
Shastri, Dwarka Das. 1968. Tattvasaṅgraha of Ācārya Shāntarakṣita with the Commentary Pañjikā of
Shri Kamalashīla. 2 vols. Varanasi: Bauddha Bharati.
Steinkellner, Ernst. 1999. “Yogic Cognition, Tantric Goal, and Other Methodological Applications
of Dharmakīrti’s kāryānumāna Theorem.” In Dharmakīrti’s Thought and Its Impact on Indian and
Tibetan Philosophy, Proceedings of the Third International Dharmakīrti Conference, Hiroshima,
November 4–6, 1997, edited by Shoryu Katsura, 349–62. Vienna: Austrian Academy of Sciences Press.
2001. “Is the Ultimate Cognition of the Yogin Conceptual or Non-Conceptual? Part I: A Critical Edition of
the Tantric Tattvasiddhi, Final Section.” In Le Parole e i marmi: Studi in onore di Raniero Gnoli nel suo
70° compleanno, edited by Raffaele Torella, 835–52. Rome: Istituto Italiano per l’Africa e l’Oriente.
———. 2014. The Edition of Śāntarakṣita’s Vādanyāyaṭīkā Collated with the Kundeling Manuscript.
Vienna: Arbeitskreis für tibetische und buddhistische Studien Universität Wien.
Steinkellner, Ernst, and M. T. Much. 1995. Texte der erkennt-nistheoretischen Schule des Buddhismus: systematische Übersicht über die buddhistische Sanskrit-Literatur II. Göttingen: Vandenhoeck and Ruprecht.
Tillemans, Tom T. F. 1982. “The ‘Neither One Nor Many’ Argument for śūnyatā and Its Tibetan Interpretations: Background Information and Source Materials.” Études de Lettres (University of Lausanne)
3: 103–28.
———. 1983. “The ‘Neither One Nor Many’ Argument for śūnyatā and Its Tibetan Interpretations.” In
Contributions on Tibetan and Buddhist Religion and Philosophy, edited by Ernst Steinkellner and
Helmut Tauscher, 305–20. Vienna: Arbeitskreis für Tibetische and Buddhistische Studien.
———. 1984. “Two Tibetan Texts on the ‘Neither One Nor Many’ Argument for Śūnyatā.” Journal of
Indian Philosophy 12: 357–88.
———. 2011. “How Far Can a Mādhyamika Reform Conventional Truth? Dismal Relativism, Fictionalism, Easy-Easy Truth, and the Alternatives.” In Moonshadows: Conventional Truth in Buddhist
Philosophy, edited by the Cowherds, 151–66. New York: Oxford University Press.
———. 2016. How do Mādhyamikas Think? And Other Essays on the Buddhist Philosophy of the Middle. Somerville, MA: Wisdom Publications.
Wangdu, Pasang, and Hildegard Diemberger. 2000. dBa’ bzhed: The Royal Narrative concerning the
bringing of the Buddha’s Doctrine to Tibet. Vienna: Austrian Academy of Sciences Press.
Westerhoff, Jan. 2015. “Nāgārjuna’s Yogācāra.” In Madhyamaka and Yogācāra: Allies or Rivals? edited
by Jay L. Garfield and Jan Westerhoff, 165–83. New York: Oxford University Press.
Wogihara, Unrai, ed. 1932–1935. Abhisamayālaṃkārāloka Prajñāpāramitāvyākhyā (Commentary on
Aṣṭasāhasrikā-Prajñāpāramitā) by Haribhadra Together with the Text Commented On. Tokyo: The
Toyo Bunko.
479