The Problem of Universals in Yogic Perception
and Tsong kha pa’s Solution
Jed Forman
(University of California, Santa Barbara)
he problem of universals is neither stranger to the West nor
the East. In essence, it concerns the relationship between
discrete objects and their properties. What, for example, is it
that makes all discrete chairs belong to a singular class of things called
“chair”? Do all chairs possess some chair-ness? That is, is there some
essence that lurks in each chair and guarantees its membership in a
more general class? Or is the property of being a chair a post facto
conceptualization? That is, is there nothing in the chair that guarantees
its being a chair other than our collective agreement that this group of
otherwise discrete objects ought to be classified under a singular
universal “chair”?
In Western philosophy, the former position is described as realist
and the latter nominalist. In Indian philosophy, the former was
championed by the Nyāya school, who argued that universals are
robust entities that co-exist with (but independently from) the
particulars that instantiate them. This was highly criticized by
Buddhists, especially Dignāga in the 5th or 6th century CE, who
subscribed to nominalism. According to Dignāga, the conceit of real
universals was just another version of a false conceptual belief in an
enduring Self (Skt. ātman, Tib. bdag) applied both to objects and
persons. He considered reality itself, by contrast, to be populated with
discrete particulars: momentary and infinitesimal particles that are
only conceptually constructed into enduring objects, which are
themselves further constructed into classes. While conceptualization
(Skt. vikalpa, Tib. rnam par rtog pa) thus cognizes these false universals,
real particulars are cognized nonconceptually (Skt. nirvikalpa, Tib. rtog
pa med pa). These in turn are the respective objects of two distinct
epistemic instruments (Skt. pramāṇa, Tib. mtshad ma), or ways of
knowing the world: inference (Skt. anumāna, Tib. rjes su dpag pa) and
perception (Skt. pratyakṣa, Tib. mgnon sum). According to Dignāga,
these are the only two epistemic instruments.
Dignāga’s nominalism creates a problem for the Buddhist
soteriological project, however. Most Buddhist schools agree that
T
Jed Forman, “The Problem of Universals in Yogic Perception and Tsong kha pa’s Solution,” Revue
d’Etudes Tibétaines, no. 55, Juillet 2020, pp. 161–181.
162
Revue d’Etudes Tibétaines
nonconceptual realization of the lack of Self is the sine qua non of
spiritual advancement. This realization came to be understood as
occurring through yogic perception (Skt. yogipratyakṣa, Tib. rnal ’byor
mngon sum). As a type of perception, yogic perception should only
perceive particulars. No-Self, however, is a type of universal, a
property that is abstracted away from discrete particulars. How, then,
can nonconceptual yogic perception perceive No-Self, a seeming
universal dependent on conceptualization?
This paper explores Tsong kha pa’s (1357–1419) unique solution to
this conundrum. Tsong kha pa argues that universals are, in fact, not
completely the product of mere conceptual concoction. Neither,
however, does Tsong kha pa argue a robust realism for universals in a
manner that abandons Buddhism’s larger anti-realist agenda.
Affording universals a quasi-real status, Tsong kha pa is able to
salvage them from the deconstruction of conceptual superimposition,
permitting their status as the object of nonconceptual yogic perception.
This position fits squarely with Tsong kha pa and his Dge lugs school’s
wider realist ontological project.
1. Dharmakīrti and Dignāga: Imagined Concepts
The notion that yogins have special perceptual powers enjoys a long
history among a myriad of Indian philosophical schools. In Buddhism,
one of the earliest explicit terminological mentions of “yogic
perception” may occur in the Nyāyānusāra by Saṅghabhadra (fl. 4th–
5th century CE), Vasubandhu’s main teacher. Therein, Saṅghabhdra
tackles the ontological question of similarity. That is, if the Buddhist
theory of momentariness necessitates that all phenomena are discrete
momentary entities with distinct causes, how can two things ever be
said to be similar in any real sense? Saṅghabhadra argues that
homogeneity (Skt. sabhāgatā) is a causal property that inheres between
any two similar things ensuring their common membership in a single
class. Furthermore, homogeneity can be known both through
inference as well as directly through yogic perception (Skt. *yogipratyakṣa, Chin. guan xing zhe xian zhen 觀行者現證).1
However, yogic perception’s first formalization in Buddhism is
most widely associated with Dignāga, 2 who, we will see, also
employed it to explain the direct perception of a shared property,
specifically, No-Self (Skt. anātman). Dignāga’s interpretation of No-Self
1
2
Saṅghabhadra 1995: 229–232. Saṅghabhadra’s text is unfortunately lost in Sanskrit,
but Collette Cox communicated to the author (April 2, 2019) that *yogi-pratyakṣa or
possibly *yogi-abhisamaya is the most felicitous reconstruction from the Chinese.
White 2012: 70–72. Pradeep Gokhale, personal communication, April 2019.
The Problem of Universals in Yogic Perception
163
extends beyond the rejection of a reified personal Self (Skt. pudgala)
superimposed on the five aggregates (Skt. pañca-skandha). 3 In his
interpretation, No-Self deconstructs the reification of all objects. While
such objects appear as res extensa, they are, in fact, discrete particulars.
Any universal projected onto these particulars—whether of an entity
extended over its supposed parts or of a type extended over its
supposed tokens—is the product of conceptual reification
superimposed on a group of particulars. The superimposition of a
personal Self on the aggregates is thus just one type of a much more
pervasive tendency to reify real discrete particles and moments into
objects with spatial and temporal extension. Belief in the Self—and
based thereon, the existence of suffering in general—is predicated on
this reification, with the realization of No-Self as its antidote, so
Dignāga argues.
Understanding No-Self intellectually, however, is insufficient for
liberation. The Buddhist path involves a process of converting the
conceptual understanding of the Buddha’s teaching that phenomena
are without Self into a direct nonconceptual realization of that same
truth afforded by yogic perception. Thus, Dignāga’s Compendium of
Valid Cognition (Skt. Pramāṇasamuccaya, Tib. Tshad ma kun btus) verse
1.6 states that “yogins see just the object, unmixed with the guru’s
instructions,” denoting a type of nonconceptual understanding of NoSelf.4 Indeed, to see No-Self otherwise—that is conceptually—is selfdefeating. Since all conceptualization involves reification, even the
concept of No-Self involves the superimposition of a type of Self.
Through yogic perception, by contrast, the aspirant directly sees that
all objects are selfless without this superimposition.
Dignāga’s explanation, however, leaves his successors in somewhat
of a thorny philosophical thicket. Saṅghabhadra can account for the
perception of homogeneity, since it is a real property that extends over
multiple objects. But on Dignāga’s metaphysics, by contrast, No-Self
cannot be said to really inhere in objects in the same fashion. Because
No-Self is the absence of a fictional entity—the Self—it is a negative
property and therefore a type of reification. It therefore must be a
conceptual superimposition—a universal. Universals, however, are
vitiated at the level of nonconceptual perception. It is unclear,
therefore, how No-Self could be an object of yogic perception, which
only perceives real positive entities and not abstract properties or
universals like No-Self. Furthermore, a real positive particular would
3
4
Though, for the story of how he came to reject of the pudgala as well, see Eckel,
Garfield, and Powers 2016: 4.
yogināṃ guru-nirdeśa-avyavakīrṇa-artha-mātra-dṛk || (Dignāga 2005: 3). I also
consulted Dignāga 2008: 4: /rnal ’byor rnams kyi bla mas bstan/ /ma ’dres pa yi don tsam
mthong/.
164
Revue d’Etudes Tibétaines
seem mutually exclusive with a negative property, such as the lack of
Self. The two could theoretically co-exist, but it would seem strange to
say they share an identity. How, then, can positive particulars cogently
instantiate a negative property like No-Self—the foundational
Buddhist truth?
In the Commentary on Valid Cognition (Skt. Pramāṇavarttika, Tib.
Tshad ma rnam ’grel), Dharmakīrti (fl. 6th–7th century CE) hedges
Dignāga’s project in order to resolve the epistemological conundrum
he inherited, relying on a psychology of mental images (Skt. ākāra, Tib.
rnam pa) to negotiate between the conceptuality of Buddhism’s
negative truth5 and Dignāga’s insistence that only positive particulars
are real and perceptible. Mental images describe mental pictures or
representations that always attend conceptual thinking and are
positive particulars. Therefore, Dharmkīrti argues that even
conceptualization has a nonconceptual component, since mental
images arise in conjunction with conceptual superimposition.6 Simply
by fixating on a concept in meditation, the conceptual overlay will
eventually fall away, leaving only a nonconceptual perception of the
attendant mental image. Quoting verse 3.284–285:7
Although considered unreal, meditative bases,
Like the ugliness [of the body], the earth, etc.,8
Can arise as a nonconceptual clear appearance
Constructed by the power of meditation.
Whether existing or non-existing,
Whatever one meditates upon intently
Will end up forming a nonconceptual cognition,
Once that meditation is perfected.9
Thus, meditation on any concept will eventually give way to a
nonconceptual cognition, since every concept has a nonconceptual
5
6
7
8
9
All conceptualization is, in fact, negative on Dignāga and Dharmakīrti. This
constitutes their apoha theory.
For a discussion of Dharmakīrti’s combination of mental images and
conceptualization, see Dunne 2004: 116–119.
For an excellent analysis on these verses, see Dunne 2007.
Ugliness is meditated upon in order to cultivate renunciation. The earth meditative
base (pṛthivī kṛtsna) involves meditating on a disk of earth until the meditator can
take control of the earth element. See Bhadantācariya Buddhaghosa 1991: 111 and
118–264 respectively for a discussion of each.
mi gtsang zad par sa la sogs/ /yang dag min pa’ang bsgoms pa yi/ /stobs kyis sprul pa rtog
med dang/ /gsal bar snang ba can du mthong/ /de phyir yang dag yang dag min/ /gang
gang shin tu bsgoms gyur pa/ /bsgom pa yongs su rdzogs pa na/ /de gsal mi rtog blo 'bras
can/ (Dharmakīrti n.d.: 129a). Also consulted in Sanskrit: aśubhā pṛthivī kṛtsna-ādyabhūtam api varṇyate | spaṣṭa-ābhaṃ nirvikalpaṃ ca bhāvanā-bala-nirmitam || tasmād
bhūtam abhūtaṃ vā yad yad eva abhibhāvyate | bhāvanā-pariniṣpattau tat sphuṭa-akalpadhī-phalam || (Dharmakīrti 1972: verses 3.284–285).
The Problem of Universals in Yogic Perception
165
mental image that attends it. If any concept—even those which
represent “non-existing” things—can become vivid in this manner,
then Dharmakīrti seems to suggest that meditation is a process of selfinduced hallucination. Still, he seems to argue that not all
hallucinations are equally fictitious. Specifically, meditation on the
attendant mental image of the concept of No-Self must be a valid
cognition (Skt. pramāṇa, Tib. mtshad ma) because that image is nondeceiving (Skt. samvādin, Tib. mi bslu ba) with respect to the goal of
liberation.10 While that mental image does not represent a particular in
the world—since although particulars do not have a Self, they do not
instantiate No-Self as a property in a realist sense—it is still an
authentic cause for enlightenment. When considering validity, the
salvific capacity of the mental image of No-Self, thus trumps its failure
to represent real particulars. This reading of Dharmakīrti aligns with
his more general doctrine of causal efficacy (Skt. arthakrīyā, Tib. don
byed), wherein valid cognition is marked by its ability to achieve
desired ends more so than its capacity to accurately represent. Thus, if
the mental image of No-Self helps one achieve enlightenment, it is
irrelevant whether it truly corresponds with the world.11
2. Tsong kha pa: Real Concepts
Even when restricting himself to the viewpoint of Dharmakīrti’s
school, Tsong kha pa did not feel comfortable abandoning the
epistemological framework of reference in favor of pure pragmatism.
Tsong kha pa, therefore, is not only charged with explaining how it is
that universals or meditation upon them lead to desired ends, but how
those universals accurately map onto reality. Tsong kha pa’s project
here was carried forward by his successors in the Dge lugs school.
Georges Dreyfus thus calls their position “moderate realism,” given
their insistence that universals are not purely fictional entities. Neither,
however, do they argue—as the Naiyāyikas do—that universals exist
independently from the particulars that instantiate them. Universals
are moderately real, impossible to disentangle from the objects to
which they belong, yet equally impossible to substantiate independent
of those particulars. 12 Because objects can be said to instantiate
10
11
12
Tib. de la sngar bshad dngos po bzhin/ /bslu ba med can gang yin de/ /bsgoms byung mngon
sum tshad mar ’dod/ /lhag ma nye bar bslad pa yin/ Skt. tatra pramāṇaṃ saṃvādi yat prāṅ
nirṇīta-vastu-vat | tad-bhāvanājaṃ pratyakṣam iṣṭaṃ śeṣā upaplavāḥ || (Dharmakīrti
n.d.: 129b and Dharmakīrti 1972: verse 3.286).
See Dunne 2007. However, Dharmakīrti’s later commentators do not similarly
eschew the issue of reference.
Dreyfus 1997: 179–182.
166
Revue d’Etudes Tibétaines
universals in a robust sense, Tsong kha pa does not have to opt for
Dharmakīrti’s pragmatism, which affords universals a teleological
role, but undermines their veridical correspondence with the world.
This reformulation of universals has significant ramifications for
Tsong kha pa’s analysis of yogic perception.
The source for Tsong kha pa’s view on yogic perception comes from
his Notes on the Perception Chapter (Tib. Mngon sum le’u’i brjed byang), a
commentary on the perception chapter of Dharmakīrti’s
Pramāṇavarttika composed from notes that Rgyal tshab rje Dar ma rin
chen (1364–1432) purportedly took from an oral teaching on the subject
given by his teacher. The relevant section concerns the taxonomy of
universals (spyi mtshan) and particulars (rang mtshan). Tsong kha pa
argues that within Dharmakīrti’s system, there is no possible third
ontological category. An interlocutor argues that an illusion—in this
case, a falling hair (skra shad), what today we call “eye floaters”—fits
into neither of these two categories. Why? “When a falling hair
appears clearly, it is not a particular, since it has no causal efficacy, nor
is it a universal, since it appears clearly, and there is not some other
type of clear appearance it could be.” 13 Despite this, Tsong kha pa
insists, “there is no necessity that [the falling hair] is some third type
of object.” Tsong kha pa gives the following rationale:
The falling hair is not an object. If the falling hair were an object, then
we would have to assert that an object exists however it appears (snang)
as per its linguistic sign (brda la byang), and, similarly, there would be
no need to establish its existence with reference to a conceived object
(zhen). Therefore, while we deny that a falling hair appearing to a sense
consciousness plagued with apparitions of falling hairs is an object,
what about the knowledge of the appearance of a falling hair? We do not
deny that the mental image of the appearance of a falling hair is an
object.14
In other words, only an existing thing must be either a universal or a
particular. Falling hairs, by contrast, are not objects, and thus there is
no necessity they fit into either category. Tsong kha pa uses the novel
Tibetan distinction between appearing object (snang yul) and
conceived object (zhen yul) to substantiate why falling hairs are illusory
13
14
skra shad sogs gsal bar snang ba’i tshe skra shad de nyid rang mtshan ma yin te/ don byed
mi nus pa’i phyir ro/ spyi mtshan ma yin te/ gsal bar snang ba’i phyir dang/ gsal ba gzhan
la rjes ’gro mi byed pa’i phyir/ (Tsong kha pa 1999a: 346).
phung gsum du thal mi ’gyur ro/ skra shad de nyid yul ma yin te/ yul yin na brda la byang
bas gang du snang ba de nyid du grub par mos shing zhen dgos pa la de ltar ma grub pa’i
phyir/ ’dir ni skra shad ’dzag snang gi dbang shes la snang ba’i skra shad nyid yul yin pa
’gog pa yin gyi/ skra shad du snang ba shes sam/ skra shad du snang ba’i rnam pa yul yin
pa ’gog pa ma yin no/ (Ibid: 347).
The Problem of Universals in Yogic Perception
167
and unreal. Namely, while their appearance qua a mental image exists,
that appearance qua the conceived falling hairs themselves—that is, in
reference to their actually being falling hairs—is false.15 We also see
Tsong kha pa drawing on Dharmakīrti’s discussion of yogic
perception, where he argued that there can be a clear appearance of an
object, “whether existing or non-existing.” As Tsong kha pa notes,
though a concept as designated by its linguistic sign may appear
(snang), this appearance in and of itself is insufficient to substantiate
the existence of its referent. If it were, any clear appearance of some
linguistic object (brda la byang) would necessitate that object’s
existence. While such appearances qua appearances exist, they fail to
do so qua their conceived objects, which do not exist. It is this latter
criterion of the conceived object that differentiates illusions from
accurate cognitions.
In other words, Tsong kha pa, like Dharmakīrti, argues that clear
appearances are not limited to existing or non-existing objects. The
yogin, for example, can develop both. 16 There is disagreement,
however, about what it means for such an appearance to be accurate.
While Dharmakīrti forgoes the question of whether appearances
accurately refer, arguing that it is pragmatically irrelevant, Tsong kha
pa avails himself of an explanatory framework foreign to Dharmakīrti
in order to incorporate reference: the distinction between an appearing
(snang yul) and conceived object (zhen yul).17 The introduction of these
two concepts radically reformulates Dharmakīrti under a type of
correspondence theory—that the truth of appearances is determined
by the felicity of their representation of some conceived object, not
merely their pragmatic efficacy in achieving desired ends.
Dharmakīrti, in fact, would argue that appearances invariantly fail
with respect to their conceived object, hallucination, or otherwise. On
his theory of momentariness, cognition of an object never occurs
simultaneously with the object itself, which (as momentary) has
disappeared by the time it appears to consciousness. The object as
conceived—that real object that is thought still to be “out there” when
it appears—ceases to exist by the time it appears. This is why reference
is irrelevant in light of causal efficacy: the real object has a causal
relationship with cognition but can never be its object proper, and so
cognition can never represent an existing object. Thus, differentiating
the validity of cognitions in relation to their conceived object is
15
16
17
This statement is not uncontroversial. Sa skya Paṇḍita (1182–1251), by contrast,
denies that invalid cognitions can have real appearances. See Stoltz 2006.
One famous story of a yogin meditating on an unreal object is found in Tā ra nā
tha 1994: 113–114, where he meditated that he had horns on his head to the point
that they were so convincing he could not fit out the entrance of his cave.
See Dreyfus 1997: 384–385 for a discussion.
168
Revue d’Etudes Tibétaines
vacuous, since all cognitions involve appearances of conceived objects
that no longer exist.18 Even more than this, the conceived object is a
conceptual entity and therefore unreal per Dharmakīrti. If we restrict
the analysis to correspondence, falling hairs are not uniquely false
with respect to a conceived object, since all objects so conceived fail to
represent reality by virtue of being conceptual. The question for
Dharmakīrti is whether such false concepts are effective.
According to Tsong kha pa, on the other hand, the conceived object
is not a categorical representational fiction. Even though this
conceived object is conceptual, concepts only misrepresent the manner
in which an object exists—superimposing permanence and a Self. They
can be valid to the degree they accurately correspond with what that
object is—a chair, a table, etc. This accurate correspondence is in turn
predicated on the existence of spatio-temporally extended objects
which the conceived object correctly captures, albeit failing insofar as
it makes such objects appear permanent and endowed with a Self.19
The falling hair, by contrast, is an illusion because its conceived object
fails in a second regard: not only is it a reification, but it fails to
correspond with any actual hairs. In other words, accurate
conceptualization is nondeceptive (mi bslu ba) in a manner that
conceiving of falling hairs is not, even though both involve a reification
error (’khrul pa).
The introduction of appearing objects and conceived objects gives
Tsong kha pa a method for differentiating false conceptual
cognitions—like that of falling hairs—from veridical ones—that of
chairs, etc. Again, this strategy relies on some notion of those
cognitions accurately or failing to correspond with real spatiotemporally extended objects—correctly conceived real universals. But
this strategy now creates a new issue for Tsong kha pa. Since it seems
suspiciously like the falling hair, how can the universal perceived in
yogic perception cogently be part of the world? Like the falling hair, it
cannot be a causally effective particular, since it is a negative property.
Nor, however, can it be a universal, because it too appears clearly in
yogic perception. Unlike the falling hair, however, No-Self cannot be a
non-object, since this would entail that the fundamental Buddhist
truth does not exist. While Dharmakīrti bites the bullet on this, arguing
whether No-Self is “real” is pragmatically irrelevant, Tsong kha pa’s
reliance on correspondence to real universals precludes this dismissal.
Tsong kha pa actually gives two solutions, one from the perspective
of the Cittamātra (Sems tsam pa) school and another from the
Sautrantika (Mdo sde pa) perspective. He first sets up the problem.
18
19
See a convincing argument from Shākya mchog ldan (1428–1507) on this point in
Dreyfus 1997: 384–385.
See an explanation of Dge lugs pa realism in this regard in Ibid: 322–326.
The Problem of Universals in Yogic Perception
169
Concerning whether this presentation of the two epistemic instruments
[as having particulars and universals as their respective objects] is from
the perspective of only the Sautrantika or is compatible with
Cittamātra, it cannot be the first. From the lower [school (Sautrantika)
on up], the reality of the two Emptinesses [of persons and phenomena]
are thoroughly settled. Thus, [both schools] have to explain how the
enumeration of the two epistemic instruments is encompassed by
nonconceptual yogic perception. Furthermore, this must be the case,
since [Cittamātra] does not give any alternate explanation, and the
Unchanging Absolute (’gyur med yongs grub) […] is definitive on its
own.20
Again, Tsong kha pa recapitulates the issue of universals and
particulars concerning yogic perception. The Cittamātra and
Sautrantika schools share a similar hurdle. Both are committed to the
exclusivity of perception’s apprehension of particulars and inference’s
apprehension of universals. Therefore, both must also explain how a
seeming universal is the object of yogic perception, such that the
knowledge produced by these two epistemic instruments is
encompassed (ya gyal du bsdus pa) by yogic perception’s insight.
Tsong kha pa next signals that he is addressing the issue from the
side of Cittamātra by mentioning the “Unchanging Absolute” (’gyur
med yongs grub). This hails from the doctrine of the three natures (Skt.
trisvabhāva, Tib. mtshan nyid gsum). The last of these, the Absolute (Skt.
pariniṣpanna, Tib. yongs grub), describes Emptiness21 in this school. As
the object of yogic perception, the Absolute further exacerbates the
problem, given that it is unchanging (’gyur med), which is antithetical
to being a causally effective particular. How, then, can it be the object
of yogic perception? Tsong kha pa raises this objection and offers a
clarification:
Someone might object that while from the perspective common to both
schools, the Unchanging Absolute has no causal efficacy, the
Cittamātrin does not concur that it is a universal, and this is
unreasonable. However, the Vijñāptimātrin 22 does not agree that if
something is unable to perform a function that it is necessarily a
20
21
22
gal te tshad ma gnyis kyi grangs nges bsgrub pa’i rnam gzhag ’di dag mdo sde pa kho na’i
dbang du byas sam/ sems tsam pa dang thun mong ba’i dbang du byas pa yin/ dang po mi
rigs te/ ’og nas gnyis stong gi de kho na nyid rgyas par gtan la ’bebs pas de mngon sum du
rtogs pa’i rnal ’byor mngon sum dbye ba’i ya gyal du bsdus pa’i tshad ma gnyis kyi grangs
nges ’chad dgos shing/ de’ang skabs ’di ma gtogs gzhan du ma bshad pa’i phyir dang […]
’gyur med yongs grub la’ang nges par tshang ba’i phyir ro/ (Tsong kha pa 1999a: 346).
Emptiness (stong pa nyid) and No-Self (bdag med dpa) are both objects of yogic
perception and as negative properties similarly problematic.
Cittamātrin (sems tsam pa) and Vijñāptimātrin (rnam rig pa) are synonymous here.
170
Revue d’Etudes Tibétaines
universal, nor that if it can, it is necessarily a particular. While [in the
Vijñāptimātrin system] the Unchanging Absolute […] is truly
established, here we are concerned with the perspective of Sautrantika
alone, and not the uncommon perspective of Vijñāptimātra.23
In other words, someone might assume that the Cittamātrin’s position
is untenable, since they deny that an unchanging entity, like
Emptiness, is a universal, and thus argue that it is a proper object of
perception. However, in that school, that the Absolute is unchanging
and unable to perform a function does not necessitate that it is a
universal. In other words, despite being unchanging, the Absolute can
still be a particular and an appropriate object of yogic perception.24
Notice that Dharmakīrti can sidestep this issue entirely. If the mental
image of the Absolute cultivated in yogic perception leads to liberation,
then this is sufficient. There is no need for that image to correspond
with any actual Absolute qua universal or particular, and indeed, due
to momentariness, mental images will invariably fail to correspond.
Other schools similarly eschew the issue of reference per the Absolute,
arguing that a theory of mental images entails that ultimately the mind
cognizes itself and that this self-cognition constitutes yogic perception
of the Absolute.25 But Tsong kha pa forgoes this strategy.26 He wants
to demonstrate that the Absolute itself—not just its mental image—is a
particular, seemingly so that yogic perception—which, as a type of
perception, can only cognize particulars—corresponds with a robust
object. In fact, later Dge lugs pa scholars understood mental images
not as mental pictures that represent objects, but (in the case of
cognizing real objects) as the direct cognition of external objects
23
24
25
26
gnyis ka’i thun mong ba’i dbang du byas na’ang ’gyur med yongs grub don byed mi nus
pa yin yang spyi mtshan du sems tsam pa khas mi len zhing mi rigs pa’i phyir zhe na/
rnam rig pas don byed mi nus pa la spyi mtshan gyis khab pa dang/ rang mtshan la don
byed nus pas khyab pa khas mi len zhing […] ’gyur med yongs grub la’ang grub pa bden
yang skabs ’dir ni mdo sde pa kho na’i dbang du byas pa yin gyi rnam rig pa’i thun mong
ma yin pa’i lugs ston pa’i dbang du ma byas te/ (Tsong kha pa 1999a: 346).
I am grateful to Dge bshes Blo bzang tshul khrims at the Central Institute for
Higher Tibetan Studies who explained to me this unique feature in Tsong kha pa’s
understanding of the Cittamātra school (personal communication, December 12,
2019).
Such a view is found among Sa skya and Bka’ brgyud thinkers. See Dreyfus 1997:
412–415.
The Dge lugs pa rejection of self-cognition (rang rig) in Prasaṅgika seems, thus, to
bleed into the discussion here. But there are parallels to this debate even within
Yogācāra, specifically between those who consider mental images unreal (Skt.
alīkākāravāda, Tib. rnam rdzun pa) and real (Skt. satyākāravāda, Tib. rnam ldan pa)
respectively. Namely, the issue is whether mental images are objects of
consciousness, which is thus not just self-cognizant, or not, meaning all duality
between mind and its object is illusory. See Kajiyama 1965: 31.
The Problem of Universals in Yogic Perception
171
themselves.27 Based on a similar understanding that mental images are
transparent to external objects, Tsong kha pa forecloses an appeal to
mental images as nonreferential entities. Otherwise, if mental images
were not transparent to non-mental objects, their status as mental
entities and their role in perception would be sufficient to establish
yogic perception as a type of self-cognition. Instead, in the same way,
Tsong kha pa differentiates illusions from accurate cognitions by way
of their correspondence with a veritable conceived object, so too does
he feel compelled to demonstrate yogic perception’s accurate
correspondence with an actual Absolute in order to substantiate its
validity.
What, then, about the Sautrāntika perspective? Tsong kha pa
continues:
From that [Sautrāntika] perspective, yogic perception encounters the
Selflessness of persons (gang zag gi bdag med) implicitly and explicitly
apprehends the causally effective particular. Therefore, it says [in
Dignāga’s Pramāṇasamuccaya] that “yogins see just the object, unmixed
with the guru’s instructions.”28 […] The Unchanging Absolute realized
by yogic perception, [by contrast], is an explicit realization and is not
realized implicitly.29
Here, the strategy is decisively different. Sautrāntika preserves the
mutual entailment of causal efficacy and particular. Thus, the object of
yogic perception must be a causally effective particular. Selflessness
(bdag med), by contrast, is a negative entity, and so cannot be an object
of perception. How, then, does yogic perception realize Selflessness?
Tsong kha pa answers, “implicitly (shugs rtogs).” 30 Through
apprehending the particular free of reification, the yogin subsequently
realizes that there can be no Self. But No-Self is not the direct object of
yogic perception. Yogic perception takes No-Self as an object
implicitly, since the recognition of particulars as ultimately real
precludes attributions of Self. This introduction of implicit cognition
into Dharmakīrti’s system is likely a Tibetan invention.31 But in Tsong
27
28
29
30
31
Dreyfus 1997: 408.
See footnote 4.
de’i dbang du byas na gang zag gi bdag med ’jal ba’i rnal ’byor mngon sum gyis gang zag
gi bdag med shugs rtogs dang/ dngos su dngos po rang mtshan gzung yul du byed pa yin
te/ rnal ’byor rnams kyis bla mas bstan/ ma ’dres pa yi don tsam mthong/ zhes bshad pa
dang […] ’gyur med yongs grub rtogs pa’i rnal ’byor mngon sum gyis ni dngos su rtogs
pa yin gyi shugs la rtogs pa ma yin no/ (Tsong kha pa 1999a: 346).
See Dreyfus 1997: 370–373 for a lucid explanation of how Dge lugs uses the notion
of implicit cognition to explain how perception perceives conceptual entities.
Rong ston (1367–1449), following Sa skya Paṇḍita (1182–1251), understands the
notion of implicit cognition as a Tibetan invention. See Rong ston 2011: 202–207.
172
Revue d’Etudes Tibétaines
kha pa’s formulation, it allows for yogic perception’s not perceiving a
universal, only the momentary particular.
However, even if this (somewhat dubiously) explains
epistemologically how yogic perception can realize No-Self, an
ontological question still remains. That is, even if yogic perception can
implicitly realize a universal, in what sense does this universal exist?
How is it that the cognition of universals—implicitly or otherwise—
accurately represents reality? For this question, we turn to another of
Tsong kha pa’s works, his Extensive Notes on Valid Cognition (Tib. Tshad
ma’i brjed byang chen mo). He first articulates the by-now-familiar
problem with universals.
Firstly, though we regard universal and particular […], positive and
negative entities […], etc., as mentally imputed [distinctions], if we thus
similarly concede that if something is a universal it is necessarily
imagined and permanent, then the common instrumental actions we
necessarily engage in in order to achieve our aims and those very goals
themselves on which our hearts are set—be it omniscience or
whatever—would refer to nothing. If that were the case, then a host of
problems occur […]. Specifically, if we argue that the determinate
objects of the conceptualization of a vase—[a negative entity]—and the
determinate realization of all things in omniscience, etc.—[positive
entities]—are both particulars, then, since those conceptualizations are
no longer erring (ma ’khrul pa), all negative and positive entities would
have to occur simultaneously [rather than in succession from
perception to conceptualization]. If we deny that they are particulars,
then it is difficult to explain how we engage with particulars at all.32
Tsong kha pa rearticulates the problem of universals. That is, if they
are actually real particulars, Dharmakīrti’s system dissolves. It is
integral to Dharmakīrti’s epistemology that the inference of concepts
occurs based on and thus after perception. This condition is even
reflected in the etymologies of both the Sanskrit and Tibetan words for
“inference”—anumāna and rjes dpag—where the prefix “anu-“ and
“rjes” both denote “after.” If universals are particulars, they would be
apprehended within and not after perception. On the other hand, if
they are pure fictions, then they have no relationship with reality, e.g.,
32
dang po ni/ spyi dang bye phrag dang […] dgag pa dang/ sgrub pa […] la sogs pa rnams
rtogs pas sgros btag su bshad pa’ang mthong zhing/ spyi yin na sgro btags yin pas khyab
pa dang de gzhin du rtags sogs pa’ang khab na/ ’bras bu don gnyer la nye bar mkho ba’i
don byed nus pa phal pa rnams dang/ mngon por ’dod pa’i don gyi gtso bo kun mkhyen la
sogs pa’i rnam gzhag bya sa med par ’gyur la/ de ltar na mi rung ba chen por ’gyur bas/
[…] khyed par du bum ’dzin rtog pa dang kun mkhyen nges pa’i rtogs pa la sogs pa’i nges
yul rang mtshan du grub na rtog pa de dag ma ’khrul bar ’gyur bas dgag sgrub thams cad
cig car bya dgos pa dang nges yul rang mtshan du grub pa bkag na rang mtshan de dag
nges pa’i yul du ’jog tshul de dag shin tu dka’ […] (Tsong kha pa 1975: 183–184).
The Problem of Universals in Yogic Perception
173
real particulars, and all conceptual reasoning is delusional. The
question is then how to keep universals and particulars sufficiently
distinct per Dharmakīrti’s system yet sufficiently related so that the
former is based on the truth of the latter. Tsong kha pa notes that if we
have no ability to ascertain particulars via universals, then we could
not reason about the ultimate nature of things. The absurdum would
result that “because grasping at a Self would not be in error, we would
have to conclude liberation is not possible.”33
Tsong kha pa offers a somewhat mind-bending solution. In essence,
though universals themselves are fictions, they still cogently refer to
particulars.
Though all of conventional reality is mere conceptual imputation, it is
assuredly epistemically warranted. All functional entities, [on the other
hand], are established as particulars. To deny this formulation of the two
truths is to grasp as contradictory the fact that although the object of
conceptualization is not a particular, the particular is an object of
conceptualization [emphasis added]. […] In this [Pramāṇavāda] system,
although being conceptually imputed necessitates not being a
functional entity, if you believe that it also necessitates that it is not
established by valid cognition, then this would fundamentally reject
any means of ascertainment. Thus, the division of the two truths is to
be explained in this manner […]34
Tom Tillemans identifies that Tsong kha pa’s position here builds off
an older Gsang phu tradition, in which although conceptual universals
themselves are reifications, the things that are those universals are not
necessarily reifications—spyi sgro btags pa yin, spyi yin na sgro btags yin
pas ma khyab.35 By employing this innovation, Tsong kha pa solves the
ontological disconnect between No-Self as a universal and the real
world as only populated by particulars. In his system, it is not
contradictory that particulars robustly—not merely as a mental
fabrication—instantiate universals, despite the fact that conceptual
thinking itself fails to perceive beyond that universal to the particular.
We could think of this almost like a one-way mirror. The particular has
a transparent relationship to the universal. But when we use inference
33
34
35
gang zag gi bdag ’dzin yang blo ma ’khrul bar ’gyur bas thar pa thob pa’ang mi srid pa
babs blang dgos so/ (See Rong ston 2011: 188).
kun rdzob mtha’ dag rtog btags tsam du rang lugs la tshad mas legs par grub pa dang
dngos po thams cad rang gi mtshan nyid kyis grub par ’jog shes pa’i bden gnyis kyi rnam
dbye ’jog shes pa’i gegs ni rtog pa’i yul rang mtshan ma yin pa dang rang mtshan rtog pa’i
yul yin pa gnyis ’gal bar ’dzin pa nyid yin no […] rigs pa ’di’i lugs la rtog btags tsam la
don byed mi nus pas khyab kyang tshad mas ma grub pas khyab par bzung na rigs pa ’di
nges pa’i gegs kyi gtso po yin pas bden gnyis kyi nam dbye la mkhas par bya ste […] (Ibid:
188–189).
Tillemans 1999: 215–216.
174
Revue d’Etudes Tibétaines
to conceptualize the universal, it occludes the particular. Thus, Tsong
kha pa solves both the perceptual and ontological problems of yogic
perception. Namely, although perception is only implicitly related to
universals epistemologically, universals themselves have a robust
ontological relationship to the real particulars of the world.
However, this is not to say that Tsong kha pa relegates his
discussion of yogic perception only to the cultivation of appearances
that appropriately refer. He also recognizes benefit in the vivid
appearance of purely fictional entities, like that of the falling hair. The
next and final section explores how Tsong kha pa argues that even the
cultivation of hallucinations can have soteriological value.
3. Tsong kha pa: Tantric Concepts
Tsong kha pa does not reserve yogic perception purely for those
perceptions that correctly refer, nor does he completely jettison the
soteriological value of the cultivation of false appearances. In
particular, he seems to afford them a role in his formulation of the
tantric Creation Stage (Skt. utpatti-krama, Tib. bskyed rim). In his Great
Treatise on the Tantric Stages (Tib. Sngags rim chen mo), Tsong kha pa
heavily relies on the Pramāṇavarttika in his chapter on that topic and
the utilization of yogic perception therein, specifically citing the same
verses 3.284–285 we explored earlier. Commenting on the first verse,
he writes:
Although the Sugatas say this in scripture, some heretics, who assert
that renunciation and freedom are impossible, disavow any instance
that demonstrates the feasibility of yogic perception. Since this
objection is pervasive, there is also—for the sake of the opponent—an
analysis of the fundamental proposition that people like aryas, etc.,
exist. Thus, [Dharmakīrti] intends the pursuit of mere habituation
[when he says] that meditation on an object, whether veritable or false,
will result in its clear appearance.36
The “heretics” in this passage are most likely a reference to the
Mīmāṃsakas, who (at least in the earliest strata) reject the possibility
of meditative insight or liberation. Because they argue no authentic
36
zhes gsung rab las ’byung bar bshad pa’i bde bar gshegs pa’i gsung yang spangs par ’gyur
zhing/ thar pa mi srid par ’dod pa’i phyi rol pa la rnal ’byor mngon sum ’byung rung du
sgrub pa’i rtags kyi khab pa nges pa’i gzhir gyur pa’i dpe phyi rol pas kyang mi bsnyon pa
la bsnyon bting bas gzhan sde la ’phags pa’i gang zag sogs yod par sgrub pa’i sgrub byed
kyi rtsa ba bcad pa yang yin no/ des na goms yul la gsal snang ’ong ba la ni yang dag pa
dang log pa’i don gang goms kyang ’dra ste goms pa tsam gyi rjes su byed pa la dgongs
nas […] (Tsong kha pa 1999b: 548).
The Problem of Universals in Yogic Perception
175
yogins exist—that only the Vedas and not any human can be an
authoritative source of spiritual knowledge—the proof for yogic
perception is purely hypothetical without an actual example, and
thereby invalid. Tsong kha pa therefore subsequently offers a proof for
yogins and their perceptions by extrapolating from verse 3.285 of the
Pramāṇavarttika, explaining that the vivid appearance of an object can
arise from its repeated conceptual mediation.
While Tsong kha pa’s proof of yogins is fascinating in its own right,
what concerns us here is how Tsong kha pa envisions the role of yogic
perception in Creation Stage practice. Tsong kha pa makes its
employment clear a little further on in the text, where he again quotes
verse 3.285 of the Pramāṇavarttika, further elaborating:
I have already explained […] that the mind will take on the mental
image of whatever object to which it habituates. First, the beginner
withdraws [the senses], and then, having amassed some
familiarization, she grabs hold of [the meditative object: the deity].
While meditating to reinforce [this object], she visualizes each and
every aspect in detail. [Then], outside of solely imagining that mental
image [of the deity], she is to cultivate a powerful mental certainty [of
being that deity], since both the clear mental image and divine pride
are necessary.37
Tsong kha pa therefore understands yogic perception as the means by
which one accomplishes the Creation Stage. It is a meditative practice
that culminates in the deity’s appearing clearly, as if real. This is
associated with the conviction that the meditator herself is also the
deity, which is described as “divine pride” (lha’i nga rgyal).
However, while Tsong kha pa argued that yogic perception of the
Unchanging Absolute constituted perception of a real object—i.e., that
it was not merely a mental image—Tsong kha pa makes no such
concession here. He first explains that the mental appearance of the
deity in meditation is not the same as an actual sensorial one.
When one has steady Deity Yoga thorough intense habituation [and
there is a clear appearance of the deity], there is no other appearance to
visual consciousness, or to the rest [of the sense consciousnesses].
Because the mental consciousness needs to be fully engaged with its
object, the power of the conditions [for meditation] is diminished as
soon as the visual or other consciousnesses arise. Therefore, no other
appearance, such as that of colors, etc., comes to mind [other than that
37
[…] yid dngos po gang dang gang la sbyar ba de dang de’i rnam par ’gyur ba […] gsungs
te sngar drangs zin to/ de la las dang po pas ni sbyor ba tshogs bsag nas bzung ste nye bar
bsdu ba’i bar rnams sgom pa na re re nas zhib tu gsal btab nas sgom pa dang rnam pa shar
ba tsam min par blo’i nges pa’i ’dzin stangs shugs can bskyed nas bya ste/ rnam pa gsal ba
dang nga rgyal ’dzin pa gnyis ka dgos pa’i phyir ro/ (Tsong kha pa 1999b: 582–583).
176
Revue d’Etudes Tibétaines
of the deity], since those [sense consciousnesses] are not operative at
that time. However, this does not mean that the Creation Stage negates
[those sensory] appearances.38
In other words, cognition of sensory appearances hinders meditation,
and so they cannot operate in tandem; mental consciousness alone
apprehends the meditative object. Again, Tsong kha pa makes use of
Dharmakīrti in making this point, citing Pramāṇavarttika verse 2.112cd:
“Attached to another object, the mind has no power, because it can
grasp nothing else.”39 This is not just a minor point over which mental
apparatus grasps a meditative object. Tsong kha pa notes that the fact
that one cannot be aware of sensory appearances during meditation
does not mean the Creation Stage negates the sensory world itself,
which would be solipsistic indeed. As a corollary, he argues that
because this meditative object is purely the domain of mental
consciousness, and not of the sense consciousnesses, its actuality is
precluded. He thus follows with:
Therefore, when one achieves the power to stop ordinary appearances
at the level of mental consciousness through the exceptional
appearances [of the deity], one gains what is necessary. Although one
will not have reached the deity in reality [emphasis added], even when the
uncontrived pride of the deity arises, one gains what is necessary
thereby.40
Tsong kha pa is therefore careful to distinguish the clear appearance
of the deity to mental consciousness from having “reached the deity in
reality.” The appearance of the deity is thus like the falling hair. It is
real as an appearance and mental image to mental consciousness, but
false in terms of its referent: an actual falling hair or actual deity, as
corroborated by the sense consciousnesses. Although the appearance
of the deity to mental consciousness is soteriologically effective toward
eliminating ordinary appearances, it fails to represent reality, a reality
that Tsong kha pa argues remains a fact of the matter despite
38
39
40
goms pa che bas lha’i rnal ’byor la mnyam par bzhag pa na mig gi shes pa la sogs pa’i snang
ba gzhan mi ’char ba ni/ yid kyi shes pa don de la rjes su zhugs dgos pas mig la sogs pa’i
shes pa skye ba’i de ma thag rkyen gyi nus pa nyams pas de dag re zhig ma skyes pas kha
dog la sogs pa’i snang ba gzhan ma shar ba yin gyi snang ba de dag bskyed rim gyis bkag
pa min te/ (Tsong kha pa 1999b: 574–575).
/rnam shes don gzhan chags pa yis/ /nus med don gzhan mi ’dzin phyir/ (Dharmakīrti
n.d.: 111b).
des na khyad par can gyi snang bas yid shes kyi ngor tha mal pa’i snang ba ’gog pa’i nus
pa thob na des dgos pa ’grub la dngos po la lhar ma song yang lha’i bcos min gyi nga rgyal
skyes na’ang des dgos pa ’grub bo/ (Tsong kha pa 1999b: 575).
The Problem of Universals in Yogic Perception
177
appearances.41
4. Back to Dharmakīrti: Dualistic Concepts
Tsong kha pa thus seems to afford two distinct roles for yogic
perception: the first is to generate a clear and vivid appearance of
meditational objects that actually exist—such as the Unchanging
Absolute—and the second to cultivate other appearances that do not—
like of oneself as a deity—both of which are soteriologically
advantageous. Dharmakīrti himself, it seems, would be suspect of this
distinction. The disparity between him and Tsong kha pa here is put
all the more in relief by a glaring difference between the Tibetan and
Sanskrit editions of the Pramāṇavarttika. While Tsong kha pa’s citation
of verse 2.112cd from the Tibetan is correctly attested in the Sde dge
Bstan ’gyur and translated above—“Attached to another object, the
mind has no power, because it can grasp nothing else”—the Sanskrit
edition puts this hemistich closer to: “When consciousness is defiled
with attachment to another object, it is because it grasps no other object
[but itself].”42 Prajñākaragupta’s (750–810) commentary confirms the
reading:
By no means does seeing [the illusion of] a subject influenced by an
object arise through some interceding, sudden [effect] other than
conceptualizations, which are the mental impressions of an obstructed
41
42
Elsewhere, however, Tsong kha pa does argue that these exceptional appearances
are definitively valid epistemic warrants (tshad ma) because they undo Selfgrasping. See Tsong kha pa 1999b: 609–610. On the other hand, because they are
valid with respect to hindering Self-grasping, it is not necessarily the case that they
are also valid with respect to reference. In other words, as an inversion of
conventional appearances, they may be correct per how things exist but not per
what exists. Again, the representational content of appearances and their
representation of that content’s existence are distinct questions for Tsong kha pa.
anya-artha-āsakti-viguṇe jñāne anartha-antara-grahāt || (Dharmakīrti 1972: verse
2.112). Also consulted Prajñākaragupta 1998: verse 2.113. The discrepancy with the
Tibetan (see note 39) is somewhat bewildering. “Āsakti” may have been misrendered “aśakti” in the Tibetan nus med, but does seem correctly translated as chags
pa. If so, then nus med may be viguṇa, which is a slightly strange translation choice,
since yon tan med pa or some variant would be more standard. The Tibetan phrase
don gzhan mi ’dzin phyir would be more appropriately artha-antara-agrahāt in
Sanskrit, as in, “it does not grasp another object,” but anartha-antara-grahāt more
felicitously means “it grasps something which is not another object.” Finally, the
Tibetan trades the locative jñāne, which denotes a conditional, for an instrumental,
also significantly changing the meaning.
178
Revue d’Etudes Tibétaines
consciousness. Thus, [subject and object] come from consciousness
alone.43
In other words, Dharmakīrti does not simply mean that mental and
sensory appearances cannot be simultaneous, but that the very notion
that some appearances represent external sensory objects is false. The
disparity between the Sanskrit and Tibetan here serendipitously traces
Tsong kha pa’s deviation from Dharmakīrti’s thought. Tsong kha pa
understands this verse to mean that sensory and mental appearances
are mutually exclusive only with respect to consciousness; it is not the
case that “that the Creation Stage negates appearances” in the world
writ large. Their mutual exclusivity is epistemological, not ontological.
This is consistent with Tsong kha pa’s larger framework that
differentiates appearances and their referents. The cessation of certain
appearances’ presence in consciousness does not necessitate the
elimination of their referents, no more than the appearance of oneself
as a deity necessitates actually being a deity.
The Sanskrit reflects Dharmakīrti’s rejection of representationalism.
That is, the notion that there is some referent to appearances is a
confusion, since the mind is actually grasping itself when it thinks it
apprehends an external object. On this understanding, the question of
what appearances accurately refer is simply ill-formed. As long as they
are tainted with conceptualization, appearances never accurately
represent the world. Appearances thus categorically fail as an
ontology, and the distinction between valid and invalid appearances
based on reference is vacuous. One wonders what Dharmakīrti would
have to say about Creation Stage practice in general.
The issue of appearances and the degree to which there is a distinct
reality that they represent accords with the larger theme of this paper:
universals and the degree to which they correspond with the world.
Dharmakīrti argues that conceptual universals and even their
attendant mental image particulars are not accurate representations.
Nonetheless, the appearance of certain mental images in yogic
perception can have a powerful soteriological effect. Tsong kha pa, on
the other hand, wants a more robustly true object for yogic perception.
Universals, including Buddhist truths, are thus part and parcel of
reality. There is no need for pragmatist apologetics to justify their
being objects of yogic perception.
43
na khalu vyavahita-vijñāna-vāsanā-vikalpānām anyena avāntara-upanipātinā śaktiviṣaya-viṣayeṇa* udayas dṛśyate | tatas vijñānāt ekakam vā […] (Prajñākaragupta 1998:
verse 2.113). *The edition gives the genitive viṣayiṇaḥ as another reading, which
seems more accurate here.
The Problem of Universals in Yogic Perception
179
Bibliography
Tibetan-Language Sources
Dharmakīrti. n.d. Pramāṇasavarttikakārikā. Tshad ma rnam ’grel gyi tshig
le'ur byas pa [A Commentary on Valid Cognition in Verse]. Sde dge
Bstan ’gyur, Toh. no. 4210. Tshad ma ce, 94a–151a.
Dignāga. 2008. Pramāṇasamuccayavṛtti. Tshad ma kun las btus pa’i ’grel pa
[A Commentary on the Compendium of Valid Cognition]. Dpe
bsdur ma Bstan ’gyur, Toh. no. 4203. Tshad ma ce, 3–30.
Rong ston. 2011. “Rigs gter rnam bshad nyi ma’i snying po” [The Heart
of the Sun: A Commentary on the Treasury of Valid Cognition]. In
The Collection of the Eighteen Renowned Scriptures: Root Texts and
Commentaries, edited by Sachen International. Kathmandu: Sachen
International. Accessed May 13, 2020. http://www.sakyalibrary.
com/Assets/files/RTON001_ཚད་མ་རིགས་པའི་གཏེར་-ི་.མ་བཤད་ཉི་མའི་2ིང་པོ1.pdf.
Tā ra nā tha. 1994. Rgya gar chos 'byung [A History of Indian
Buddhism]. Chengdu: Si khron mi rigs dpe skrun khang.
Tsong kha pa. 1999a. “Mngon sum le’u’i brjed byang” [Notes on the
Perception Chapter]. In Gsung ’bum: Tsong kha pa [The Collected
Works of Tsong kha pa], edited by Dar ma rin chen, vol. ba, 342–
395. Xining: Mtsho sngon mi rigs dpe skrun khang.
———. 1999b. “Sngags rim chen mo” [The Great Treatise on the
Tantric Stages]. In Gsung ’bum: Tsong kha pa [The Collected Works
of Tsong kha pa], edited by Dar ma rin chen, vol. ga. Xining: Mtsho
sngon mi rigs dpe skrun khang.
———. 1975. “Tshad ma’i brjed byang chen mo” [The Extensive Notes
on Valid Cognition]. In Gsung ’bum: Tsong kha pa [The Collected
Works of Tsong kha pa], edited by Dar ma rin chen, vol. 22, 152–
245. New Delhi: Nga dbang dge legs bde mo.
Sanskrit-Language Sources
Bhadantācariya Buddhaghosa. 1991. The Path of Purification
(Visuddhimagga). Translated by Bhikkhu Ñāṇamoli. Kandy:
Buddhist Publication Society.
Dharmakīrti 1972: Miyasaka, Yūsho, ed. 1972. “Pramāṇavārttika-kārikā:
180
Revue d’Etudes Tibétaines
Sanskrit and Tibetan.” Acta Indologica 2, 1–206.
Dignāga 2005: Steinkellner, Ernst, ed. 2005. Dignāga's
Pramāṇasamuccaya, Chapter 1: A Hypothetical Reconstruction of the
Sanskrit Text with the Help of the Two Tibetan Translations on the Basis
of the Hitherto Known Sanskrit Fragments and the Linguistic Materials
Gained from Jinendrabuddhi’s Ṭīkā. Vienna: Österreichische
Akademie der Wissenschaften.
Prajñākaragupta. 1998. “Prajñākaragupta’s Pramāṇvārttikabhāṣyam
Manuscript B. Facsimile Edition.” In Sanskrit Manuscripts of
Prajñākaragupta’s Pramāṇvārttikabhāṣyam: Facsimile Edition, edited by
Shigeaki Watanabe, 1–81. Patna: Bihar Research Society and
Naritasan Institute for Buddhist Studies.
English-Language Sources
Dunne, John D. 2004. Foundations of Dharmakīrti’s Philosophy. Boston:
Wisdom Publications.
———. 2007. “Realizing the Unreal: Dharmakīrti’s Theory of Yogic
Perception.” Journal of Indian Philosophy 34 (6), 497–519.
Dreyfus, Georges B. J. 1997. Recognizing Reality: Dharmakīrti's
Philosophy and its Tibetan Interpretations. Albany: State University of
New York Press.
Eckel, Malcolm David, Jay L. Garfield, and John Powers. 2016. “The
Subject Matter of Investigation of the Percept: A Tale of Five
Commentaries.” In Dignāga's Investigation of the Percept: A
Philosophical Legacy in India and Tibet, edited by Douglas Duckworth,
Malcolm David Eckel, Jay L. Garfield, John Powers, Yeshe
Thabkhas, and Sonam Thakchöe, 3–37. Oxford: Oxford University
Press.
Kajiyama, Yuichi. 1965. “Controversy between the Sākāra- and
Nirākāra-Vādins of the Yogācāra School—Some Materials.” Journal
of Indian and Buddhist Studies (Indogaku Bukkyogaku Kenkyu) 14 (1),
429–518.
Saṅghabhadra. 1995. Disputed Dharmas: Early Buddhist Theories on
Existence: An Annotated Translation of the Section on Factors
Dissociated from Thought from Saṇghabhadra's Nyāyānusāra.
Translated by Collett Cox. Tokyo: International Institute for
The Problem of Universals in Yogic Perception
181
Buddhist Studies.
Stoltz, Jonathan. 2006. “Sakya Pandita and the Status of Concepts.”
Philosophy East and West 56 (4), 567–582.
Tillemans, Tom J. F. 1999. Scripture, Logic, Language: Essays on
Dharmakīrti and His Tibetan Successors. Boston: Wisdom
Publications.
White, David Gordon. 2012. “How Big Can Yogis Get? How Much Can
Yogis See?” In Yoga Powers: Extraordinary Capacities Attained through
Meditation and Concentration, edited by Knut A. Jacobsen, 61–76.
Boston: Brill.
v