Apple, James 2010-Spiritual Exercises of Middle Way PDF

Buddhist Spiritual Practices
Thinking with Pierre Hadot on

Buddhism, Philosophy, and the Path
Edited by David V. Fiordalis
Mangalam Press
Berkeley, CA
Mangalam Press
2018 Allston Way, Berkeley, CA USA
www.mangalampress.org
Copyright © 2018 by Mangalam Press. All rights reserved.
No part of this work may be copied, reproduced, published, distributed, or stored

electronically, photographically, or optically in any form without the prior written
permission of the publisher.
ISBN: 978-0-89800-117-4
Library of Congress Control Number: 2018930282
Mangalam Press is an imprint of Dharma Publishing.
The cover image depicts a contemporary example of Tibetan Buddhist instructional

art: the nine stages on the path of “calming” (śamatha) meditation. Courtesy of Exotic
India, www.exoticindia.com. Used with permission.
♾ Printed on acid-free paper. This paper meets the requirements of ANSI/NISO

Z39.48-1992 (Permanence of Paper).
Printed in the USA by Dharma Press, Cazadero, CA 95421
10 9 8 7 6 5 4 3 2 1
Table of Contents
Acknowledgments
ix
Introduction
David V. Fiordalis
1
Some Remarks on Hadot, Foucault, and

Comparisons with Buddhism
Steven Collins
21
Schools, Schools, Schools—Or,

Must a Philosopher be Like a Fish?
Sara L. McClintock
71
The Spiritual Exercises of the Middle Way:

Reading Atiśa’s Madhyamakopadeśa with Hadot
James B. Apple
105
Spiritual Exercises and the Buddhist Path:

An Exercise in Thinking with and against Hadot
Pierre-Julien Harter
147
The “Fecundity of Dialogue” and
the Philosophy of “Incompletion”
Maria Heim
181
Philosophy as a Way to Die:

Meditation, Memory, and Rebirth in Greece and Tibet
Davey K. Tomlinson
217
Learning, Reasoning, Cultivating:

The Practice of Wisdom and the Treasury of Abhidharma
David V. Fiordalis
245
Bibliography
291
Contributors
327
Selected Titles from Dharma Publishing

331
The Spiritual Exercises of the Middle Way:
Reading Atiśa’s Madhyamakopadeśa with Hadot
James B. Apple
T his chapter provides a close reading of the Special Instructions on

the Middle Way (madhyamakopadeśa) of Atiśa Dīpaṃkaraśrījñāna
(982-1054 CE) using the approach of Pierre Hadot (1922-2010) to
interpreting philosophia as a way of life. The paper first outlines the
attractiveness of Pierre Hadot’s program for reading ancient philoso-
phy and his notions of spiritual exercises and philosophia. The paper
then addresses recent critiques of interpreting Buddhist thought as
philosophia within the socio-historical context of Indian Buddhism. In
response to such criticisms, a reading of Atiśa’s Special Instructions on
the Middle Way (hereafter, Special Instructions) furnishes a case study
for Buddhist spiritual exercises within a way of life that brings about
self-transformation.
The Attraction of Hadot’s Approach

Pierre Hadot’s program outlines ancient philosophy as comprised
of spiritual exercises within a way of life that brings about self-
transformation.1 Part of the attraction to this approach, as emphasized
by Pierre Force, is Hadot’s strongly contextualist and historicist
stance that is also able to support as entirely legitimate a presentist
1
I undertand “spiritual” in the sense Foucault speaks of “spirituality” as “the
subject’s attainment of a certain mode of being and the transformations that
the subject must carry on itself to attain this mode of being.” Thomas Flynn,
“Philosophy as a Way of Life. Foucault and Hadot,” Philosophy & Social Criticism
31.5-6 (2005): 620.
105
106 Buddhist Spiritual Practices
appropriation of ancient texts.2 This approach is influenced by

Hadot’s early understanding and application of Wittgenstein in his
analysis of ancient texts. Examining ancient texts’ literary structure,
interpreting their “pragmatic” modes of communication to oneself
and others, and understanding their “language games” for apparent
inconsistencies are influences from Hadot’s reading of Wittgenstein.
Along with these characteristics, Phillippe Hoffmann outlines three
major qualities in the numerous works of Pierre Hadot: a rigorous
historical approach to the history of philosophy, clarity and simplicity
of expression without unnecessary technical jargon, and a concern
for joining both a “scientific” style and detailed presentation to the
educated public.3 In addition to these qualities of his approach,
I am interested in the work of Pierre Hadot as it provides a model
for thinking about ancient thought that is an alternative to analytical
philosophy. Hadot himself did not favor the methods of analytical
philosophy, which are quite often purely logical and indifferent to
philology, history, and contextualization, as well as to existent-
ial aspects of texts and their proposals. As recently pointed out by
Vincent Eltschinger:
Since the early seventies analytic philosophy has
become the paragon of philosophical reflection.
In spite of its overall aversion for the history of
philosophy, the analytical approach has remained
. . . the dominant paradigm in the historiography
of Buddhist epistemology . . . analytically oriented
scholarship . . . has generally disregarded texts as
organic wholes, socio-historical contexts and the
dogmatic frameworks in favor of an unhistorical,
2
Pierre Force, “The Teeth of Time: Pierre Hadot on Meaning and Mis-
understanding in the History of Ideas,” History and Theory 50 (2011): 20.
3
Phillippe Hoffmann, “In memoriam. « Pierre Hadot (1922-2010) »,” in Pierre
Hadot, Discours et Mode de Vie Philosophique (Paris: Les Belles Lettres, 2014),
298: Cet ouvrage manifeste avec éclat les trois qualités majeures de nombreux
livres de Pierre Hadot, jusqu’à ses derniers écrits: rigueur de la démarche historique
en histoire de la philosophie, clarté et simplicité absolues de l’e expression (sans
technicité inutile ni jargon), souci de conjoindre à la fois le style “scientifique” et le
style d’un exposé ad extra, tourné vers le public cultivé.
The Spiritual Exercises of the Middle Way107
comparative and, at times, formal approach to logical

quantification, linguistics and ontological theory.4
For a number of scholars of Indian and Tibetan Buddhist scholasticism
the emphasis on an approach through the lens of analytical philo-
sophy reflects, as Eltschinger describes, “their promoters’ aspiration
to see Indo-Tibetan scholastic productions finally recognized,
against resistant Anglo-Saxon and continental European prejudices,
as properly and genuinely philosophical.”5 In this regard, Matthew
Kapstein has recently remarked that the emphasis on making Indian
and Tibetan philosophies as critical as analytical philosophy has been
“an effort at overcompensation” where such scholars have “bracketed
the tradition’s own interest in spiritual progress right out of the
equation.”6 The importance of Hadot’s approach, as emphasized by
Kapstein, is that “ . . . Hadot directs us to envision philosophy itself
in this context as a soteriological enterprise.”7 A Buddhist sense of
soteriology is one in which specific practices and prescribed patterns
of behavior have transformative power and will lead, somehow
necessarily, to specific religious goals. As I have previously noted, the
theological term soteriology is derived from the earlier Greek medical
tradition’s understanding of the root sōtēr as meaning “to heal,” a
point that can be compared to Hadot’s understanding of philosophia as
therapy and the Buddha’s traditional role as doctor and his teachings
of the Dharma as medicine.8 Comparison based on a decontextualized
4
Vincent Eltschinger, Buddhist Epistemology as Apologetics: Studies on the
History, Self-understanding and Dogmatic Foundations of Late Indian Buddhist
Philosophy (Vienna: Verlag der Österreichischen Akademie der Wissenschaften,
2014), 2.
5
Eltschinger, Buddhist Epistemology as Apologetics, 2-3.
6
Matthew Kapstein, “Stoics and Bodhisattvas: Spiritual Exercise and Faith in
Two Philosophical Traditions,” in Michael Chase, Stephen R. L. Clark, and Michael
McGhee, eds., Philosophy as a Way of Life: Ancients and Moderns: Essays in Honor
of Pierre Hadot (Malden, MA: Wiley-Blackwell, 2013), 100-101.
7
Kapstein, “Stoics and Bodhisattvas,” 100.
8
James B. Apple, Stairway to Nirvāṇa (Albany: State University of New York
Press, 2008):191-192, note 4; See also Robert E. Buswell and Robert M. Gimello,
Paths to Liberation: the Mārga and Its Transformations in Buddhist Thought
(Honolulu: University of Hawaii Press, 1992), 2-3 and 310. On the Buddha
approach rooted in analytical philosophy, more often than not, disre-

gards the soteriological context in which Buddhist thought takes place.
Buddhists think with the basic cosmological presumptions that all
conditioned things are dissatisfactory and suffering, including aging,
sickness, and repeated rebirth and redeath. Buddhist authors take
quite seriously the concept of saṃsāra, endless rebirth and redeath,
where all possible worlds have come into existence and dissipated
countless times. For the Buddhist thinker, there are not new things
in the conditioned world created by ignorance. In the case study
presented below on Atiśa’s Special Instructions, all appearances are
false, deceptive, and arise from ignorance, while all worldly activities
are equivalent to a ball of spit.9 Yet, Atiśa’s prescriptions will also advo-
cate for practices that lead, in his understanding, to a transformation
of awareness that is beneficial to oneself and others. The general
contours of Hadot’s approach offer the possibility of providing for the
cultural and historical context of Buddhist thought as opposed to the
decontextualized and ahistorical focus that an analytical philosophical
approach often warrants.
Hadot’s Spiritual Exercises and Philosophia

Hadot is well-known for developing the notions of spiritual exercises
and philosophia in his approach to reading ancient philosophy.
According to Hadot, in order to interpret properly the texts of ancient
philosophy that have come down to us, we must understand their
role in the life of the ancient schools. These texts were mostly written
for students who had already chosen to follow the way of life prac-
ticed in a particular school; their purpose was not to set arguments,
but rather to lead disciples along a path of spiritual progress (and
sometimes to induce members of a larger public to enter onto that
path). Accordingly, the reader must approach ancient philosophical
compared to a doctor, see Phyllis Granoff, “The Buddha as the Greatest Healer:
The Complexities of a Comparison,” Journal Asiatique 299 (2011): 5-22.
9
See also James B. Apple, “Atiśa’s Open Basket of Jewels: A Middle Way
Vision in Late Phase Indian Vajrayāna. An Annotated Translation of the
Ratnakaraṇḍodghaṭamadhyamakopadeśa,” The Indian International Journal of
Buddhist Studies 11 (2010): 117-198.
texts not as insufficiently rigorous precursors to the modern system-

atic treatise, but rather as elaborations of demanding and formative
“spiritual exercises.”10 As explained by Pavie, “spiritual exercises”
are, for Hadot:
an expression that refers to any practice designed to
transform in oneself or in others, the manner of living,
the way of looking at things. This is both a discourse,
whether interior or exterior, and an implementation
. . . . Spiritual exercises are considered as tools, as
means, they do not constitute in themselves, a finality
. . . . These are the methods taught by the masters that
allow disciples to improve, to change. More than a
commitment allowing one to access well-being, it is
a true conversion, passing from one state to another,
a state from which moreover there is no return, since
one is converted.11
Such exercises include techniques of reading texts, meditation, and
physical exercises. Hadot also draws a distinction between “philo-
sophical discourse” and a “philosophical way of life,” and he argues
that the former should properly be regarded as one of several means
for actualizing the latter: “Philosophical discourse, then, originates
in a choice of life and an existential option—not vice versa.”12 As
10
James B. Apple, “Can Buddhist Thought be Construed as a Philosophia,
or a Way of Life? Relating Pierre Hadot to Buddhist Discourses on
Self-cultivation,” Bulletin of the Institute of Oriental Philosophy 26 (2010):
192-193.
11
Xavier Pavie, “Préface. Exercices spirituels, discours et mode de vie,” in Hadot,
Discours et Mode de Vie, 8-10: . . . “exercices spirituels”, expression qui désigne toute
pratique destinée [9] à transformer, en soi-même ou chez les autres, la manière de
vivre, de voir les choses. C’est à la fois un discours, qu’il soit intérieur ou extérieur,
et une mise en œuvre. . . . Les exercices spirituels sont considérés comme de outils,
des moyens, ils ne constituent pas, en eux-mêmes, une finalité. . . . Ce sont ces
méthodes enseignées par les maîtres qui permettent aux disciples de s’améliorer, de
se transformer. Plus qu’un engagement permettant d’accéder à un mieux-être, c’est
une véritable conversion, un passage d’un état à un autre. Un état dont d’ailleurs on
ne revient pas, puisque l’on est converti.
12
Pierre Hadot, What is Ancient Philosophy? Michael Chase, trans. (Cambridge,
MA: Harvard Belknap, 2002), 3.
Mark Shiffman puts it in his review of Hadot’s book, What is Ancient

Philosophy?, “The student chooses the school and the way of life it
embodies, and the school helps him to uncover the implications and
rational foundations of his choice and to conform himself to it more
thoroughly.”13 Hadot states further, “ . . . philosophical discourse
must be understood from the perspective of the way of life of which
it is both the expression and the means. Consequently, philosophy is
above all a way of life, but one which is intimately linked to philo-
sophical discourse.”14 Hadot emphasizes that one of the fundamental
themes of his works is the distance which separates philosophy
from wisdom. Philosophy, he notes, is merely a preparatory exer-
cise for wisdom. As we will see with Atiśa, rational inquiry and
reasoning procedures are preparatory exercises for non-
discursive wisdom.
For Hadot, the ancient Greek term philosophia, or the practice
of philosophy, embodies an “art of living” or way of life in ancient
Hellenistic culture. In Hadot’s view, philosophia is “a lived, experienced
wisdom, and a way of living according to reason,” where wisdom is “a
state of complete liberation from the passions, utter lucidity, knowl-
edge of ourselves and of the world” that provides one with a “vision of
things as they are.” Wisdom is the aim of the ancient philosopher, and
achieving wisdom is a therapy that heals the philosopher, bringing
about a transformation of the person involving liberation from things
such as “worries, passions, and desires.” The ancient philosopher
engages in spiritual exercises that employ reason “designed to ensure
spiritual progress toward the ideal state of wisdom,” analagous to
“the athlete’s training or to the application of a medical cure.”15
Indeed, although Hadot’s use of wisdom may not exactly correspond
to Buddhist notions of wisdom, exercises of reason conducive to the
actualization of transformational insight, or prajñā, is how I have
characterized Buddhist spiritual exercises in a Hadotian sense, and
what we will see in the work of Atiśa.16
13
Mark Shiffman, “Interpreting Ancient Texts.” Modern Age 45 (2003): 370.
14
Hadot, Ancient Philosophy, 3-4.
15
See Hadot, Way of Life, 103 and 58-59.
16
Apple, “Can Buddhist Thought be Conceived as a Philosophia?,” 196.
Problems in Interpreting Buddhist Thought as Philosophia within

the Socio-Historical Context of Indian Buddhism
For Hadot, “ . . . one must not only analyze the structure of . . . ancient
philosophical texts;” one must also situate them in the “living praxis
from which they emanated.” Hadot then advocates:
. . . in order to understand the works of the philo-
sophers of antiquity we must take account of
all the concrete conditions in which they wrote,
all the constraints that weighed upon them: the
framework of the school, the very nature of philo-
sophia, literary genres, rhetorical rules, dogmatic
imperatives, and traditional modes of reasoning.17
Hadot’s approach is based on his meticulous study of ancient Greek
and Latin texts. However, with respect to Indian Buddhist formations,
identifying the concrete conditions in which Indian authors wrote
their texts and the living practices from which they emanated is
exceedingly difficult, if not outright impossible. Vincent Eltschinger
summarizes this issue as follows:
Recourse to the works of P. Hadot requires that one
take them seriously by trying to restore Buddhist
literature to the “educational,” “psychological,” and
“spiritual” praxis to which they could be registered,
but also, and perhaps above all, to the institutional
life that gives them their legitimacy and their signif-
icance. [539] But it is fair to note that ancient India
poses seemingly insurmountable problems for the
historian whom a “problems and arguments” type
approach would not satisfy. First, because our texts
almost never let themselves show their historical con-
ditions of production, they are never released from
the grounds of argumentation and pure reasoning;
secondly, because archaeologists and historians on
one side, philologists and historians of ideas on the
17
Hadot, Way of Life, 19 and 61.
other, ignore one another more profoundly than ever

before in this area of study . . . 18
One problem with applying Hadot’s approach is that it is quite often
the case that we do not know the historicity of the Indian Buddhist
texts we study, why they were composed, how they were used, and
whether the ideals stated in these texts were actually followed.
Another problem is Hadot’s identification of spiritual exercises and
philosophia with so-called schools, such as the Stoics and Epicureans,
and whether such classifications of schools are applicable for exam-
ining practices in the socio-historical context of Indian Buddhism.
I will return to this issue below.
With regard to Indian Buddhism, we have almost no biograph-
ical data about Buddhist philosophers. As Eli Franco points out, we
have no idea if such so-called philosophers Dignāga, Dharmakīrti,
Dharmottara, or Prajñākaragupta even practiced meditation.19 If we
look at socio-historical studies of Indian Buddhist monastic practices
in the context of their institutional structures, although recitation and
meditation are the ideal duties of a monk (dve bhikṣukarmāṇi dhyānam
adhyayanam), many other duties and obligations were expected of a
resident monk.20 Daniel Boucher provides a nice summary of Gregory
18
Vincent Eltschinger, “Pierre Hadot et les ‘Exercises Spirituels’: Quel Modèle
pour la Philosophie Bouddhique Tardive?,” Asiatische Studien/Études Asiatiques
LXII.2 (2008): 538-539: Le recours aux travaux de P. Hadot exige qu’on les prenne
au sérieux en s’efforçant de restituer les littératures bouddhiques à la praxis
pédagogique, “psychagogique” et “spirituelle” dans laquelle elles purent s’inscrire,
mais aussi, et peut-être surtout, à la vie institutionnelle qui leur confère leur
légitimité et leur signification. [539] Or il est juste de relever que l’Inde ancienne
pose des problèms apparemment insurmontables à l’historien que ne satisferait
pas une approche de type “problems and arguments”. D’abord, parce que nos textes
ne se laissent pratiquement jamais reconduire à leurs conditions historique de
production, ne quittant jamais le terrain de l’argumentation et du raisonnement
purs; ensuite, parce qu’archéologues et historiens d’un côté, philologues et
historiens des idées de l’autre, s’ignorent plus superbement que jamais dans ce
champ d’études; . . .
19
Eli Franco, “Meditation and Metaphysics: On their Mutual Relationship
in South Asian Buddhism,” in Eli Franco and Dagmar Eigner, eds., Yogic
Perception, Meditation and Altered States of Consciousness (Vienna: Verlag der
Österreichischen Akadémie der Wissenschaften, 2009): 119-120.
20
Gregory Schopen, Figments and Fragments of Mahāyāna Buddhism in India.
More Collected Papers (Honolulu: University of Hawaiʻi Press, 2005), 68ff; and
Schopen’s understanding of the socio-historical situation of middle

period Indian Buddhism, roughly the first half of the first millennium,
when he states Indian Buddhist institutions were:
. . . characterized by a highly organized, sedentary
monasticism with a complex administration governed
by an equally complex legal system. Monks living in
these monasteries were bound in a tangled web of
relationships to lay donors and their fellow monks,
relationships that required the constant negotiation
of property rights and ritual obligations.21
Regarding meditation or spiritual practices in such a monastic
institution, a recent study of the Mūlasarvāstivādavinaya by Jeffrey
Bass,22 a student of Schopen’s, indicates that this monastic code never
provides a clear picture of where and when meditation would have
been practiced within a monastery. When the practice of medita-
tion is mentioned, the discussion is directed toward the forest as a
place of practice. The Mūlasarvāstivādavinaya narratives are silent
about any meditation sessions taking place in the daily schedule
of Mūlasarvāstivādin monks. Rather, the practice of recitation is dis-
cussed in great detail and the monastic code outlines why, where, how,
and when recitation was performed in a monastery. Bass describes
four important features of recitation in Mūlasarvāstivādavinaya
narratives: (1) recitation is part of the daily monastic schedule;
(2) recitation is practiced within the monastery; (3) this vinaya
provides detailed instructions for the practice of recitation; and
(4) recitation is an obligatory practice for a monk residing in a
monastery.23 In this way, the practice of recitation represented
“On Monks and Menial Laborers, Some Monastic Accounts of Building Buddhist
Monasteries,” in Buddhist Nuns, Monks, and Other Worldly Matters: Recent Papers
on Monastic Buddhism in India (Honolulu: University of Hawaiʻi Press, 2014),
265-266.
21
Daniel Boucher, Bodhisattvas of the Forest and the Formation of the Mahāyāna:
A Study and Translation of the Rāṣṭrapālaparipṛcchā-sūtra (Honolulu: University
of Hawai’i Press, 2008), 67.
22
Jeffrey W. Bass, “Meditation in an Indian Buddhist Monastic Code.” Ph.D.
Dissertation, University of California at Los Angeles, 2013.
23
Bass, “Meditation,” 158-179.
in Mūlasarvāstivādavinaya narratives has four features that the

practice of meditation does not have. In sum, the evidence for the
practice of meditation in north Indian Mūlasarvāstivādin mon-
asteries based on its monastic code shows that it is not outlined,
emphasized, or required. Rather, meditation is to be practiced away
from the hustle and bustle of the monastery and in the forest in
middle period Indian Buddhism.
Vincent Eltschinger has discussed the problems of utilizing
Hadot’s work as applied to the socio-historical context of Buddhist
thought and practice after the sixth century.24 He has specifically
criticized Matthew Kapstein and Sara McClintock in their attempt
to relate Hadot’s philosophia to late Indian Tibetan Buddhist
works, specifically the Tattvasaṅgraha, an eighth century work
of Śantarakṣita, and its commentary (pañjikā) by his student,
Kamalaśīla. He argues that these works were primarily apologetic
and polemical in their content and context of composition as they
are works of pramāṇa, texts comprised of logical argument and
epistemology for defending the Buddhist dharma and refuting
opponents. In a related and more recent work, Eltschinger explains,
. . . there is also very little doubt that these intel-
lectuals were Buddhist monks active in Buddhist
educational and ritual centers as specialists of the
“science of [justificative] reasons(/evidences)”
(hetuvidyā); that the early descriptions of the
hetuvidyā reflect its essential connection to, and
function as, positive and negative apologetics on
behalf of Buddhism, a connection that could only gain
in strength in a contextof exacerbated religious rival-
ry; that at least since the sixth century CE Buddhism
was the object of orthodox Brahmanical hostility and
had to struggle mainly against Śaivism, for economic
patronage and political support; . . .
Eltschinger continues, noting that “the socio-historical matrix
[consisting of] (religious pluralism, Brahmanical hostility, competi-
tion for patronage), the identity of the opponents (rival salvational
24
Eltschinger, “Quel modèle?”
systems with strong apologetical concerns expressing themselves

through linguistic and epistemological theory), the doctrinal foun-
dations, and issues at stake” were the cultural conditions for what
Eltschinger calls “Buddhist epistemology as apologetics.”25
Eltschinger also notes that, as Indian Buddhist logical-
epistemological discourses are often intermixed with soteriological
doctrines and terminology, it is quite problematic to simply trans-
pose models of spiritual exercises to Buddhist literary productions.26
According to Eltschinger, what is identified as “Buddhist philosophy”
in its Indic socio-institutional context is apologetics. Along these
lines, in Tibet from the 13th century onward, Cabezón has classified a
great amount of so-called Tibetan philosophical works as polemical.27
In brief, it seems that in either Indian or Tibetan socio-historical
contexts the majority of works that scholars consider as philosophy
are actually comprised of either apologetics or polemics.
Eltschinger emphasizes that the great monastic complexes
of Nālandā, Valabhī, and Vikramaśīla, although welcoming serious
Buddhist intellectual study, apparently did not promote meditation
or the practices of religious recluses.28 He states that exegetical as-
pects of commentary were marginal among logicians and Buddhist
theorists of knowledge and that this intellectual current focused on
logical-epistemological concerns and did not have any theological-
religious interests other than apologetics.
One of the problems with utilizing Hadot for examining
Śāntakarakṣita’s Tattvasaṅgraha and Kamalaśīla’s Tattvasaṅgraha-
pañjikā is, as Eltschinger points out, that “such a hypothesis fails to
demonstrate how these works reflect an actual situation of teaching,
that is to say: that of a teacher trying to train and transform an audi-
ence by pedagogical and psychological strategies expressed in the
work itself, particularly in terms of literary genres and rhetorical
25
Eltschinger, Buddhist Epistemology as Apologetics, 4.
26
Eltschinger, “Quel modèle?,” 525-526.
27
José Ignacio Cabezón, Freedom from Extremes: Gorampa’s “Distinguishing the
Views” and The Polemics of Emptiness (Boston: Wisdom Publications, 2007).
28
Eltschinger, “Quel modèle?,” 536-537.
techniques.”29 This approach of utilizing Hadot, for Eltschinger, also

fails to demonstrate how these works, and the teachings that they
reflect, are part of the institutional reality of a school, along with
the attitudes and behaviors for practicing non-discursive spiritual
exercises. Eltschinger notes that an interpretation that follows
Hadot must not neglect key institutional and sociological factors of
philosophical activity. Eltschinger then summarizes his critique:
Indian Buddhism therefore confronts us with the
following situation. On the one hand, there are many
sects whose doctrinal and disciplinary specificities we
have more or less well documented; on the other, phil-
osophical discourses are more or less well known as
well, but sectarian institutional anchoring escapes us.
In other words, these two orders of reality, institutional
and philosophical, coincide or overlap only in very rare
cases in [533] the present state of our knowledge. A
historiographical program based on P. Hadot requires
that one or more sects are identified where one could
specify both a lifestyle and a discourse in which we
would be willing to recognize philosophy. There is no
question here of challenging something like a “Hadot
model” but to call, at first, to conduct the research
necessary to identify a credible intersection between a
sectarian institution and philosophical discourse.”30
29
Eltschinger, “Quel modèle?,” 526: cette hypothèse manque encore de montrer
en quoi ces oeuvres reflètent une situation concrète d’enseignement, c’est-à-
dire: celle d’un maître cherchant à former et à transformer un auditoire par des
stratégies pédagogiques et psychagogiques s’exprimant dans l’oeuvre elle-même,
notamment sur le plan des genres littéraires et des techniques rhétoriques.
30
Eltschinger, “Quel modèle?,” 532-533: Le bouddhisme indien nous confronte
donc à la situation suivante. D’un côté, des sectes nombreuses dont les spécificités
disciplinaires et doctrinales nous sont plus ou moins bien documentées; de l’autre,
des discours philosophiques plus ou moins bien connus eux aussi, mais dont l’ancrage
institutionnel sectaire nous échappe. En d’autres termes, ces deux ordres de réalité,
l’institutionnel et le [533] l’état actuel de nos connaissances. Or un programme
historiographique s’inspirant de P. Hadot requiert que soi(en)t identifiée(s) une ou
des secte(s) telle(s) que la/les spécifieraient à la fois un mode de vie et un discours
Eltschinger’s critique thus presents a challenge to which the present

case stdy offers one possible response.
Case Study: Reading Atiśa’s Special Instructions with Hadot

The Special Instructions on the Middle Way by Atiśa Dīpaṃkaraśrījñāna
provides a viable case study of Buddhist spiritual exercises within
Hadot’s project for valuing philosophy as a way of life. Atiśa31 is famous
for his journey to Tibet and his teaching there for thirteen years
where his teachings on Mahāyāna and Vajrayāna Buddhist thought
and practice came to influence all subsequent traditions of Buddhism
in Tibet. This case study addresses a number of concerns found in
Eltschinger’s critique of adopting Hadot in the study of Buddhist texts
and for which my earlier article on Nāgārjuna provided a test case.32
The present study is based on four works: 1) Atiśa’s Special Instructions,
of which a translation is provided in the appendix to this chapter;33
2) Prajñāmukti’s Commentary on [Atiśa’s] Special Instructions of the
en quoi nous serions disposés à reconnaître de la philosophie. Il n’est ici nullement

question de récuser quelque chose comme un “modèle Hadot”, mais d’appeler, dans
un premier temps, à conduire les recherches nécessaires à l’identification d’une
intersection crédible entre institution sectaire et discours philosophique.
31
As noted by Helmut Eimer, Testimonia for the Bstod-pa brgyad-cu-pa, An
Early Hymn Praising Dīpaṃkaraśrījñāna (Atiśa) (Lumbini: Lumbini International
Research Institute, 2003), 47, note 1, and Geshe Lhundup Sopa, et al., Peacock
in the Poison Grove: Two Buddhist Texts on Training the Mind (Boston: Wisdom,
2001), 24, n. 2, the form Atiśa is derived from the Sanskrit atiśaya, “eminent,
superior” (Tib. phul (du) byung (ba)) rather than Sanskrit ati + īśa, “the great
Lord,” which is not permitted by the rules of Sanskrit grammar. Tibetans often
refer to Atiśa as jo bo, “(The) Lord.”
32
Apple, “Can Buddhist Thought be Construed as a Philosophia?”
33
Madhyamakopadeśa (Dbu ma’i man ngag). Atiśa Dīpaṃkaraśrījñāna and
Tshul khrims rgyal ba, trans. Tôh. no. 3929 (also, no. 4468). See Derge Tanjur, vol.
101 (ki), folios 95v.1-96r.7. In Tshul khrims rin chen, Bstan ’gyur (sde dge). Tibetan
Buddhist Resource Center W23703. 213 vols. Delhi: Delhi karmapae choedhey,
gyalwae sungrab partun khang, 1982-1985. For a scan, see the website of the
Tibetan Buddhist Resource Center, < https://www.tbrc.org/#!rid=W23703>.
An English translation is also found in James B. Apple, “A Study and Translation
of Atiśa’s Madhyamakopadésa with Indian and Tibetan Commentaries,” Acta
Tibetica et Buddhica 7 (2014): 1-82.
Middle Way (madhyamakopadeśavṛtti);34 3) Collection on the Two

Realities (bden gnyis kyi ’bum) by an anonymous bKa’-gdams-pa
(hereafter, Kadampa) author;35 and 4) Explanation of [Atiśa’s] Special
Instructions of the Middle Way, The System of Potawa and His Spiritual
Son (dbu ma’i man ngag gi bshad pa / pu to yab sras kyi lugs) (here-
after, Potawa’s Middle Way) by an anonymous Kadampa author.36
The analysis of these works is also supplemented with
anecdotes from the early biographies of Atiśa and his numerous works
and translations that are preserved in Tibetan. This case study on Atiśa
provides a unique example of Buddhist thought as a spiritual exercise
in that biographical and historical information for Atiśa’s instruction
is available that has not been preserved for other Indian Buddhist
scholars and teachers. The accounts of the journeys of Tibetan trans-
lator-monks to bring Atiśa to western Tibet, along with accounts
of Atiśa’s life and his own writings, supplemented with recently
discovered Kadampa manuscripts, provide evidence for the practice
of Buddhist philosophia in eleventh century India and Tibet.
I base my analysis of the works of Atiśa and his early Kadampa
commentaries on recently published manuscripts of the so-called
Collected Works of the Kadampas (bka’-gdams-pa gsung-’bum).37 These
manuscripts, hidden away in the basement storerooms of Drepung
monastery and the Potala for over four hundred years, are disturbing
34
Madhyamakopadeśavṛtti (Dbu ma’i man ngag ces bya ba’i ’grel pa). Tôh. no.
3931. Derge Tanjur, vol. 101 (ki), folios 116v.7-123v.2. Translated by Prajñāmukti
(shes rab thar pa) and Tshul khrims rgyal ba. For an English translation, see
Apple, “Atiśa’s Madhyamakopadésa,” 6-21.
35
Collection on the Two Realities (bden gnyis kyi ’bum), found in Dpal brtsegs
bod yig dpe rnying zhib ’jug khang, Bka’ gdams pa gsung ’bum (Zi-ling: Krung-
go’i-bod-rig-pa-dpe-skrun-khang, 2003), vol. 19, 335-369. See Apple, “Atiśa’s
Madhyamakopadésa,” 21-23 for a description of the manuscript and 25-65 for a
complete English translation.
36
This is a short work explaining the Atiśa’s Special Instructions of the Middle
Way (madhyamakopadeśa) according to the lineage of Po-ta-ba rin-chen gsal
(1027-1105) and his spiritual son Sha-ra-ba yon-tan grags (1070-1141), con-
tained in volume 19, pages 317-334, of the bKa’ gdams pa gsung ’bum
37
Dbyangs-can-lha-mo, et al, Bkaʼ gdams gsung ʼbum phyogs bsgrigs bzhugs
so. 90 volumes (Chengdu: Si-khron Dpe-skrun Tshogs-pa, Si-khron mi-rigs dpe
skrun-khang, 2006-2009).
in that they bring to light exegetical modes of thinking about Indian

Madhyamaka that Hadot might characterize as a “history of creative
mistakes.”38 In this instance, these earlier, and mostly unknown,
Kadampa manuscripts of Atiśa’s teachings and his followers reveal an
understanding of Madhyamaka thought and exegesis which is, in many
respects, totally opposite from the views of post-sixteenth century
Gelukpa (dge-lugs-pa) thinkers. These texts illustrate how unique the
Gelukpa presentation of Madhyamaka is from the standpoint of Atiśa
and the majority of early Kadampa thinkers. These texts implicitly
demonstrate just how reliant a number of modern interpreters of
Madhyamaka are upon Gelukpa understandings of Madhyamaka and
its related practices, and just how different modern interpreters are
in their soteriological understanding of Madhyamaka in the context of
Buddhist philosophia.39
The Special Instructions of Atiśa, along with the aforemen-
tioned commentaries, furnish evidence for Madhyamaka (“Middle
Way Philosophy”) and its practice in India and Tibet during the early
eleventh to twelfth century. Although all four texts were composed
in Tibet (see below), the base text by Atiśa and the brief commen-
tary (vṛtti) by Prajñāmukti were written by Indian authors initially
in Sanskrit, while the Collection on the Two Realities and Potawa’s
Middle Way are by anonymous Tibetan authors. The texts demonstrate
the distinctions between how an Indian Buddhist authored text was
commented upon by an Indian commentator such as Prajñāmukti, who
was a contemporary of Atiśa, and how later Tibetan commentators
who belonged to the early twelfth-century lineage of Kadampa
38
Hadot, Way of Life, 71-77. See also Force, “Teeth of Time,” 30-31.
39
Shirō Matsumoto, “The Mādhyamika Philosophy of Tsong-kha-pa,” Memoirs
of the Research Department of the Toyo Bunko 48 (1990), 17-47, and Chizuko
Yoshimizu, “The Madhyamaka Theories regarded as False by the Dge lugs pas,”
Wiener Zeitschrift für die Kunde Südasiens 37 (1993), 201-227, have demonstrated
the uniqueness of Tsong-kha-pa and his Dge-lugs-pa followers understanding of
Madhyamaka thought. David Seyfort Ruegg, The Buddhist Philosophy of the Middle:
Essays on Indian and Tibetan Madhyamaka (Boston, MA: Wisdom Publications,
2010) has argued against Matsumoto’s analysis. However, the recent Kadampa
manuscript evidence favors Matsumoto’s and Yoshimizu’s conclusions. See James
B. Apple, “An Early Tibetan Commentary on Atiśa’s Satyadvayāvatāra,” Journal of
Indian Philosophy, 41.3 (2013): 263-329, and “Atiśa’s Madhyamakopadésa.”
followers of Atiśa interpreted the text. In brief, Prajñāmukti is concise

and to the point, providing the reader explanatory glosses on most
of the words and phrases found in Atiśa’s basic text. The anonymous
Kadampa commentators, on the other hand, provide extended expla-
nations to unpack the overall doctrinal meaning of Atiśa’s text.
The Special Instructions and the Entry to the Two Realities
(Satyadvayāvatāra)40 are considered by traditional Gelukpa historians
to be the two foremost textual teachings (gzhung) on the view (lta
ba) within Atiśa’s works.41 However, an early Kadampa commentary
on the latter attributed to Rnal ’byor pa shes rab rdo rje (ca. 1125)
understands the former to be a text on meditation (sgom pa).42 The
colophon to the Tibetan translation of the Special Instructions, as
well as a number of traditional historical sources, mention that this
teaching was given by Atiśa in Lhasa at the request of Rngog legs pa’i
shes rab (11th century).43 Tibetan historical sources state that, based
on Rngog legs-pa’i shes-rab’s request for Madhyamaka teachings
(dbu ma’i chos), Atiśa and his translator-disciple Nag-’tsho lo-tsā-ba
tshul-khrims rgyal-ba (1011-1064) first translated Bhāviveka’s Blaze
of Reasoning (rTog-ge ’bar-ba, Skt. Tarkajvālā), the auto-commentary
40
On this text, see Apple, “Atiśa’s Satyadvayāvatāra.”
41
Las chen kun dga’ rgyal mtshan (1432-1506), Bka’ gdams kyi rnam par thar
pa bka’ gdams chos ’byung gsal ba’i sgron me (Lha sa: Bod ljongs mi dmangs dpe
skrun khang, 2003), 10: / lta ba gtso bor ston pa ni jo bo nyid kyis mdzad pa’i bden
pa gnyis la ’jug pa dang / dbu ma’i man ngag la sogs pa yin la. Texts on the “view”
(lta ba) are, along with practice (spyod pa) and integration (zung ’brel), belong to
the textual (zhung) lineage of teachings. The textual lineage belongs to a broader
classification that includes advice (gdams ngag) and special instructions (man
ngag).
42
Rnal ’byor pa shes rab rdo rje, who was a direct disciple of Sha-ra-ba yon-tan
grags (1070-1141), explains the Madhyamakopadésa to be a text on meditation
(sgom pa) in his bden gnyis kyi rnam par bshad pa in Bka’ gdams gsung ’bum
phyogs sgrigs thengs dang po, vol. 21, fol. 2b7: sgom pa rtan la ’bebs pa’i dbang du
byas na / dbu ma’i man ngag . . .
43
See appendix for a translation of the colophon. Rngog legs pa’i shes rab,
also known as “Gsang phu ba,” was a direct disciple of Atiśa and later founded
the early Kadampa monastery of Gsang-phu ne’u-thog around 1073 C.E. See
Leonard van der Kuijp, “The Monastery of Gsang-phu Ne’u-Thog and Its Abbatial
Succession from ca. 1073 to 1250,” Berliner Indologische Studien 3 (1987): 103-
127, particularly 105.
on his Madhyamakahṛdayakārikā, at Lhasa’s main temple, the ’Phrul-

snang gtsug-lag-khang.44 Atiśa then taught Bhāviveka’s text, followed
by lectures on the smaller and greater Special Instructions. The smaller
Special Instructions, the Madhyamakopadésa, was written in Lhasa,
while the greater text, the Ratnakaraṇḍodghaṭamadhyamakopadeśa,
had already been composed in India.45 This account indicates that
Atiśa utilized Bhāviveka’s thought as an introduction to Madhyamaka,
including the Madhyamakaratnapradīpa and the Blaze of Reasoning
(see below), followed by Candrakīrti’s thought embodied in the
Special Instructions on the Middle Way.46
In relation to its genre and content, the Special Instructions is
a brief text on the practice of Madhyamaka in meditation. The term
upadeśa, “special instructions,” in the title of Atiśa’s basic text is an
expression that has a long history in Indian Buddhism with different
connotations over the centuries. As Étienne Lamotte has remarked
in his study of the Mahāprajñāpāramitāśāstra, upadeśa is the name
of the twelfth and last member of the “twelve-membered” word
of the Buddha (dvādaśāṅgabuddhavacana) and generally signifies
44
The historical sources include: ’Brom ston rgyal ba’i ’byung gnas (1004-
1064) (attributed), Jo bo rje dpal ldan a ti sha’i rnam thar bka’ gdams pha chos
zhes bya ba bzhugs so (Zhin ling: mtsho sngon mi rigs dpe skrun khang, 2012),
168-169; Bya ’dul ’dzin btson ’grus ’bar (1091-1166/1100-1174), Jo bo chen
po rje lha cig gi rnam par thar ba bzhugs so (Lha sa: bod ljongs mi dmangs dpe
skrung khang, 2014), 353; Rnal ’byor pa shes rab rdo rje (ca. 1125), bden gnyis
kyi rnam par bshad pa, fol. 2b6-7; mChim nam kha’ grags (1210-1285), Jo bo rin
po che rje dpal ldan a ti sha’i rnam thar rgyas pa yongs grags bzhugs so (Lha sa:
bod ljongs mi dmangs dpe skrung khang, 2014), 166-167. Las chen kun dga’
rgyal mtshan mentions that Rngog legs pa’i shes rab invited Atiśa to Lhasa but
does not mention the teaching of Madhyamaka (Rnam par thar pa, 97). On the
historical and philological issues in the Tibetan translation of the Tarkajvāla and
other works of Bhāviveka, see now: He Huanhuan and Leonard van der Kuijp,
“Further Notes on Bhāviveka’s Principal Oeuvre,” Indo-Iranian Journal 57 (2014):
399-352.
45
See Apple, “Atiśa’s Open Basket of Jewels.”
46
As David Seyfort Ruegg, Three Studies in the History of Indian and Tibetan
Madhyamaka Philosophy. Studies in Indian and Tibetan Madhyamaka Thought,
Part I (Vienna: Arbeitskreis fūr Tibetische und Buddhistische Studien Universität
Wien, 2000) notes: “In Dīpaṅkaraśrījñāna’s time and circle, Bhavya’s and
Candrakīrti’s schools of Madhyamaka were apparently not clearly differentiated
by distinct designations and they were evidently being studied side by side” (17).
“instruction” or “teaching.”47 A number of scholastic Indian Buddhist

texts preserved in the Chinese Tripiṭaka have upadeśa in their title,
such as the upadeśas on the Saddharmapuṇḍarīka attributed to Vasu-
bandhu.48 The Tibetan Tanjur has dozens of texts containing upadeśa
in their titles. Atiśa authored and translated several texts with upadeśa
in the title such as the Sūtrārthasamuccayopadeśa (mdo’i sde’i don kun
las btus pa’i man ngag, Tôh. no. 3957) and the Ekasmṛtyupadeśa (dran
pa gcig pa’i man ngag, Tôh. no. 3928).49 The term upadeśa translated
into Tibetan as either gdams ngag or man ngag generally means,
as Kapstein notes, “the immediate, heartfelt instructions and admo-
nitions of master to disciple concerning directly liberative insight
and practice.”50 Atiśa’s Special Instructions, therefore, can be said to
offer special guidance or instructions concerning the practice of
Middle Way philosophy, a genre generally distinct from apologetics
or polemics. Several different lineages of this type of instruction and
practice on Madhyamaka were brought into Tibet. Atiśa’s own lineage
of Special Instructions was commented upon at least up until the
thirteenth century, as Skyo-ston smon-lam tshul-khrims (1219-1299)
wrote a brief text on this topic entitled Explanation of Lord [Atiśa’s]
Middle Way Special Instructions (jo bo rje’i dbu ma’i man ngag gi bshad
pa). Another lineage of similar instructions, Guidance on the Great
Middle Way (dbu ma chen po’i khrid) was brought into Tibet by Zla-ba
rgyal-mtshan (12th century).51 mChims nam-mkha’ grags (1210-
1285), a Kadampa author, also wrote a commentary on this lineage
of instruction entitled Guidance on the Middle Way (dbu ma’i khrid),
47
Étienne Lamotte, Le Traité de la grande vertu de sagesse de Nāgārjuna. Vol. 3
(Louvain: Institut orientaliste, 1970), vii-viii.
48
See Taishō 1519 and 1520. For an e-version of the Chinese Buddhist canon,
see the Chinese Buddhist Electronic Text Association, <http://www.cbeta.org/>.
See also the SAT Daizōkyō Text Database, <http://21dzk.l.u-tokyo.ac.jp/SAT/
index_en.html>.
49
For the former, see Derge Tanjur, vol. 101 (ki), folios 338v.7-340v.7; for the
latter see, vol. 101 (ki), folios 94v.4-95v.1.
50
Matthew Kapstein, “gDams ngag: Tibetan Technologies of the Self,” in José
Ignacio Cabezón and Roger R. Jackson, eds., Tibetan Literature: Studies in Genre
(Ithaca: Snow Lion, 1996), 275.
51
Kapstein, “Tibetan Technologies of the Self,” 282.
different than the lineage from Atiśa. Other lineages of Middle Way
practice instructions also existed in Tibet and this genre of Middle
Way instructions influenced later Tibetans scholars such as Red-
mda’-ba gzhon nu blo gros (1349-1412) and Tsong-kha-pa blo-gzang
grags-pa (1357-1419) who composed their own Middle Way guidance
instructions (dbu ma’i khrid). The historical relations between these
lineages is a topic for future research.
As mentioned, Atiśa wrote or gave a lecture on the Special In-
structions in Lhasa based on the request from his disciple Rngog legs
pa’i shes rab. Atiśa therefore had been in Tibet at least seven years
at the time of this teaching.52 Based on Atiśa’s biography, and texts
attributed to him found in the Collected Works of the Kadampas, we
know that he gave a number of teachings on Madhyamaka in India
and Tibet before giving this teaching in Lhasa. I highlight a brief
chronology of these teachings and their context before discussing
the Special Instructions itself.
First, the narratives of Atiśa’s life do not depict him learning
about meditation or Madhyamaka in Indian Buddhist monasteries
alone. Rather, as the young prince *Candragarbha (zla ba’i snying
po), Atiśa visits his teachers at forest retreats or while they were
living in caves. The idealized portraits of Atiśa’s life perserved in the
Potala palace visually depict such meetings with teachers in the forest
retreat.53 Among Atiśa’s teachers related to Madhyamaka or meditation,
Bodhibhadra is mainly connected with a monastery, that being Nālandā,
but his primary teacher of Madhyamaka, Avadhūtipa, is a yogi who
resides in the forest. In his Bodhimārgapradīpapañjikā, written later
in western Tibet at the request of a king, Atiśa explicitly states that
52
According to Las chen kun dga’ rgyal mtshan, Atiśa spent thirteen years in
Tibet, including the first three at mNga’ ris, four years in Dbus-gtsang, and six
years in sNye-thang (de ltar na jo bos mnga’ ris su lo gsum / dbus gtsang gzhan
du lo bzhi / snye thang du lo drug ste/ bod du lo bcu gsum ’gro ba’i don mdzad nas
/ . . . (Rnam par thar pa, 97). See also Alaka Chattopadhyaya, Atīśa and Tibet; life
and works of Dīpaṃkara Śrījñāna in relation to the history and religion of Tibet
(Calcutta: Indian Studies Past & Present, 1967), 330-366, and Vetturini, bKa’
gdams pa School, 89.
53
Phun tshogs tshe brtan, Mnyam med jo bo rje dpal ldan a ti sha’i rnam par
thar pa phyogs bsdus dad pa’i ’jug ngogs (Pe-cin: Krung-goʼi Bod rig pa dpe skrun
khang, 2011), 5-8.
he will not give instructions about meditation or śamatha, for as

he explains,
Instruction in meditation is the kind which relies
upon the personal instruction of an experienced holy
teacher. And that is because details of the teaching on
calmness and insight must be explained, and because
of the difficulty of learning meditation just from
reading books.54
Instruction on meditation must come directly from one’s spiritual
teacher or guru. This implies that during Atiśa’s time instruction
on the practice of meditation was given in a guru-paramparā,
directly from teacher to disciple in a lineage of teaching. Atiśa
gives the Madhyamaka instruction in the form of an upadeśa,
“pith instructions,” based on his earlier and much more extensive
Ratnakaraṇḍodghāṭa, written in India.55 The model of providing
an upadeśa may have been influenced by teachings that Atiśa received
in India, particularly the upadeśa on apratiṣṭhita [madhyamaka]
darśana, one of three pith instructions that Atiśa received
from his teacher Avadhūtipa mentioned in the colophon of the
Tibetan translation of Atiśa’s Sūtrasamuccayasañcayārtha.56
The particular upadeśa on Madhyamaka that Atiśa received
from Avadhūtip has not been preserved, and Atiśa’s two
54
Bodhimārgapradīpapañjikā in Derge Tanjur, vol. 111 (khi), folios 241a4-
293a4: ji ltar bsgom pa ni ’dir ma brjod de gzhung mang pa’i ’jigs pa dang / bla
ma dam pas nyams su myong ba’i man ngag la brten pa rigs kyi / sgom pa’i man
ngag ni yi ger gnas pas shes par dka’ ba’i phyir dang/ zhi ba tu ni zhi gnas dang
lhag mthong bstan pa las bzhad pa’i phyir ro / zhes gsungs so / (275b); See also
Richard Sherburne, ed. and trans., The Complete Works of Atīśa Śrī Dīpaṁkara
Jñāna, Jo-bo-rje: The “Lamp for the Path” and “Commentary”, together with the
newly translated “Twenty-five Key Texts” (New Delhi: Aditya Prakashan, 2000),
208-209.
55
Apple, “Atiśa’s Open Basket of Jewels.”
56
Sūtrasamuccayasañcayārtha, extended colophon not in Derge Tanjur but in
Bstan’ gyur gser bris ma, mdo ’grel a, fol. 513r: . . . lha khang ke ru’i khyams smad
kyi ban de bdag gyi zhus te gdams ngag dang bcas te gnang ngo/ jo bo’i bla ma a
wa dhū ti pas rab tu mi gnas pa’i lta ba dang / las mtha’ sems bskyed pa’i cho ga
dang / mdo kun las btus pa’i don man ngag tu byas pa ’di gsum stabs gcig tu gnang
ba lags so. See also Chattopadhyaya, Atīśa and Tibet, 462.
Madhyamaka upadeśas are preserved only in the Tibetan Tanjur.57

The early biographical evidence and colophons preserved in
Atiśa’s works indicate that Atiśa used the Madhyamakaratnapradīpa
and Tarkajvālā as instructional manuals for teaching Madhyamaka
as a guest lecturer at the monastery of Somapuri.58 Atiśa attributes
the authorship of the Madhyamakaratnapradīpa to the sixth-
century Bhāviveka but most likely this work was a tenth-century
handbook for introducing Madhyamaka by an unknown author. This
information is important in that a great amount of Atiśa’s understand-
ing of Madhyamaka thought and practice is found in the Madhyama-
karatnapradīpa, and as mentioned above, Atiśa used the Tarkajvālā to
introduce Madhyamaka while teaching in Lhasa.
Hubert Decleer has provided a translation of the Kadampa
accounts of the Tibetans’ negotiations to bring Atiśa to Tibet from
Vikramaśīla monastery, as well as Atiśa’s journey to Nepal on the
way to western Tibet.59 I briefly add to these accounts by noting that
the Collection on the Two Realities notes in its beginning section that
Atiśa had a dispute with Ratnākaraśānti (ca. 970–1030), traditionally
considered to be one of Atiśa’s teachers. Decleer’s translation fur-
nishes the Tibetan story of King Lha bla-ma Ye-shes ’od offering his
head’s weight in gold as ransom for Atiśa to come to Tibet.60 Yet the
anecdote in the Collection on the Two Realities presents an alternative
view from Atiśa’s side, indicating a disagreement based on the fact
that the Yogācāra Ratnākaraśānti did not approve of Atiśa’s teaching
57
The Tanjur also preserves Ratnākaraśānti’s Madhyamakālaṃkāropadésa,
but this text is different in content from the upadeśas of Atiśa.
58
On the Madhyamakaratnapradīpa, see Krishna Del Toso, “Some Problems
Concerning Textual Reuses in the Madhyamakaratnapradīpa, with a Discussion
of the Quotation from Saraha’s Dohākośagīti,” Journal of Indian Philosophy 43.4
(2015): 511-557. On the Tarkajvālā, see Eimer, Testimonia, 42, lines 177-186.
59
Hubert Decleer, “Atiśa’s Journey to Tibet,” in Donald Lopez, ed., Tibetan
Religions in Practice (Princeton: Princeton University Press, 1997), 157-177;
Decleer, “Master Atiśa in Nepal: The Tham Bahīl and the Five Stūpas’ Foundations
according to the ’Brom ston Itinerary,” Journal of the Nepal Research Centre 10
(1996): 27-54.
60
On this point, see also Kurtis R. Schaeffer, Matthew Kapstein, and Gray Tuttle,
eds., Sources of Tibetan Tradition (New York: Columbia University Press, 2013),
176-181.
of Madhyamaka to his Tibetan students at Vikramaśīla. Here again, we

do not know the context for this dispute, whether it was personal or
institutional, or how insitutional politics may have influenced the cur-
riculum at Indian universities, or even the reasons underlying Atiśa’s
decision to travel to western Tibet at the age of sixty.
Be that as it may, by the time Atiśa arrived in western
Tibet after a three year journey there, the social expectations for
the content of his teaching had already been established through
writings and imperial ordinances. The Ordinance of Lha la ma
ye shes ’od, which Cabezón classifies as the earliest polemical
work in the history of Tibetan literature,61 establishes that certain
tantras are not word of the Buddha, while the vehicle of per-
fections (phar phyin gyi theg pa) consisting of basic Mahāyāna
Buddhist teachings and the already imperially decreed Madhyamaka
teachings of Nāgārjuna are suitable forms of Buddhadharma. The
main concerns among the Tibetan kings and important men (mi
chen) among the ordained Buddhists was with Vinaya and the prac-
tice of Tantra. Ronald Davidson has outlined the social constraints on
Atiśa’s visit in Tibet, noting that Atiśa seldom founded any temples
or institutions and had restrictions placed on his teaching sched-
ule.62 This was primarily due to the fact that Atiśa was a monk of the
Lokottaravāda section of the Mahāsāṃghika-vinaya, and most, if not
all, his disciples were following the Mūlasarvāstivāda-vinaya that
had been imperially established centuries earlier during the time
of King Relpachen. Several times in his Extensive Biography (rnam
thar rgyas pa), Atiśa complains that he is not allowed to teach his
beloved Mahāsāṃghika-vinaya, vows of secret mantra, or the dohā
songs of realization that were popular during this time in Bengal.63
Davidson notes that there is a distinction between the content of
Atiśa’s public teaching schedule and what texts he translated and
taught with his close disciples. We know there were constraints on
61
Cabezón, Freedom from Extremes, 21.
62
Ronald Davidson, Tibetan Renaissance. Tantric Buddhism in the Rebirth of
Tibetan Culture (New York: Columbia University Press, 2005).
63
mChim nam kha’ grags, Rnam thar rgyas pa, 170-173; Davidson, Tibetan
Renaissance, 111.
Atiśa’s teaching Vinaya and Tantra during his stay in Tibet, but what
about Madhyamaka?
A section of Atiśa’s biography that is also preserved in a num-
ber of Tibetan histories—in an episode that I call the “interrogation at
mNga’-ris”—documents Atiśa’s meeting with hierarchs after Atiśa’s
first arrival in Western Tibet. Atiśa is asked, “Do you accept the
Madhyamaka-Yogācāra according to the ancient ways?” He gives the
rather vague and cagey reply: “I posit things according to scripture
and reasoning.” Several other questions are given and one infers that
the questions from the Tibetan interlocutors presume a different
understanding of Madhyamaka from Atiśa’s. One may infer that the
Madhyamaka presumptions of Atiśa’s interlocutors were most likely
based on the texts and lineages of teachings of Śāntarakṣita and Ka-
malaśīla from the earlier imperial period and that Atiśa’s responses
were based on Candrakīrti’s system, with which Tibetans at this time
may not have been familiar.64
64
See Apple, “Atiśa’s Satyadvayāvatāra,” 299, n. 70. This is a reference to Atiśa’s
meeting with hierarchs of mNga’-ris after first arriving in Tibet and being ques-
tioned on his view of Madhyamaka. Although the Blue Annals does not mention
this event, the biographies of Atiśa provide brief statements on the questions of
various scholars from Dbus and gTsang and Atiśa’s answers. See Helmut Eimer,
Rnam thar rgyas pa: Materialien zu einer Biographie des Atiśa Dimpaṃkaraśrī-
jñāna (Wiesbaden: O. Harrassowitz, 1979), vol. 2, 191-194; Lokesh Chandra, ed.
Biography of Atiśa and his disciple ḥBrom-ston (Delhi: International Academy of
Indian Culture, 1982), vol. 2: ka 57b and following. The 15th century Bka’ gdams
rin po che’i chos ’byung rnam thar nyin mor byed pa’i ’od stong of bSod nams lha’i
dbang po notes this event as well. See Gianpaolo Vetturini, The bKa’ gdams pa
School of Tibetan Buddhism (Ph.D. Dissertation: School of Oriental and African
Studies, 2007 [Revised 2013]), vol. 1, 65-66. A succinct account is found in Dpa’
bo gtsug lag phreng ba, Mkhas pa’i dga’ ston, 345: “When many scholars asked
the Lord [Atiśa] the view, asking ‘What do you assert?’ and ‘Do you accept the
Madhyamaka, Yogācāra, and so forth according to the ancient ways?’ [He replied,]
‘I posit things according to scripture and logic.’ When asked by a single minded
one from mNga’-ris named Blo gsal, ‘How can this be conferred in common to
all as you do not teach according to assertions?’ [Atiśa replied that] ‘as I am a
follower of the Buddha, what I say is in accord with the thoughts of those to be
trained.’ The others asked, ‘Well then, what do you assert?’ [Atiśa replied that,]
‘I do not assert anything.’ [They then asked,] ‘Since [you] do not have any asser-
tions ultimately, what do you accept conventionally?’ [Atiśa] continued on that
all these [conventionals] are seen like hairs in the vision of one with eye disease.
When asked whether one clears away or does not clear away appearances, [Atiśa
Along these lines, in this section of his biography Atiśa

provides an early brief description and summary of the practice of
Madhyamaka when he states,
For the followers of Ācārya Nāgārjuna, first in the con-
text of conventional reality, one ascertains all things
by means of the two realities; in the context of practice
during meditative equipoise ultimately one is free
from all elaborations; and in the post-meditative state
one construes objects to be like illusions. At the time
of the result when the two realities are actualized,
the form body is conventional reality and the dharma
body is ultimate reality.65
This is basically a brief summation of Atiśa’s longer and shorter
Special Instructions. Before examining the shorter version, I note
Atiśa’s attitude toward the practice of self-cultivation, the practice of
debate, and the use of reasoning.
During his three year residence in western Tibet, Atiśa
translated a number of texts and composed several works as
said,] ‘one purified of eye disease [sees appearances] like cooked rice but you
should analyze . . . ” (mkhas pa mang pos jo bo la lta ba gang bzhed zhus pas khyed
rang gang ’dod gsungs te dbu sems sogs gtam rnying ltar ’di ’dod zhus pa na / kho
bo yang de ltar ’dod de la lung rigs ’di lta bu bdog gsungs / mnga’ ris pa blo gsal
zhes bya ba blo gros zla med pa zhig gis go nas nyid kyi bzhed pa mi gsung bar kun
la mthun ’gyur gnang ba ji ltar lags zhus pas / nga yang sangs rgyas kyi rjes su slob
pa yin pas gdul bya’i bsam pa dang mthun par lan btab pa yin gsungs / gzhan zhig
gis ’o na nyid ji ltar bzhed zhus pas / nga la ’dod pa med gsungs / don dam du bzhed
pa mi mnga’ yang kun rdzobs tu ji ltar bzhed zhus pas / rab rib can gyis skra shad
mthong ba ltar ’di thams cad de ltar gnas pa yin gsungs / snang ba sel lam mi sel
zhus pas ’bras chan la rab rib dag lta bu yin gyi khyed rang dpyod dang gsungs . . . )
The text continues with Atiśa being asked whether a subject appears or does not
appear in the perspective of a valid reasoning consciousness and whether jñāna
exists at the Buddhabhūmi.
65
mChim nam kha’ grags, Rnam thar rgyas pa, 133-134: slob dpon klu sgrub
rjes ’brang dang bcas pas / dang po tha snyad kyi dus su yang chos thams cad bden
gnyis kyis gtan la ’bebs la / nyams su len pa’i dus su yang / [134] mnyam gzhag du
don dam pa spros pa thams cad dang bral ba dang rjes thob tu kun rdzob sgyu ma
lta bu yul du byed pas bden gnyis su nyams su leng zhing / ’bras bu’i dus su yang
bden gnyis mngon du byed de / gzugs sku kun rdzob dang chos sku don dam pa’i
bden pa’o.
well. Among these texts composed by Atiśa, he emphasizes in his

Mahāyānapathasādhanavarṇasaṃgraha that “the marvel of under-
taking the purification of one’s own mind is taught from the Sage’s
sūtras” and disparages the practice of debate when he states,
Neglecting one’s hard to tame mental stream while
practicing argument in order to learn debate [or]
engaging in the explanation of the teaching in every
moment of the day and night for worldly things such
as fame and so forth, life quickly passes on without
purpose and one degenerates from the supreme path.66
In brief, Atiśa and his early Kadampa followers side with Eltschinger’s
assessment of the role of pramāṇa in Buddhist discourse. Atiśa and
his early followers accept the use of logic and epistemology as part of
the five fields of knowledge to refute non-Buddhists and Buddhists.
For Atiśa, the science of epistemology and logic (hetuvidyā) was “a
profane secular science that is common to the Buddhist and other
Indian non-Buddhist schools such as the Naiyāyikas.”67 Pramāṇa is
only utilized at the level of the conventional to refute opponents. But
as I discuss below, for the followers of Atiśa “reasoning” (yukti) sig-
nifies an “internal” Buddhist form of critical analysis that is different
from hetuvidyā, the “external” epistemological devices used to defend
Buddhist Dharma and defeat non-Buddhist opponents. In terms of
Hadot, this use of reasoning is part of the mastery of inner discourse
and dialogue that leads to transformative judgements.68
According to the Extensive Spiritual Biography of Atiśa, after
residing in mNga’-ris for three years, Atiśa travels east across Tibet
66
Mahāyānapathasādhanavarṇasaṃgraha, in Derge Tanjur, vol. 111, folios
299a5-302b6. See 301a-b: [40] / gdul dka’ sems rgyud yin btang nas / / grags
sogs ’jig rten chos kyi phyir / / nyin mtshan gyi ni dus kun tu / / rtsod pa slob phyir
rtsod pa sbyong / [D 301b] / nyan {P,nyams} bshad la sogs chos la ’jug / / tshe ni
don med myur du ’da’ / / mchog gi lam las de nyams ’gyur. See also Sherburne,
Complete Works, 452-453.
67
Helmut Krasser, “Are Buddhist Pramāṇavādins non-Buddhistic? Dignāga and
Dharmakīrti on the impact of logic and epistemology on emancipation,” Hōrin 11
(2004): 129-146, particularly 130.
68
Pierre Hadot, The Inner Citadel: The Meditations of Marcus Aurelius. Michael
Chase, trans. (Cambridge, MA: Harvard University Press, 1998), 55-56, 88.
spending four years in the central area of Dbus-gtsang, and six years in
sNye-thang. In the narrative accounts of Atiśa’s travels he is depicted
as translating texts (gyur), performing rituals (cho ga), giving blessings
of temples, teaching manuals of deity visualization (sādhana, sgrub
thabs), and giving advice for practice (gdams ngag), including advice
on how to practice meditation and on how to practice Madhyamaka.
Taught in Lhasa at the request of his disciple Rngog legs pa’i
shes rab, the Special Instructions contains Atiśa’s advice for self-trans-
formation through Madhyamaka philosophia. In Tibetan catalogs,
the work is given the Sanskrit title, Madhyamakopadeśa, even though
the reconstructed Sanskrit title in all Tibetan versions is madhyama-
upadeśa, instructions on the middle or the center. Potawa’s Middle
Way explains that while all four major traditions of the Buddha
teach a middle way, the instructions of Atiśa concern the middle way
between the two extremes of existence and non-existence based on
the framework of the two realities.69 The Special Instructions, after
formulaic statements regarding the languages of translation, the
translator’s homage, and the author’s homage, may be analyzed as
consisting of instructions on cultivating the three wisdoms (prajñā)
of learning (śrutamayī), reflection (cintāmayī), and meditation
(bhāvanāmayī) within the context of meditative equipoise (mnyam
bzhag, samāhita) and post-meditative (rjes las thob, pṛṣṭhalabdhajñāna)
wisdom construed through the purviews of conventional (kun
rdzob) and ultimate (don dam) reality.70 The instructions conclude
69
See dbu ma’i man ngag gi bshad pa/ pu to yab sras kyi lugs, 320.6-321.3.
70
On samāhitajñāna and pṛṣṭhalabdhajñāna see John J. Makransky, Buddhahood
Embodied: Sources of Controversy in India and Tibet (Albany, NY: State University
of New York Press, 1997), 97ff.; Orna Almogi, Rong-zom-pa’s Discourses on
Buddhology: A Study of Various Conceptions of Buddhahood in Indian Sources
with Special reference to the controversy surrounding the existence of Gnosis
(jñāna: ye shes) as presented by the eleventh-century Tibetan scholar Rong-zom
Chos-kyi-bzang-po (Tokyo: The International Institute for Buddhist Studies of the
International College for Postgraduate Buddhist Studies, 2009), 163ff.; Giuliana
Martini, “A Large Question in a Small Place: The Transmission of the Ratnakūṭa
(Kāsyapaparivarta) in Khotan,” Annual Report of the International Research
Institute for Advanced Buddhology at Soka University 14 (2011): 151-152, n37.
For more on the three wisdoms (prajñā) and their cultivation, see also David
Fiordalis’ contribution to the present volume.
with brief statements on the status of Buddhahood after one attains

the vajra-like concentration (vajropama-samādhi).
The first paragraph of the Special Instructions—for the com-
plete translation see the appendix—from the sentence beginning with
“Conventionally, all things . . . ” to the sentence ending “something the
size of the tip of a hair that is split a hundred times cannot be grasped,”
according to both Prajñāmukti and our Kadampa commentator
indicates the training in the wisdoms of learning and reflection in
relation to conventional reality and ultimate reality. The instructions
indicate that this exercise initially takes place at the level of the
deluded whose vision is narrow, the arvāgdarśana (tshu rol thong ba),
ordinary individuals who cannot understand the two realities nor cog-
nize emptiness.71 At the level of reflection, karmic cause and effect are
considered real as they appear. The phrase “as it appears” (ji ltar snang
ba) occurs in Atiśa’s Satyadvayāvatāra (vs. 21) and is found in works
attributed to Bhāviveka, Jñānagarbha, and Śāntarakṣita.72 The works
71
The Sanskrit equivalents for tshu rol thong ba are arvāgdarśana, arvāgdṛś,
or aparadarśana. As noted by Ryusei Keira, 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 (Wien: Arbeitskreis für tibetische
und buddhistische Studien Universität Wien, 2004), 94, Kamalaśīla explains
in his Tattvasaṃgrahapañjikā that people of narrow vision (tshu rol thong
ba) have three types of direct perception—sense cognition (indriyajñānam),
mental [cognition](mānasaṃ), and reflexive cognition (ātmasaṃvedana), but
such people do not have yogipratyakṣa which directly understands emptiness
(śūnyatā). Atiśa and the bKa’ gdams pa commentary will repeatably mention
that the direct perception and inferences of those with narrow vision cannot
understand the two realities nor cognize emptiness.
72
Satyadvayāvatāra, verse 21: kun rdzob ji ltar snang ba ’di // rigs pas brtags
na ’ga’ mi rnyed // ma rnyed pa nyid don dam yin // ye nas gnas pa’i chos nyid
do /. On the “conventional that appears just as it is” see Malcolm David Eckel,
Jñānagarbha’s Commentary on the Distinction between the Two Truths: An Eighth
Century Handbook of Madhyamaka Philosophy (Albany: State University of New
York Press, 1987), 110-111, n. 7. The phrase ji ltar snang ba = yathādarśana also
occurs in the Madhyamakārthasaṃgraha attributed to Bhāviveka, see Krishna
Del Toso, “Il Madhyamakārthasaṃgraha di Bhāviveka: introduzione, edizione del
testo tibetano e traduzione annotata,” Escercizi Filosofici 6 (2011): 347-365, p.
360. The term yathādarśana may be a contracted form of yathānudarśana found
in Dharmakīrti’s Pramāṇavarttika 3.357ab, for which see Keira, Mādhyamika and
epistemology, 38-46.
of Atiśa and the Kadampa commentaries will repeatedly stress that

the appearances from causes and effects are perceived as real at the
level of conventional reality until reaching the path of vision (darśana-
mārga). The Special Instructions then mentions that, when the conven-
tional as it appears is examined with the great reasons or reasonings,
one gains an ascertainment (niścaya) that nothing can be grasped,
not even minute things, or, as the Kadampa commentary explains,
that nothing is established. The great reasons refers to four reasons
that Atiśa explains in his Bodhimārgapradīpapañjikā73 and also in a
Kadampa manuscript attributed to him called the General Explanation
of, and Framework for Understanding, the Two Realities (bden gnyis
spyi bshad dang/ bden gnyis ’jog tshul).74 The latter text clarifies
73
Atiśa will state four great reasons (gtan tshigs chen po bzhi) in his
Bodhimārgapradīpapañjikā (Sherburne, Complete Works, 230-236), including the
reason refuting production according to the tetralemma (mu bzhi skye ’gog gi gtan
tshigs, catuṣkoṭyutpādapratiṣedhahetu), the diamond-splinters reason (rdo rje
gzegs ma’i gtan tshigs, vajrakaṇahetu), the reason of being neither one nor many
(gcig du bral gyi gtan tshigs, ekānekaviyogahetu), and the reason consisting in
dependent arising (rten ’brel gyi gtan tshigs, pratītyasamutpādahetu). Atiśa leaves
out the reason refuting the production of existent and nonexistent things (yod
med skye ’gog gi gtan tshigs, *sadasadutpādapratiṣedhahetu) that is discussed by
earlier Mādhyamikas like Kamalaśīla (Keira, Mādhyamika and epistemology, 13).
74
In Dpal brtsegs bod yig dpe rnying zhib ’jug khang, Jo bo rje dpal ldan a ti sha’i
gsung ’bum (“Collected Works of Atiśa”), 697-751 (Zi-ling: Krung-go’i-bod-rig-pa-
dpe-skrun-khang, 2006), 706.1-12: “The cause and effect which is intrinsically
established is refuted. The cause and effect of mere appearance [706] is not
refuted. This reasoning of dependent-arising (rten ’grel gyi rigs pa) contains four
reasonings within it. The mutual relation of cause and effect that lacks intrinsic
nature is the very nature of dependent-arising. That an effect occurs when
causes accumulate is the very nature of dependent-arising. The reasoning of the
nature of things (dharmatāyukti) and the reasoning of dependence (apekṣayukti)
of an effect on a cause [demonstrates that it is] suitable for an effect to arise
from a cause. Since it is suitable for reasoning, the reasoning that relates to
demonstration of a proof (upapattisādhanayukti) is said to establish both cause
and effect by the two means of valid cognition. The reasoning of the nature of
things [706.5] is like the body, nature, or shape. That [reasoning] contains within
it at the same time the four reasonings. All the great reasons (he tu chen po) are
grouped within the reasoning of dependent-arising. The diamond-slivers (rdo rje
gzegs ma) is from the point of view of the cause. The negation of existence or
non-existence is through analysis from the perspective of the effect. Free from
the one and the many is in terms of intrinsic nature when examining both the
cause and effect. Moreover, this is the reasoning of dependence (apekṣayukti).
that these reasons are based on the reasoning of dependent-arising

and that all four reasons are accepted as consequences that non-
implicatively negate the intrinsic existence of things, but do not negate
the mere appearance of causes and effects. The mere appearances
that arise from causes and effects are overturned through antidotes
cultivated while practicing the path. Atiśa’s General Explanation
specifies that the object of negation of reasoning is a conceived object
(zhen yul) based on conceptualization that imputes things as either
existent or nonexistent. The object negated by reasoning (rigs pa’i
dgag bya) consists of conceptual thought which imputes objects as
existing with own-character (rang gi mtshan nyid) (708.20-709.1).
Atiśa’s General Explanation offers an early distinction between objects
negated by an antidote (gnyen po’i dgag bya) while implementing
the path, and objects negated by reasoning (rigs pa’i dgag bya)
when searching out the inherent existence of something. Thus,
Atiśa in the first paragraph of the Special Instructions indicates
the “reasoning at the level of reflection” (yukti-cintā-mayī) stage of
this spiritual exercise.
The second paragraph, beginning with the phrase “While sitting
in a cross-legged position . . . ” through to “for as long as the enemies
or thieves of phenomenal marks and conceptual thought does not
arise,” indicates how to cultivate the wisdom arising from meditation
(bhāvanāmayī-prajñā). The stages of meditation are indicated in this
and the following paragraph of the Special Instructions, including
the application, the actual practice, and what occurs in the post-
meditative state. In this dense paragraph, Atiśa instructs that while
seated cross-legged one should contemplate two kinds of entities.
The great reasons (he tu chen po) are accepted as consequences. Therefore, this
reasoning of dependence is the principal. If all things within saṃsāra and nirvāṇa
were intrinsically established, the activity of the nature [706.10], persisting with
a nature that intrinsically exists, by being unfabricated and without change,
would be permanent. Since all things within saṃsāra and nirvāṇa are produced
by causes and conditions, they artificially occur. During the time that a cause does
not cease, the effect does not exist. At just the point when the cause ceases, the
effect is produced. As the effect is produced immediately all at once, the cause
in that way ceases, and this implies that it is impermanent.” For a description of
this text and complete translation, see James B. Apple, “An Early Bka’-gdams-pa
Madhyamaka Work Attributed to Atiśa Dīpaṃkaraśrījñāna,” Journal of Indian
Philosophy 44.4 (2016): 619-725.
The classification of entities into two kinds is also mentioned in Candra-

kīrti’s Pañcaskandhaprakaraṇa and in the Madhyamakaratnapradīpa
in its instructions on rough or “Gross Yoga” (rags pa’i rnal ’byor).75
The Collection on the Two Realities commentary on this point (fol.
11a-b) provides the details that one should practice in solitude and
outlines the corporeal details of body posture, where to place the
eyes, and how to set one’s mouth, as well as the length of time for
meditation sessions.
Atiśa next instructs that one then contemplate material and
non-material entities. Prajñāmukti and our Kadampa commentator
indicate that this contemplation is to be carried out by examining
with reasoning and that these two kinds of entities include all objects
of knowledge. For Madhyamaka thinkers like Atiśa, reasoning (yukti,
nyāya) “designates, in a restrained sense, the fundamental principle
or proposition that enounces the law of causality discovered by the
Buddha that has issued by inductive reasoning, proceeding to a direct
and personal experience.”76 Atiśa does not explicitly state if reasoning
75
Christian Lindtner, “Candrakīrti’s Pañcaskandhaprakaraṇa: Tibetan text,”
Acta Orientalia 40 (1979): 87-145, p. 113: chos la bdag med pa ni chos rnams kyi
ngo bo nyid med pa’o// de la mdor bsdu na chos ni rnam pa gnyis te / gzugs can
dang gzugs can ma yin pa’o / / de la gzugs can la yang rnam pa gnyis te / ’byung ba
dang ’byung ba las gyur pa’o / / gang gzugs can ma yin pa la yang rnam pa gnyis
te / ’dus byas dang ’dus ma bya pa’o / / de la gzug can ma yin pa ’dus byas la yang
rnam pa gsum ste / sems dang sems las byung ba dang / sems dang mi ldan pa’o /
/ gang yang zugs can ma yin pa ’dus ma byas pa ni rnam pa bzhi ste / nam mkha’
dang / so sor brtag pa’i ’gog pa dang / so sor brtags pa ma yin pa’i ’gog pa dang /
chos rnams kyi chos nyid do. For the Madhyamakaratnapradīpa, see Derge Tanjur,
vol. 97 (tsha), 259b3-289a7. See particularly 279a63-b2: slob dpon zla ba grags
pa’i zhal nas / dngos ni rnam pa gnyis te / gzugs can dang / gzugs can ma yin no
/ / gzugs can ni gnyis te / ’byung ba dang / ’byung ba las gyur pa’o / / gzugs can
ma yin pa’i chos ni gnyis te / ’dus byas dang / ’dus ma byas so / / ’dus byas kyi chos
ni gsum ste / sems dang / sems las byung ba dang ldan pa ma yin ba’o / / ’dus ma
byas kyi chos ni bzhi ste / so sor brtag pa’i ’gog pa dang / so sor brtags pa ma yin
pa dang / nam mkha’ dang / chos rnams kyi de bzhin nyid do zhes gsungs mod kyi /
’on kyang ’di dngos po’i chos bsdus par gyur na ni ’di ltar gnyis te / gzugss can dang
/ gzugs can ma yin pa’o / / [279b1].
76
Cristina Scherrer-Schaub, Yuktiṣāṣṭikāvṛtti: Commentaire à la soixantaine
sur le raisonnement ou Du vrai enseignement de la causalité par le Maître indien
Candrakīrti (Bruxelles: Institut belge des hautes études chinoises, 1991), 221, n.
398, and 245, n. 471.
is a pramāṇa or not, but he does demonstrate his understanding

of reasoning as a weapon that dissolves conceptual thought in his
works. Prajñāmukti, the Indian commentator, states that “Reasoning
is a valid cognition which invalidates” (D, 122b), while the early
Kadampa author of the Dbu ma’i man ngag claims (folio 10a) that
“reasoning” (rigs pa, yukti) is neither a direct perception nor an
inference. Be that as it may, the scope of reasoning for these schol-
ars applies to the investigation of the ultimate ontological status
of things and not their mere appearance, which is unexamined
according to Atiśa.77
Based on this understanding of reasoning, or yukti, one
examines material entities, that is, things imputed to be comprised of
collections of atoms, analytically breaking them down by performing
a mereological analysis based on their directional parts. The commen-
tators provide examples of this reasoning procedure from the works
of Śrīgupta, Śāntarakṣita, and Jñānagarbha. Through this reasoning
procedure one ascertains that material entities are not established
and they no longer appear after being dissolved through reasoned
analysis. The instructions then turn toward examining non-material
things, namely the mind, and includes the four aggregates other than
material form. The Special Instructions instructs that,
. . . the past mind has ceased and perished. The mind
of the future has not yet arisen or occurred. Even the
mind of the present is extremely difficult to examine:
it has no color and is devoid of shape, since it is sim-
ilar to space, it is not established, and since it is free
of unity and multiplicity, unproduced, and having a
luminous nature and so forth, when it is analyzed and
examined with the weapon of reasoning, one realizes
that it is not established.
This statement is close to the instructions for examining the mind
that Atiśa gives in his longer version of the Special Instructions,
the Ratnakaraṇḍodghāṭa, where he emphasizes that, “The mind
is without color, without form, by its own nature clear light, and
77
Apple, “Atiśa’s Satyadvayāvatāra,” 292.
unarising from the beginning.”78 This same sequence of qualities is

cited in the Madhyamakaratnapradīpa.79 The Ratnakaraṇḍodghāṭa
cites Nāgārjuna’s Bodhicittavivaraṇa and the Aṣṭasāhasrikā-
prajñāpāramitā to support its claims about the mind being luminous
and unestablished.80 The Collections on the Two Realities states that,
There is natural luminosity and through breaking
down through these reasonings [things] are not
made empty but since they are naturally unproduced,
when they are not elaborated by elaborations or not
conceptualized by conceptions like “this is form,”
“this is not form,” “low” and “excellent,” “large” and
“middle-sized” and so forth, it is called “luminous.”81
The Kāśyapaparivarta (paragraph 98) and possibly other sūtras,
such as the Mañjuśrī-nairātmyāvatārasūtra, influenced Atiśa in his
discussion about the mind.82 However, what is of interest about Atiśa
is that he gives these instructions in the shorter Special Instructions,
as well as in portions of the longer Ratnakaraṇḍodghāṭa, in his own
formulation and integration of Buddhist sūtras and śāstras.
Atiśa then indicates that, just as material and non-material
entities do not have any nature and are not established, “wisdom
itself, without appearance and luminous, is not established with any
nature at all.” Atiśa compares the reasoning process to two sticks,
which after rubbing together and generating a fire, burn up and
become non-existent. Although Atiśa does not state his textual source,
78
Apple “Atiśa’s Open Basket of Jewels,” 125; for the Tibetan text, see
Izumi Miyazaki, “Annotated Tibetan Text and Japanese Translation of the
Ratnakarandodghaṭa–nāma–madhyamakopadeśa of Atiśa.” Memoirs of the
Department of Literature, Kyoto University 46 (2007): A1–A126, p. 6: sems ni kha
dog med pa/ dbyibs med pa/ rang bzhin gyis ’od gsal ba/ gdod nas ma skyes pa’o.
79
Madhyamakaratnapradīpa in Derge Tanjur, vol 97 (tsha), 280a2.
80
Apple “Atiśa’s Open Basket of Jewels,” 127-128.
81
Apple, “Atiśa’s Madhyamakopadésa,” 53; Bden gnyis kyi ’bum (fol. 12b): yang
na rang bzhin gyis ’od gsal ba ste/ rigs pa de dag gis bzhigs nas stong par byed pa
ni ma yin gyi/ rang bzhin skye ba med pa yin pas gzugs can dang de ma yin pa’am/
dman pa dang gya nom pa’am/ che ’bring la sogs pa’i spros pas ma spros shing
rtogs pas ma rtogs pas na ’od gsal ba zhes bya ste.
82
Giuliana Martini, “Large Question,” 151-152, and n. 37.
he draws this example from the Kāśyapaparivarta which is cited in

the Madhyamakaratnapradīpa.83 Kamalaśīla also cites this sūtra in
his Bhāvanākrama and Avikalpapraveśadhāraṇīṭīkā as an example to
illustrate that although the analysis of reality is indeed the nature of
conceptual thought, it will nevertheless be consumed by the fire of
correct wisdom produced by it.84
In his longer Ratnakaraṇḍodghāṭa, Atiśa states that “the
wisdom of individual analysis (so sor rtog pa’i shes rab) itself turns
into clear light.” Prajñāmukti is even clearer in his commentary on
the Special Instructions:
That the very wisdom which individually discriminates
is not established either [means that] [analytical] cog-
nition negates itself. Since wisdom is a particularity of
83
Kāśyapaparivarta (’od srung gi le’u, Toh. 87, dkon brtsegs, vol. cha, 133a.7-
133b.1); “Kāśyapa, it is as follows: For example, wind rubs together two sticks
of wood, from that, fire emerges, and once arisen, the two sticks are consumed.
Similarly, Kāśyapa, when one has correct individual analysis [of things, through
its force] a Noble being’s faculty of wisdom arises. Once produced, correct
individual analysis itself is consumed” (’od srung ’di lta ste / dper na shing gnyis
rlung gis drud pa / de las me byung ste/ byung nas shing de gnyis sreg pa de bzhin
du ’od srung yang dag par so sor rtog pa yod na ’phags pa shes rab kyi dbang po
skye ste / / de skyes pas yang dag par so sor rtog pa de nyid sreg par byed do / /
de la ’di skad ces bya ste/ dper na shing gnyis rlung gis drud pa las / / ma byung
nas ni de nyid sreg par byed / / de bzhin gshegs rab dbang po skyes nas kyang /
/ so sor rtog pa de nyid sreg par byed. Tibetan and Chinese edited by Alexander
von Staël-Holstein, Kāçyapaparivarta: A Mahāyanasūtra of the Ratnakūṭa Class
(Shanghai: Commercial Press, 1926; reprint, Tokyo: Meicho-fukyūkai, 1977),
102, paragraph 69; Sanskrit of this passage not extant, although a citation of
Sthiramati is preserved in the Madhyāntavibhaga-ṭīkā. For the text of Sthiramati’s
Madhyāntavibhāga-ṭīkā, see Martini, “Large Question,” 147: tad yathā, Kāśyapa,
kāṣṭhadvayaṃ pratītyāgnir jāyate iti jātaś ca samānas tad eva kāṣṭhadvayaṃ
dahati. evam eva, Kāśyapa, bhūtapratyavekṣāṃ pratītyāryaṃ prajñendriyaṃ
jāyate jātaṃ ca tām eva bhūtapratyavekṣāṃ dahatīti. Martini’s translation: “Just
as, Kāśyapa, from a pair of firesticks fire is born and, as soon as fire is born, it
burns up that very couple of pieces of wood, exactly so, Kāśyapa, in dependence
on analytical examination of reality the faculty of wisdom is born and, once it is
born, it burns up exactly that very analytical examination of reality” (147).
84
See David Seyfort Ruegg, Buddha-nature, Mind and The Problem of Gradualism
in a Comparative Perspective: On the Transmission and Reception of Buddhism in
India and Tibet (London: School of Oriental and African Studies, 1989), 94-95
and n. 179.
an entity, when an entity is not established, the very

wisdom itself is also not established, just like when
a tree is not established, the wood and so forth are
negated. [121a] As it is said, “In this regard, a fire which
burns fuel, having burned its fuel does not remain.”85
These passages indicate that, for Atiśa and his followers, reasoning is
a conventional process that dissolves itself when seeking to establish
the existence of an object. Analytical reasoning which dissolves itself
is, in a Hadotian reading, philosophy that is preparatory for wisdom,
more specifically, for non-conceptual wisdom (nirvikalpa-jñāna). The
texts suggest a difference between prajñā, or discernment, at the
level of learning and reflection, utilizing reasoning (rigs pa’i shes rab),
and the non-conceptual gnosis that comprises jñāna. The Collections
on the Two Realities contains a number of reasonings, derived from
Jñānagarbha’s Satyadvayavibhaṅga, which lead the reader through
merelogical forms of analysis to dissolve conceptual thought that
reifies things and their relations.
Analysis dissolving conceptual thought while meditating gen-
erates other factors as well. Atiśa notes that in this process faults to
achieving concentration, such as laxity and excitement, are eliminated.
As Potawa’s Middle Way (330.6) explains, in emptiness the faults of
laxity and excitement are no longer established. The aim is to achieve
a non-conceptual realization in which awareness does not apprehend
anything at all. All recollection (dran pa, smṛti) and mental engage-
ment (yid la byed pa, manasikāra) are also eliminated, indicating that
concepts which objectify the past and the future are abandoned. The
Madhyamakaratnapradīpa in outlining its bhāvanākrama chapter
also advocates that bhāvanā should be cultivated to free oneself from
smṛti and manasikāra.86 The attainment of a state of non-conceptuality
(nirvikalpa) is advocated in a number of Mahāyāna Buddhist texts.
85
Apple, “Atiśa’s Madhyamakopadésa,” 15; Madhyamakopadeśavṛtti (D, fol.
120b7-121a1): so sor rtog pa’i shes rab de nyid kyang mi ’grub ste zhes bya bas ni
bdag nyid kyi rtog pa ’gog par byed do / / shes rab ni dngos po’i bye brag yin pas
dngos po ma grub na shes rab de nyid kyang mi ’grub ste / shing ma grub na sha pa
la sogs pa bkag pa bzhin no / / [121a] / / de yang bsreg bya tshig pa’i me / / bsreg
bya tshig nas mi gnas ltar / / zhes pa dang . . .
86
Ruegg, Buddha-nature, 207.
However, unlike other meditation practices imported into Tibet such

as the Great Completion (rdzogs chen) or the practitioners of non-men-
tation (amanasikāra) which advocate non-mind and non-mentation in
their spiritual exercises, The Collections on the Two Realities empha-
sizes that Atiśa’s instructions “develop non-conceptual concentration
in a way which cuts off attachment by means of not finding when
searching through reasoning [13b].” The technique of not finding
when searching through reasoning is mentioned in verse 21 of Atiśa’s
Satyadvayāvatāra, and as Tillemans notes, “there is a quasi-consensus
amongst [Madhyamaka] commentators on . . . unfindability under
analysis.”87 The point in Atiśa’s Special Instructions is that this unfind-
ability eliminates attachment and other negative afflictions. The
instructions then prescribe that consciousness should reside in this
non-conceptual state resulting from analysis so long as the enemies
or thieves, which are phenomenal marks and conceptual thought,
do not arise. The Collections on the Two Realities (13b) explains that
phenomenal marks and conceptual thought are cognitive objects
which scatter awareness away from non-conceptual concentration.
Atiśa’s instructions in the third paragraph briefly explain
the process of slowly arising from meditation and conducting
post-meditative virtuous activities. In a comparable citation in his
Ratnakaraṇḍodghāṭa, Atiśa states,
In this manner, for a long time and with humility,
the bodhisattva should continuously practice the
teachings previously explained, even when not in
contemplation, [and] at the time of meditative stabi-
lization, should cultivate the space-like vajra-samādhi
previously explained. When one has a little clarity
towards the ultimate mind of awakening, and does
not feel one’s own body as existent, one should pacify
the defilements a little, and view all worldly activities
and verbal conventions, all inner and outer objects,
87
Tom J.F. Tillemans, “Trying to be Fair to Mādhyamika Buddhism,” in K.
Preisendanz, ed. Expanding and Merging Horizons Contributions to South Asian
and Cross-Cultural Studies (Vienna: Austrian Academy of Sciences, 2007), 509.
This article has been reissued in Tillemans, How do Mādhyamikas Think? (Boston:
Wisdom, 2016), 19-46.
as like misty vapor (ban bun lang long) with subtle

form. Then, a vast, pervasive, smooth, light, joyful, and
blissful awareness will occur.88
The Collections on the Two Realities provides more details describing
the precise way that one should arise from meditation, as well as
the activities that one should perform. After devoted practice for a
long time one will be able to perceive reality. Based on the Avikalpa-
praveśadhāraṇī, Atiśa states that bodhisattvas see reality in meditative
equipoise and then in the post-meditative state perceive things like
illusions. Direct perception of reality in meditative equipoise causes
one to lessen one’s attachment to objects in the post-meditative state
and the texts provide the example of seeing objects like illusions to
illustrate that one no longer perceives things as substantially existent.
The final paragraph of instructions, beginning with the phrase
“From the point of time when . . . ,” indicates achieving the state of
Buddhahood through attaining the vajra-like concentration (vajra-
upama-samādhi). The vajropama-samādhi is also mentioned in Atiśa’s
Ratnakaraṇḍodghāṭa as indicating the point when one has attained
Buddhahood. At this point, for Atiśa, based on Candrakīrti’s under-
standing of the state of being a Buddha, all mind and mental factors
are cut off and one has fully transformed into the sphere of dharma
(dharmadhātu), being directly fused with the dharmakāya. Because
of dwelling in the dharmakāya, directly fused with reality for as long
as space endures, a Buddha does not possess subsequent attainment.
Atiśa’s understanding of the state of Buddhahood is emphatically
based on the texts of Nāgārjuna,89 and his Kadampa commentators
indicate that this is also the system of Candrakīrti.90 But in brief, the
implication of the last paragraph of instructions is that the cultivation
of the Mahāyāna’s Middle Way is a spiritual exercise that is practiced
in this lifetime, and in future lifetimes, until one attains Buddhahood
for the sake of all beings.
88
Apple, “Atiśa’s Open Basket of Jewels,” 160-161.
89
Apple, “Atiśa’s Open Basket of Jewels,” 163-165.
90
See Apple, “Atiśa’s Satyadvayāvatāra,” and “An Early Bka’-gdams-pa
Madhyamaka Work.”
Conclusion
To conclude, Atiśa’s Special Instructions on the Middle Way
(Madhyamakopadeśa) can function as a spiritual exercise in a Hadotian
manner in that it is focused on prescriptive instructions that evoke a
spiritual transformation, enabling the aspirant to “traverse a certain
itinerary in the course of which he will make spiritual progress.”91 It is
a work which demonstrates “a teacher trying to train and transform an
audience by pedagogical and psychological strategies.” The evidence
for Atiśa receiving and teaching this lineage of Madhyamaka practice,
as embodied within the Special Instructions, furnishes a record of an
actual situation of teaching that demonstrates a credible intersection
between a teacher and his disciples. However, the social context of
Atiśa’s teaching in Tibet moves us away from the institutional context
that a stringent reading of Hadot might demand. In coming to Tibet,
Atiśa is stripped of his social and ritual duties, which he had while
residing in his Indian monastery of Vikramaśīla. Ironically, when
Atiśa arrived in Tibet, because he followed the Mahāsāṃghika-vinaya
instead of the Mulasarvāstivādavinaya already established in Tibet,
he had no other responsibility than to teach Mahāyāna Buddhism in
a purified dogmatic manner. In this way, Atiśa was like a prestigious
guest lecturer visiting a modern university. A prestigious guest
lecturer, rather than assessing student performance and assigning
grades, attending committee meetings, or negotiating with admin-
istrators and colleagues concerning administrative or pedagogical
duties, ostensively teaches and instructs on just the subject matter
at hand. Likewise, while in Tibet, Atiśa instructed on just the prac-
tices of Mahāyāna Buddhism in its Madhyamaka aspects rather than
being caught up in the tangled web of relationships and obligations
of an Indian Buddhist monastery. Perhaps then, the conditions for
Buddhist philosophia, or way of life, is best indicated by Atiśa’s teacher
Bodhibhadra who states: “For those of great learning the happy place
of aging and growing old is in the inner purity of the forest.”92
91
Hadot, Way of Life, 64.
92
Cited in Bodhibhadra’s Samādhisambhāraparivarta (ting nge ’dzin gyi
tsogs kyi le’u), 81b, as from the ’phags pa nyan thos kyi so sor thar pa’i mdo. For
Bodhibhadra’s text, see Tôh no. 3924. Derge Tanjur, vol 101 (ki), folios 79v7-91r6.
Translated by Vinayacandrapa and Chos kyi shes rab.
Appendix
Madhyamakopadeśa, Derge Tanjur, volume Ki, folios 95b1-96a.7
In the Indian language: madhyama-upadeśa-nāma
In the Tibetan language: dbu ma’i man ngag ces bya ba
[Special Instructions of the Middle Way]
I bow down to the Protector of the World
I bow down to that supreme holy person,

whose light rays of speech
Opens the lotuses of the hearts
Of all the deluded like me without exception.
The special instructions of the Mahāyāna’s Middle Way are as fol-

lows. Conventionally, all things, from the perspective of the deluded
whose vision is narrow, including all presentations of cause and
effect and so forth, are real according to how they appear. Ultimately
or actually, when the conventional as it appears is closely exam-
ined and clarified by the great reasonings, one should thoroughly
understand with certainty that even something the size of the tip of
a hair that is split a hundred times cannot be grasped.
While sitting in a cross-legged position on a comfortable
seat, [contemplate] for a while as follows: there are two kinds
of entities, material and non-material. In this regard, material
entities are collections of minute particles. When these are closely
examined and broken up according to their directional parts, not
even the subtlest [part] remains and they are completely without
appearance. Non-material is the mind. In regards to this, the past
mind has ceased and perished. The mind of the future has not
yet arisen or occurred. Even the mind of the present is extremely
difficult to examine: it has no color and is devoid of shape, since
it is similar to space, it is not established, and since it is free of
unity and multiplicity, unproduced, and having a luminous nature
and so forth, when it is analyzed and examined with the weapon
of reasoning, one realizes that it is not established. In this way,
when those two are not established as having any nature at all and
do not exist, the very wisdom which individually discriminates is

not established either. For example, through the condition of fire
occurring by rubbing two sticks together, the two sticks are burned
up and become non-existent. Just as the very fire [D96a1] which
has burned subsides by itself, likewise, when all specific and gener-
ally characterized things are established as non-existent, wisdom
itself, without appearance and luminous, is not established with
any nature at all. All faults such as laxity and excitement and so
forth are eliminated. In this interval of meditation, consciousness
does not at all conceptualize, does not apprehend anything at all.
All recollection and mental engagement are eliminated. Conscious-
ness should reside in this way for as long as the enemies or thieves
of phenomenal marks and conceptual thought do not arise.
When you wish to arise, slowly release from the cross-
legged position and stand up. Then, with mind which sees all
things like illusions,93 do as many virtuous deeds as you are able
with body, speech, and mind. Accordingly, when one practices with
devotion, for a long time, and uninterruptedly, then those with
good fortune will see reality in this very life and all things will be
directly realized, effortlessly and spontaneously, like the center of
space. Through the attainment [of wisdom] after [meditation], all
things are understood to be like illusions and so forth.
From the point of time when one has realized the vajra-like
concentration onwards, [Buddhas] will not have any subsequent
attainment as they are settled in meditative equipoise at all times.
I will not speak here regarding the reasonings and scriptures
that make statements such as, “If it is not like that, what is the
difference from bodhisattvas?” Through the power of gathering
the accumulations and making aspiration prayers for countless
aeons for the welfare of others, [Buddhas] will become just as
those who are to be taught wish [them to be]. There are many scrip-
tures and reasonings [on this topic], but I will not elaborate upon
them here.
93
Literally, “with an illusion-like mind.”
The [text] called Special Instructions of the Middle Way

composed by the paṇḍita Dīpaṃkaraśrījñāna is completed. The
Indian master himself and the great editor, translator and monk,
Tshul-khrims rgyal ba, translated, edited, and set the final version
at the ’Phrul-snang temple in Lhasa.

Apple, James 2010-Spiritual Exercises of Middle Way PDF

Caricato da

Informazioni sul documento

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Apple, James 2010-Spiritual Exercises of Middle Way PDF

Caricato da

Copyright:

Formati disponibili

Buddhist Spiritual Practices

Thinking with Pierre Hadot on

Edited by David V. Fiordalis

Copyright © 2018 by Mangalam Press. All rights reserved.

No part of this work may be copied, reproduced, published, distributed, or stored

Mangalam Press is an imprint of Dharma Publishing.

The cover image depicts a contemporary example of Tibetan Buddhist instructional

♾ Printed on acid-free paper. This paper meets the requirements of ANSI/NISO

Printed in the USA by Dharma Press, Cazadero, CA 95421

Some Remarks on Hadot, Foucault, and

Schools, Schools, Schools—Or,

The Spiritual Exercises of the Middle Way:

Spiritual Exercises and the Buddhist Path:

Philosophy as a Way to Die:

Learning, Reasoning, Cultivating:

Selected Titles from Dharma Publishing

T his chapter provides a close reading of the Special Instructions on

The Attraction of Hadot’s Approach

appropriation of ancient texts.2 This approach is influenced by

comparative and, at times, formal approach to logical

approach rooted in analytical philosophy, more often than not, disre-

Hadot’s Spiritual Exercises and Philosophia

texts not as insufficiently rigorous precursors to the modern system-

Mark Shiffman puts it in his review of Hadot’s book, What is Ancient

Problems in Interpreting Buddhist Thought as Philosophia within

other, ignore one another more profoundly than ever

Schopen’s understanding of the socio-historical situation of middle

in Mūlasarvāstivādavinaya narratives has four features that the

systems with strong apologetical concerns expressing themselves

techniques.”29 This approach of utilizing Hadot, for Eltschinger, also

Eltschinger’s critique thus presents a challenge to which the present

Case Study: Reading Atiśa’s Special Instructions with Hadot

en quoi nous serions disposés à reconnaître de la philosophie. Il n’est ici nullement

Middle Way (madhyamakopadeśavṛtti);34 3) Collection on the Two

in that they bring to light exegetical modes of thinking about Indian

followers of Atiśa interpreted the text. In brief, Prajñāmukti is concise

on his Madhyamakahṛdayakārikā, at Lhasa’s main temple, the ’Phrul-

“instruction” or “teaching.”47 A number of scholastic Indian Buddhist

he will not give instructions about meditation or śamatha, for as

Madhyamaka upadeśas are preserved only in the Tibetan Tanjur.57

of Madhyamaka to his Tibetan students at Vikramaśīla. Here again, we

Along these lines, in this section of his biography Atiśa

well. Among these texts composed by Atiśa, he emphasizes in his

with brief statements on the status of Buddhahood after one attains

of Atiśa and the Kadampa commentaries will repeatedly stress that

that these reasons are based on the reasoning of dependent-arising

The classification of entities into two kinds is also mentioned in Candra-

is a pramāṇa or not, but he does demonstrate his understanding

unarising from the beginning.”78 This same sequence of qualities is

he draws this example from the Kāśyapaparivarta which is cited in

an entity, when an entity is not established, the very

However, unlike other meditation practices imported into Tibet such

as like misty vapor (ban bun lang long) with subtle

I bow down to the Protector of the World

I bow down to that supreme holy person,

The special instructions of the Mahāyāna’s Middle Way are as fol-

do not exist, the very wisdom which individually discriminates is

The [text] called Special Instructions of the Middle Way

Potrebbero piacerti anche