Translating

the Karmapas’ works

Śrīlakṣmīṅkarā’s Advayasiddhi
Śrīlakṣmīṅkarā’s
Advayasiddhi
— gNyis su med par
grub pa’i sgrub thabs —

I
Copyright © 2021 All rights reserved

Editing, Translation of Tibetan and Sanskrit texts, annotations

and introduction: Julian Schott, Torsten Gerloff

Layout: Lajos Dömötör

Drawings on the front cover: Katarína Blaháková
Foto di copertina posteriore: Marcin Muchalski

Cover design: Hana Schott

Back cover photo: Marcin Muchalski

Hanno collaborato all’edizione italiana:

Project manager: Sergio Pestarino
Publication in Italy in: I libri di Marpa

ISBN 978-88-904236-8-0

II
 Śrīlakṣmīṅkarā’s
Advayasiddhi
— gNyis su med par
grub pa’i sgrub thabs —

Critically edited, translated and annotated

by
Torsten Gerloff and Julian Schott

Preface
by
Caroline Gammon

III
IV
Contents
Acknowledgements . . . . . . . . . . . . . . . . . . . . . VI
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . VII
Symbols and Abbreviations . . . . . . . . . . . . . . . . IX
1. Introduction: The Text and its Author. . . . . . . . . . . 1
2. Sources and Previous Scholarship . . . . . . . . . . . . . 9
2.1 Primary Sanskrit Sources. . . . . . . . . . . . . . . 9
2.2 The Tibetan Translation(s) . . . . . . . . . . . . . . 11
2.3 Previous Scholarship . . . . . . . . . . . . . . . . . 14
3. Critical Editions and Annotated Translation. . . . . . . . 18
Conventions. . . . . . . . . . . . . . . . . . . . . . . . 18
Conspectus Siglorum . . . . . . . . . . . . . . . . . . 20
Sanskrit. . . . . . . . . . . . . . . . . . . . . . . . 20
Tibetan. . . . . . . . . . . . . . . . . . . . . . . . . 21
Editions and Translation . . . . . . . . . . . . . . . . . . . 23
Bibliography. . . . . . . . . . . . . . . . . . . . . . . . . 98
Appendices. . . . . . . . . . . . . . . . . . . . . . . . . 103
Additional passage in T (184,5–186,1) — Lo rgyus of
Lakṣmīṅkarā’s Advayasiddhi . . . . . . . . . . . . . . 103
Translation — Lo rgyus of Lakṣmīṅkarā’s
Advayasiddhi . . . . . . . . . . . . . . . . . . . . . . . . 104
Additional passage in T (189,2–190,3)
— Description of the AS . . . . . . . . . . . . . . . . 107
Additional passage in T (189,2–190,3) — Lo rgyus of
the *Sahajasiddhi. . . . . . . . . . . . . . . . . . . . 108
Semi-critical Edition of the canonical
Tibetan Translation . . . . . . . . . . . . . . . . . . . 109

V
Acknowledgements
We would like to express our thanks to the Italian Buddhist
Union (“Piano Economico Progetto 8x1000 U.B.I.”) for their
support, and to Sergio Pestarino for his administrative efforts
and organisation. Thanks also are owed to the Albagnano Heal-
ing Meditation Center (AHMC) in Italy, the place where the
idea of working on the Advayasiddhi initially evolved. We are
especially indebted to the late Lama Gangchen, who supported
and nurtured the idea of working on this project, as well as
Lama Caroline Gammon, whose sincere hope it had been to
fulfill Rinpoche’s wish to publish a small booklet containing
this text.
Moreover, we would like to express our gratitude towards Ha-
runaga Isaacson, our dear Sanskrit teacher, who kindly read
through the editions and had several much-appreciated remarks
and references.
Last but not least, we would like to thank Katherina B for the
beautiful cover picture she has drawn for us as well as for her
efforts in finding advice on how to do so, Paul Partington for
going through the English, and our Italian friends, Sergio and
Kika, for making the effort of translating this work into Italian,
and thus making it available in yet another language.

VI
Preface
Regarding Lakṣmīṅkarā, there is a lot to say about her.  

The elusive Lakṣmīṅkarā has followed me in this life like the

shadow of my body. I have entirely given up trying to determine
an accurate historical account of Lakṣmīṅkarā. I have many
preferred interpretations of course, such as: While king In-
drabhūti was preoccupied with the responsibilities of kingship,
it was his little sister who realised the truth of the Buddha’s
Dharma and gained siddhis. That it was Lakṣmīṅkarā who in-
spired and initiated her older brother Indrabhūti into the path.
Whether she lived in Śrī Laṅka or the Swat Valley, Oḍḍiyā-
na, her sahajayāna approach, glimpsed in her hagiographies,
rings true to me when the hairs stand up on the back of my
neck and the mind feels expansive and unfettered. It gives me
great pleasure to write this preface to this translation of the
Advayasiddhi. 

I would like to congratulate Julian and Torsten on their comple-

tion of this beautiful scholarly critical edition of Mahāsiddhā
Lakṣmīṅkarā’s famous work, the Advayasiddhi. 

I hope they find the sponsorship to be able to continue to make

many other books like this one on the works of the Indian
mahāsiddhas. Till now there has been a lot of confusion re-
garding what the mahāsiddhas actually said, but fortunately
now we have a new generation of excellent Buddhist scholars
like Julian and Torsten. They are equally comfortable trans-
lating both Sanskrit and Tibetan, and thanks to them we can
finally hear the actual voices of Lakṣmīṅkarā and the other
siddhas clearly.

My most recent sighting of Lakṣmīṅkarā happened in 2018,

when Tsawa Tamdrin Rinpoche, a Mongolian Lama and a close
disciple of the late Guru Deva Rinpoche, came to visit Lama
Gangchen Rinpoche in Italy. Tsawa Tamdrin told us about his
auspicious expedition to the Golden Mountains in the Mon-
golian Steppes. He and his attendant Sherab went to explore

VII
some very remote meditator’s caves. Discarded outside of one
of these caves was the body of a yogin, fixed in the cross-legged
meditation posture. The body was what we would crudely term
a “mummy” in the West. It is thought to be three centuries old,
based on the style of his clothing. This body was neither dead
nor alive. His hair was lustrous black, his skin completely per-
fect with the plumpness of a ripe peach and with cherry lips.
And his teeth and nails possessed the glimmer of pearls. The
overall constitution of his body was neither cold nor warm.
The yogin’s body had been unceremoniously dumped outside
of the cave by robbers who had looted the meditator’s cave.
The thieves had even stolen the buttons from the yogin’s shirt.
The cave had been completely ransacked of the altar and ritual
implements and the only thing that remained, the true treasure,
worth more than all the riches of the devas, was a previously
unknown text with the pages stuck together with age, entitled,
in Tibetan, the “Completion Stage of Lakṣmīṅkarā”.

This text is now being restored in Ulānbātor.

Oṁ Aḥ Guru Buddhinī Lakṣmīṅkarā Kāya Vāk Citta Guṇa

Karma Sarva Siddhi Hūṁ

Best wishes,
Lama Caroline, spiritual daughter of TYS Lama
Gangchen Rinpoche, 
Albagnano Healing Meditation Center, Verbania, It-
aly, 16th January 2021.

VIII
Symbols and Abbreviations

𝝨 all sources excluding that mentioned separately

† crux mark
a.c. ante correctionem
BHSD Edgerton’s Buddhist Hybrid Sanskrit Dictionary
cf. confer
cp. compare
conj. conjecture
corr. correction
CTRC China Tibetology Research Center, Beijing
e.g. exempli gratia (“for example”)
em. emendation
etc. et cetera
f(f). (forth)following
fol. folio(s)
IASWR Institute for the Advanced Study of World Religions
ibid. ibidem (“in the same place”)
i.e. id est (“that is/namely”)
MS(S) manuscript(s)
MTM multiple-text-manuscript(s)
MW Sir Monier-William’s Sanskrit Dictionary
NAK National Archives Kathmandu
NGMCP Nepalese-German Manuscript Cataloguing Project
NGMPP Nepalese-German Manuscript Preservation Project
nt. note
om. omission/omitted
p.c. post correctionem
sil. silent
Skt. Sanskrit
s.v. sub verbum (“under the word/entry”)
TBRC Tibetan Buddhist Research Centre
Tib. Tibetan
Tōh. Tōhoku number of the sDe dge version of the bsTan
’gyur (Ui 1934)
v(v) verse(s)

IX
X
1. Introduction: The Text and its Author
Lakṣmīṅkarā’s Advayasiddhi — “The Non-Dual Accomplish-
ment” (henceforth: AS) — was composed in Sanskrit in the leg-
endary kingdom Oḍḍiyāna sometime around the ninth century
of our Common Era.1 This sādhanopāyikā can be classified as
a prakaraṇa type of condensed practice manuals and should
be counted among the most important writings of the female

1
As it holds true for almost all of the mahāsiddhas, the precise identification and
dating of Lakṣmīṅkarā remains difficult, and further research waits to be conducted.
It seems that there were at least three, possibly even five (see Dimitrov 2000), different
persons known by the names Lakṣmī, Śrīlakṣmī, Lakṣmīṅkarā, lCam dPal mo, dPal
ldan dPal mo, dPal ldan lha mo, dGe slong ma dPal mo etc.: 1) the female tantric master
Lakṣmīṅkarā from Oḍḍiyāna, also known as Śrīlakṣmī (Tib. dPal ldan dpal mo / dPal
ldan lha mo), 2) dGe slong ma dPal mo, the founder of a smyung gnas tradition practiced
in Tibet, and 3) Śrīlakṣmī, a Kashmiri scholar who composed a commentary on Nāg-
ārjuna’s Pañcakrama (Tōh. 1842). The life stories and lineage histories of these persons,
it seems, became conflated at a relatively early point of time. Shendge, in an extensive
footnote to her introduction to the editio princeps of the AS, offers an overview of the
information that can be derived from the hagiographical accounts of Tāranātha and Gos
lo tsā ba gZhon nu dpal (Shendge 1964: 11–13 note 18). Abhayadatta’s account of La-
kṣmīṅkarā’s life can be found in Grünwedel 1916: 219–220, Robinson 1979: 250–253,
and Dowman 1985: 372–375. For studies of dGe slong ma dPal mo’s life one may refer
to Vargas-O’Bryan 2001 and Ujeed 2016. A short version of Lakṣmīṅkarā’s life story (lo
rgyus) is also contained in the ’Bri gung bka’ brgyud chos mdzod chen mo, one of the ex-
tra-canonical sources used in this publication for the edition of the Tibetan translation of
the AS. A preliminary edition and translation of this lo rgyus which shows strong resem-
blance to the life stories transmitted in the smyung gnas tradition of dGe slong ma dPal
mo is given in the Appendices. Regarding the dating of Lakṣmīṅkarā, Shendge (1964:
11, note 18) and Dowman (1985: 390–391) date her into the 9th century CE. In Dowman
(1985: 375), moreover, it is remarked that there is another Lakṣmī, the Kashmiri nun
Lakṣmī, who must have lived in the 11th century CE, implying that these Lakṣmīs, i.e. the
authoress of the AS and the later Lakṣmī, are to be distinguished. A similar distinction is
made in Ujeed (2016: 145, 152) wherein Lakṣmī (authoress of the AS) is placed into the
9th, dGe slong ma dPal mo (*Bhikṣunī Lakṣmī) into the 10th, and the Kashmiri Lakṣmī
into the 11th century CE (based on Dimitrov 2000). Péter-Dániel Szántó, on the other
hand, writes that ‘Although Śrīlakṣmi ̄ or Lakṣmīkara ̄ is presented by the tradition as
having lived at least three centuries earlier, it is unlikely that she dates from before the
late 10th century.’ (Szántó 2015: 760). Unfortunately, he states neither the grounds for
his estimation nor the sources that place her in the late 7th century. Apart from the close
relationship of the tantric master Lakṣmīṅkarā to other mahāsiddhas from Oḍḍiyāna
(see note 13), the dates known for the translators of some of Lakṣmīṅkarā’s works, i.e. the
AS and the *Sahajasiddhipaddhati (SSP, Tōh. 2261; see Kragh 2010), ranging from the
late 10th up to the middle of the 11th centuries CE, might also be taken as a slight support
for dating the authoress of the AS into the 9th rather than the 10th or 11th centuries CE.
The Kingdom of Oḍḍiyāna, also referred to as Oḍiyāna and Uḍ(ḍ)iyāna (Tib. O rgyan /
U rgyan), was in all likelihood located in the Swāt valley of present-day Pakistan. For a
discussion regarding the identification and location of Oḍḍiyāna, the reader may refer to
Donaldson 2001: 8–15 and Gerloff 2020: Part 1, 628 note 1.

1
tantric master Lakṣmīṅkarā.2 Although consisting of only 36
verses composed in anuṣṭubh meter, this brevity does by no
means delimit the contents of this remarkable text. Its various
implications require the reader to be familiar with a fair num-
ber of closely related works, many of which are to be counted
among the perhaps most prominent scriptures of Indo-Tibetan
Tantric Buddhism, all together bearing testimony to the earlier
formation of the Indo-Tibetan mahāmudrā doctrine.
The main intent of the AS, as becomes clear from the opening
verses of the AS and the translated passage below from the ’Bri
gung bka’ brgyud chos mdzod chen mo (henceforth: ’Bri gung
chos mdzod), is the accomplishment of the state of Vajrasattva.
Thus, as regards the [title] “Non-dual Accomplishment” (gNyis su
med par grub pa; Skt. Advayasiddhi), ‘non-dual’ (gnyis su med pa;
Skt. advaya) expresses (ni) ‘means and wisdom’ (thabs dang shes
rab; Skt. prajñopāya). ‘Means’ (thabs; Skt. upāya) is the gnosis
that is free from conceptualisations (rnam par rtgog pa dang bral
ba’i ye shes; Skt. nirvikalpakajñāna)[, and] ‘wisdom’ (shes rab;
Skt. prajñā) is luminosity by nature (rang bzhin gyis ’od gsal; Skt.
prakṛtiprabhāsvara[tā]). ‘Non-dual’ is the ‘vajra’ (rdo rje) which
is yoked in union (zung du ‘ jug pa; Skt. yuganaddha), and when
the ‘being’ (sems dpa’; Skt. sattva) has emerged [from the union]
in a bodily form, that one is Vajrasattva (rdo rje sems dpa’). In
his sādhana, the basis for accomplishment (bsgrub gzhi) is the
reality (de kho na nyid; Skt. tattva) that is the common ground of
oneself [and] all sentient beings. After having been introduced to
it thanks to the kindness of the teacher, the one who is constantly
absorbed in meditative concentration becomes accomplished. [...]
The capacity to bring to completion is accomplishment (grub pa;
Skt. siddhi).3
With regard to this accomplishment of the state of Vajrasattva,
the AS is specifically concerned with certain elements of the
practice in connection with ritual worship and offering on the
stage of self-empowerment (svādhiṣṭhānakrama) within the

2
The importance of the AS among the other works of Lakṣmīṅkarā has been pointed
out in the Encyclopaedia of Buddhism (1961–1965, Vol. I: 236f.) which contains a short
description of the text. Another promising work of Lakṣmīṅkarā that awaits study is the
above-mentioned SSP, in which Lakṣmīṅkarā herself supplies an account of her own line-
age (see Kragh 2010). To our present knowledge, this work does not survive in Sanskrit.
A third work from Lakṣmīṅkarā’s hand is a Vajrayoginīsādhana (Tōḥ. 1547). For other
works ascribed to Lakṣmīṅkarā, cf. Robinson (1979: 306) and Dimitrov (2000: 22 note 28).
3
See Appendices: Additional passage in T (189,2–190,3), §1.

2
larger context of the so-called completion-stage practice (ut-
pannakrama, lit. “stage of the arisen”).4
(…) the sādhanopāyikā called Non-dual Accomplishment has
been issued from the lotus-mouth of the Glorious Lakṣmī; it has
as its source the stage of self-empowerment [and] as its central
core the reality of all yogatantras (…).5
In this context, an important aspect of the AS is what can best be
described with the terms vrata and caryā, i.e. the (prescribed)
behaviour and conduct of a tantric practitioner.6 This ‘code of
conduct’, e.g. what is to be done and not to be done, what and
who is and is not to be worshipped, etc., is addressed in another
passage of the ’Bri gung chos mdzod.7 Providing an interesting
statement about how this ‘code of conduct’ as prescribed in the
AS is to be understood in the context of a person’s practice, the
’Bri gung chos mdzod further states:
Regarding that, however, there is a path with concepts (rnam par
rtog pa’i lam), and a path which is free from concepts (rnam par
rtog pa dang bral ba’i lam). For as long as the path mainly con-
sists of concepts (ji srid du rnam par rtog pa gtso bor gyur pa’i

4
The idea that self-empowerment is used particularly in relation to ritual offerings
as part of the completion-stage practices is likewise mentioned in the commentary to
Kongtrul’s bsKyed rdzogs gnad bsdus. Regarding that, see Harding 2002: 114–115. For
a brief overview and useful remarks regarding the concept and development of the svād-
hiṣṭhānakrama, see Shenge 1964: 5–11, English 2002: 372ff. et al.
5
(…) akhilayogatantratattvagarbhā śrīlakṣmīmukhakamalād viniḥsṛtā svād-
hi-ṣṭhānakramodayā advayasiddhir nāma sādhanopāyikā (…). The entire Sanskrit co-
lophon is given at the end of the edition. The expression akhilayogatantra here is to be
understood in the light of a threefold classification of tantra into caryā-, kriyā- and yoga-
tantras. The further sub-classification of the tantras, including the later mahāyoga- and
yoginītantras, does not seem to have been fully established at the time of Lakṣmīṅkarā.
The same holds true for other contemporary works such as Indrabhūti’s Jñānasiddhi (cp.
JS 1.3, 7.6, 12.1, and 21.5) and Vilāsavajra’s Nāmamantrārthāvalokinī (see Tribe 2016:
11–13). The fact that these works contain significant parallels and explicit references to
tantric scriptures that later on became classified as mahāyogatantras and yoginītant-
ras (or: yoganiruttatatantras), such as the Guhyasamāja, Hevajra, and Sarvabuddha-
samāyoga, clearly indicates that statements such as akhilayogatantra, aśeṣayogatantra
and sarvayogatantra are — from a later point of view — not to be understood as referring
to the yogatantras exclusively, but also to the māhayoga- and yoginītantras.
6
This has already been expressed in Krug 2018b. In his dissertation, Krug points out
that the AS is primarily concerned with the concept of “self-consecration” at the ‘stage
and Vajrayāna Buddhist forms of caryā and vrata ascetic practice’ (125 ff.), ‘samaya
offerings’ (177) and ‘sectarian identity’ (329 ff.).
7
See Appendices: Additional passage in T (189,2–190,3), §1a: bya ba dang bya ba ma
yin pa gang yang rung ste | de’i grongs su (read grangs su ?) ’gro ba ni mchod pa’o || de lta
bu’i rnal ’byor gyi ngor mi rung ma ni gang yang med pa’o || bzung ’dzin gyis rnam par
rtog pas rgyun du byas pa’i spang blang ldog ’jug rnams ni ’dir mi dgos pa’o ||

3
 lam), for that long ‘true [tantric] performance’ (dngos su spyod
pa), inasmuch as it is based merely on a firm conviction (yid kyi
mos pa tsam las), is not permitted (ma gnang). Once there is main-
ly non-conceptuality (ji srid rnam par mi rtog pa gtso bor gyur
pa), then ‘true [tantric] practice’ (dngos su spyod pa) is also per-
mitted (yang gnang ba yin).8 Classifications of what is to be done
and not to be done, and of what is necessary and what is not, etc.
(bya ba dang mi bya ba dang dgos pa dang mi dgos pa sogs kyi
rnam grangs ni) are taught according to the scripture (gzhung nas
bstan). This is the concise meaning of the Advayasiddhi.9
To sum up: Lakṣmīṅkarā’s AS constitutes a short practice man-
ual that — from within a “completion-stage view” — is mainly
concerned with discerning correct from incorrect tantric ritual,
practice and view by using various prescriptions, some of which
may appear to provoke or even oppose common ‘codes of con-
duct’. However, it is important to note that this ‘directness’ is
not a manoeuvre to simply counteract certain norms, but one to
display the view and conduct of “Non-Dual Accomplishment”.
This state, by definition, can no longer be bound to any norms,
rules or regulations, and it ought to be free from all concepts in
the widest sense. In Lakṣmīṅkarā’s own words:
gamyāgamyavikalpaṃ tu
bhakṣyābhakṣyaṃ tathaiva ca |
peyāpeyaṃ tathā mantrī
kuryān naiva samāhitaḥ ||21||
The mantrin, remaining in meditative equipoise, should indeed
not make concepts about whom to approach and not to ap-
proach, what to eat or not eat, and likewise what to drink and
not to drink. AS 21

After this brief overview of the main subject matter

and of what may be seen as the central aim of the AS, a few ob-
servations regarding the contextual relations of the work and its
author may be stated. First of all, it has to be noted that almost
all verses of the AS are either equal, or at least significantly
close, to a handful of other tantric texts, most notably to Padma-

8
These statements can be understood as referring to the context of the so-called un-
mattavrata conduct addressed in, for instance, verses 18–23. About this, see notes on
verses 18 and 19 in the translation.
9
See Appendices: Additional passage in T (189,2–190,3), §1b.

4
vajra’s Guhyasiddhi and Indrabhūtis’s Jñānasiddhi, the works
of two contemporary authors whose writings seem to have been
a major inspiration for the authoress of the AS.10 Although the
exact relationship of these authors remains to be studied fur-
ther,11 it is evident that the Indrabhūtis, presumably all three of
them, and Lakṣmīṅkarā had an intimate relationship, while she
also must have had close connections to other important histor-
ical figures such as Padmavajra and Anaṅgavajra the Elder.12

10
The close relationship of the AS, GS and JS can also be established on the grounds
of the subject matter of these works, as has already been pointed out in Krug 2018b,
wherein he states that ‘Three of The Seven Siddhi Texts, Padmavajra’s Guhyasiddhi,
Indrabhūti’s Jñānasiddhi, and Lakṣmīṅkarā’s Advayasiddhi, refer to themselves as a
‘sādhana of vajrasattva’’ (p. 142).
11
Here we would like to draw the reader’s attention to the ongoing research project
‘Indrabhūti’s Jñānasiddhi’, funded by the DFG (Deutsche Forschungsgemeinschaft). In
the course of this project the relation of the above authors and their works will be further
investigated. Regarding this, see also Gerloff and Schott 2020, wherein some indications
underlying this strong and intimate relation of Indrabhūti and Lakṣmīṅkarā are presen-
ted.
12
The major difficulty in establishing the exact relationship between these mahāsidd-
hhas consists in their precise identification. The hagiographical accounts mention at least
three different Indrabhūtis and three different people known by the name Padmavajra. In
addition to this, there also seem to have been at least two persons named Anaṅgavajra.
To complicate the matter even further, we find accounts for different relations, that of
father and daughter, of brother and sister, of teacher and disciple, and of consorts. These
relations are not all necessarily mutually exclusive, and the confusion may partially be
owed also to ambiguous terminology in some of the accounts. The primary evidence
for solving this conundrum is the internal evidence preserved in the texts themselves,
both in their Sanskrit originals as well as in the Tibetan translations. Besides the texts
together with which the AS is commonly transmitted in the Sanskrit and Tibetan sour-
ces, this includes the SSP, Lakṣmīṅkarā’s commentary to Indrabhūti’s *Sahajasiddhi
(Tōh. 2260), and further works belonging to the Hevajra system. The second source of
information consists a) of the more general hagiographical accounts of (historiographic)
writers such as Abhayadatta (Grünwedel 1916: 219–220, Robinson 1979: 250–253, and
Dowman 1985: 372–375), Tāranātha (bKa’ babs bdun ldan, Grünwedel 1914: 32–33 51,
59, 168; Templeman 1983: 25), and Gos lo tsā ba gZhon nu dpal (Blue Annals; Roerich
1995: 362–363, 390), and b) of the more specific lineage histories, etc. preserved in other
collections such as the ’Bri gung chos mdzod and the Sa skya bka’ ’bum. As regards the
’Bri gung chos mdzod, it may be noted here that several passages in volume one contain
further information on the relationship between the three Indrabhūtis and Lakṣmīṅkarā,
namely the hagiographical accounts to the JS (1), AS (2) and the *Sahajasiddhi (3): 1)
sngon U rgyan du rgyal po Indrabhūti (antri b.ho d.hi T) chen po zhes bya ba | sangs
rgyas la dPal gsang ba ’dus pa zhu ba por gyur ba dang | dei’ rjes su U rgyan gyi (gyis
T) rgyal po gsang sngags kyi grub pa brnyes pa Indrabhūti mtshan gsol ba dgu byung
zhes kyang grags (gras T) | Indrabhūti (intra b.ho d.hi T) che ’bring chung gsum du grags
pa las | che ba lCam Lakṣmīṅkarā’i (lcam la khyim ka ra’i) yab | ’bring po rgyal srid
ma spangs par spros bcas kyi spyod pas grub pa mchog gi dngos grub brnyes | chung ba
lCam dpal mo nyid kyi rjes su bzung nas dur khrod du shin tu spros med kyi spyod pa
mdzad pas | Bla ma dGe slong gNyid log pa zhes bya’i | La bai’ na (zla ba’i na T) bza’
can zhes pa’i Dur khrod pa zhes kyang bya’o || […] (T 122); 2) […] [Lakṣmīṅkarā] rgyal
po Indrabodhi’i sras mor sku ’khrungs | […]. (T 184); and 3) […] de la Bla ma dGe slong
gNyid log pa’am | sa’i bdag po dBang po’i Blo Indrabhūti (intra b.hu ti T) chung ba

5
This relation and its further exploration is not only of high signif-
icance for the history of the mahāsiddhas in general, but is also
required for a more nuanced understanding of their treatises. The
textual situation seems to imply that a thorough understanding
of the AS can only be gained through a clear understanding of
the other sources to which the AS is closely connected, in which
many aspects of the contents of the AS are described in greater
detail, and with which it has been transmitted. The important
role of Lakṣmīṅkarā’s opus for the study of the Indo-Tibetan
mahāmudrā teachings and the aspects of its contextualisation
within other scriptures (e.g. those that constitute the Grub pa sde
bdun), is reflected by the fact that already in the 11th century Sub-
hāṣitasaṃgraha — a remarkable compendium of excerpts from
tantric scriptures — a significant portion of the AS is quoted to-
gether with passages from all of the other texts that form the
Grub pa sde bdun.13 That the Grub pa sde bdun forms an integral
aspect in the study of the Indo-Tibetan mahāmudrā teachings of
the bKa’ brgyud school has readily been pointed out on several
occasions,14 and it is once more reflected in the verse preceding
the final paragraph of the ’Bri gung chos mdzod’s addendum. It
states that [subsequent] to having listened to the Advayasiddhi,
the listening to, i.e. study of, the *Sahajasiddhi follows,15 and it
thus invites further investigation of the role of Lakṣmīṅkarā, her
writings and relations.

dang | La ba’i na bza’ can dang | Dur khrod pa rnams mtshan gyis rnam grangs yin la |
des mdzad pa’i Lhan cig skyes grub kyi gzhung dang | de’i ‘grel pa lCam dpal mo nyid
kyis mdzad pa yod pa | […] (T 190). For a preliminary edition and translation of (2), see
Appendices. In view of the fact that we find accounts for both relations, i.e. Indrabhūti
as the father of Lakṣmīṅkarā and as her brother, combined with the fact that there is also
uncertainty regarding which Indrabhūti this relation refers to, one might — following the
implications in the ’Bri gung chos mdzod — consider the hypothesis that Lakṣmīṅkarā
(authoress of the AS and SSP) was the daughter of Indrabhūti I, the sister and disciple
of Indrabhūti II (author of the JS), and the aunt/elder sister and teacher of Indrabhūti III,
also known as Bla ma dGe slong gNyid log pa (author of the *Sahajasiddhi). In the SSP,
Lakṣmīṅkarā herself mentions her relation to Indrabhūti as that of a sister, yet without
clarifying to which of the Indrabhūtis (see Kragh 2010: 196 note 6).
13
These citations in the Subhāṣitasaṃgraha (ed. Bendall 1905) further support the
theory that these works were transmitted together in Indian sources, in a more or less
fixed form, already prior to the 12th century CE.
14
Gerloff and Schott 2020: 249–250, Shendge 1964: 5, Kragh 2010 et al.
15
Cp. Appendices: Additional passage in T (189,2–190,3): §2. Note that there exist at
least two works named Sahajasiddhi. Indrabhūti’s *Sahajasiddhi, which is referred to
here and which, to our knowledge, does not survive in Sanskrit, is not to be confused
with the Sahajasiddhi of Ḍombīheruka (Tōh. 2223) which is extant in Sanskrit (ed. Rin-
poche and Dwivedi 1987).

6
We hope that this new edition of the Sanskrit text, presented
along with the edition of the extra-canonically transmitted ver-
sion of its Tibetan translation — the first to be published in
modern times — together with the new English translation, our
remarks, annotations, and few observations, may contribute to
the collective pursuit of reaching a better and more complete
understanding of these fascinating masters, their lives and
works. It is our heartfelt hope that this may inspire others to
engage with some of the many important questions that still
wait to be answered.

7
8
2. Sources and Previous Scholarship

2.1 Primary Sanskrit Sources

The AS survives to us in Sanskrit, its original Indian language,
in at least five manuscripts, all of which are of a rather recent
date and all of which are to be classified as multiple-text-man-
uscripts (MTM). For the present edition, three of these five
MTM have been employed.16
B: Baroda MS 13124 (= BaMS). A multiple-text- m a n u -
script on grey-black paper kept in the Central Library
Baroda, Oriental Institute, University Baroda, India.
137 folios (ca. 27.5 x 9.5 cm, ‘Newari script’, seven lines
per side, no string hole), undamaged and complete. This
manuscript, which can tentatively be dated to the 19th to
20th centuries CE,17 contains a total of ten works and has
been the single source of the editio princeps. The AS is
contained in folios 35r4 –37r6 (69–73).
K 1: NAK 5-45 (= Saga). A multiple-text-manuscript on
‘Nepalese paper’ in the possession of the National Ar-
chives Kathmandu, Nepal. 49 folios (ca. 32 x 12 cm,
Devanāgarī script, 13 lines per side, no string hole),
undamaged and complete. A total of ten texts are
contained in this MTM. The AS is contained in fo-

16
The two MSS not included in the present study are Sa kha (Śāntarakṣita Library MS
‘21549’, Sarnath: 15 folios, Devanāgarī script, incomplete) and Sagha (Śāntarakṣita Li-
brary MS ‘21537’, Sarnath: 45 folios, Devanāgarī script, complete). A more elaborate de-
scription of the following MSS is forthcoming in Gerloff and Schott 2023 “Indrabhūti’s
Jñānasiddhi: A Revised Critical Edition of the Sanskrit Text and its Tibetan Translation”
(provisional title). For a more detailed description, Krug (2018: 375–383) may also be
referred to. MS K1 is also described in Isaacson and Sferra (2015: 127ff).
17
Shendge (p. 12) dates the manuscript to ‘round about the latter half of the 14th cen-
tury (from the data given by Bendall, in the Cambridge Catalogue of Buddhist Sanskrit
MSS, Cambridge, 1883)’. The script, with its typical “medial” ‘e’ and ‘o’ consisting in
a wavy line (pṛṣṭhamātra), the characteristic style of the initial ‘i’, and other features
indeed correspond to the scripts often classified as ‘Newari’. Nonetheless, when taking
into consideration the script together with the material of the writing support it is rather
feasible to assume that the manuscript originated in Nepal in between the 19th and 20th
centuries CE.

9
 lio 15r2–v12. In the final colophon, no details regarding
the scribe, the place and time of the copying etc. are
mentioned. The manuscript can provisionally be dat-
ed to the 19th or 20th century CE. Photographs of this
MTM are available at the former NGMCP (= NGMPP
A 0134-2 and A 0915-3).
K2: NAK 4-71 (= Saka). Another multiple-text-manuscript on
‘Nepalese paper’ in the possession of the National Archives
Kathmandu, Nepal. 41 folios (ca. 30 x 14.5 cm, 11 lines per
side, no string hole), undamaged and incomplete. Parts of
four texts are contained in this MTM. The AS is found in
folios 20r9–21v4. The date and provenance of this codex can
only be estimated on the basis of the script (Devanāgari )̄ ,
material of the writing support, and its present location. It
is likely to have been produced either in northern India or
Nepal by a northern Indian hand, presumably in the late
19th or, perhaps more likely, in the 20th century CE. Pho-
tographs of this MTM likewise are available at the former
NGMCP. These, however, are split into two parts: NGMPP
A 0112-5 (fol. 3–20) and A 0137-4 (fol. 21–41).

Regarding the above manuscripts, MS K1 stands out as the best

manuscript, wherein many of the scribal mistakes have been
corrected by a second hand. MS B gives the impression of be-
ing the oldest MS in this group, with an overall fair quality. MS
K2, on the other hand, is the most corrupt witness, with many
mistakes and scribal errors. When compared with the other
two witnesses, a rather close relationship with MS B can be
observed. Judging from the scribal mistakes, it is rather unlike-
ly that MS K2 is a direct apograph of MS B. Apart from these
three manuscripts, both of the previous editions of the AS (see
Previous Studies) have been consulted in the course of the pres-
ent study and are reported in the apparatus. Also worth men-
tioning here is a palm-leaf manuscript preserved in the Tibetan
Autonomous Region in the PR China. This MS might possibly
contain a commentary, a ṭīkā, on the AS. Unfortunately, it is not
accessible at the moment.18

18
A copy of this MS should be available at the CTRC, Beijing. According to the so-cal-
led “Sandhak Catalogue” (p. 22, no. 42.2 [bod 25 / da 0404]), this palm-leaf MS is titled
Advayasiddhiṭīkā (“gNyis med grub pa’i rgya cher ’grel pa”) and belongs to the Potala. It

10
Another important Sanskrit source for the study of Lakṣmīṅkarā’s
work, apart from the MSS mentioned above, consists in the tex-
tual material preserved in other texts in the form of direct quo-
tations (SS: AS 4, 10–11, 15, 17–18, 30, 33) and close parallels in
related works (GS, JS, SP, CVP, GST, HT et al.), most notably in
the GS (AS 4, 7, 8, 9, 15–16ab, 19, 22, 24 and 25).
Regarding the transmission history of the work, at least three
stages, or layers, can be detected on the basis of the available
textual material in the MTM, quotations and the extra-canoni-
cal versions of the Tibetan translation:

1. The original reading of the lost archetype

2. An intermediate layer, wherein the older witnesses of
the SS and the Tib. translation of the AS agree against
what is preserved in the younger Sanskrit MSS (cp. vv.
17–18)
3. The text as it is preserved in its present form in the
MTM

2.2 The Tibetan Translation(s)

Lakṣmīṅkarā’s AS is transmitted in Tibetan translation under
the title gNyis med grub pa (short for: gNyis su med par grub
pa’i sgrub thabs) in the bsTan ’gyur section of the Tibetan can-
on (Tōhoku 2220, Ōtani 3064) and outside of it in the text-col-
lections of at least three branches of the bKa’ brgyud school,
the ’Bri gung (T), ’Brug pa (B) and Karma (G) bKa’ brgyud.
While the canonical translation does not contain a translators’
colophon, the extra-canonical versions B and G state that the
AS has been translated by the Indian scholar Śraddhākara-
varman and the Tibetan monk Rin chen bZang po (958–1055
CE). This translation constitutes the oldest surviving witness
of the AS.19 Editions of the canonical translation are published

is reported to comprise six folios. An author is not mentioned in the catalogue. A Tibetan
translation has not been identified.
19
The only other text within the corpus of the Grub pa sde bdun that has also been
translated by Śraddhākaravarman and Rin chen bZang po (TBRC: P753) is Indrabhūti’s
JS. Here, it has been observed that quite a few of those readings which do not agree with
the readings transmitted in the ‘modern’ Sanskrit MTM are in fact confirmed by an older
palm-leaf MS that has come to light in the past decades (cf. Gerloff and Schott 2020).
A similar observation can be made in the passages of the AS that are cited in the SS,

11
in Shendge 1964 and in Rinpoche and Dwivedi 1987 (“Sarnath
edition”). No edition of the extra-canonical version has been
published so far.
When examining the canonical and extra-canonical Tibetan wit-
nesses, it is not only obvious that these represent two different ap-
proaches to translation, but further the impression arises that these
could be based on somewhat different versions of the Sanskrit text.
While it is evident that the sample(s) for the extra-canonical ver-
sion must have been very close to the Sanskrit witnesses available
to us, this is not the case for the canonical translation, which di-
verges both from the Sanskrit and the extra-canonically transmit-
ted versions to a degree that seems indeed out of the ordinary.20 In
fact, these two are so different that the canonical version has been
excluded from the edition presented in this publication.
An example to illustrate this difference between the canonical
and the extra-canonical translations can be found, for instance,
in verse seven, a rather well-known verse, the content of which
is attested in various forms in a number of scriptures. In San-
skrit the verse reads:
yena yena hi badhyante jantavo raudrakarmaṇā |
sopāyena tu tenaiva mucyante bhavabandhanāt ||
The verse as translated, rather literally and correctly, by Śrad-
dhākaravarman and Rin chen bZang po in the extra-canonical
translation, reads as follows:
skye bo rnams ni mi bzad las ||
gang dang gang gis ’ching ’gyur ba ||
thabs dang bcas pa de nyid kyis ||
srid pa’i ’ching ba las grol ’gyur ||
Comparing this with the canonical translation, it becomes ob-
vious that it is not the same translation as that of Śraddhākara-
varman and Rin chen bZang po:

the main source of its edition had been a now-inaccessible palm-leaf MS from the 15th
century CE (cf. Bendall 1905: 1). The life dates of the translators, as may be mentioned
here, mark the AS as well as JS as texts of an earlier translation period.
20 The discrepancy between the Sanskrit version and the canonical Tibetan transla-
tion has already been observed in the previous editions of the Tibetan translation, as it
becomes evident when looking at the notes both in Shendge’s edition as well as in the
Sarnath edition. In the latter, the editors have supplied useful alternative translations of
the Tibetan according to their edition and understanding of the Sanskrit text.

12
 ’di ltar sems can ’ching ba ni ||
drag po’i las kyis sems can rnams ||
de ltar thabs dang ldan pa yis ||
’khor ba dag las ’grol bar byed ||
In this example it is obvious that we have two different kinds of
translations, namely a ‘literal’ one (the extra-canonically trans-
mitted one), and a so-called ‘meaning translation’ (the transla-
tion preserved in the bsTan ’gyur).21 In the context of the Grub
pa sde bdun, no other instance of such a strong discrepancy
between the canonically and extra-canonically transmitted ver-
sions have been observed so far.
Other examples not only show the different approaches of the
two Tibetan translations, but also seem to suggest the possibili-
ty that different Sanskrit exemplars may have formed the basis
of these translations. One of these examples is found in verse
twenty:
jugupsāṃ naiva kurvīta sarvavastuṣu mantravit |
vajrasattvaḥ svayaṃ tatra sākṣādrūpeṇa saṃsthitaḥ ||
In the extra-canonical version, this Sanskrit verse is translated,
rather faithfully and literally, in the following manner:
dngos po kun la sngags shes pas ||
smad pa dag tu mi bya’o ||
rdo rje sems dpa’ rang bzhin der ||
mngon sum gzugs kyis yang dag bzhugs ||
Unlike the version above, the canonical translation here diverg-
es to a degree that we cannot conclude anymore that the two
were based on the same Sanskrit version of the AS:
thams cad dag ni kun la yang ||
rtog pa dag ni mi bya ste ||
rdo rje sems dpa’i bdag nyid du ||
dngos po ma lus gyur par bsam ||
Thus, it has to be suspected that the canonically and extra-ca-
nonically transmitted versions of the AS are not only different
types of translations but also translations of different attesta-

21
The same observation about the canonical Tibetan translation has been made by
Shendge (1964: 13).

13
tions. A further hint that we indeed have different translations
is the fact that the colophon of the canonical version does not
mention the aforementioned translators.
As a consequence of all this, the Tibetan edition pre-
sented in the following is exclusively based on the extra-canon-
ical sources. For referential purposes a ‘Semi-critical Edition
of the canonical Tibetan Translation’, including the readings of
sDe dge (D), Peking (P) and those chosen in the previous edi-
tions (Ba and Sa), is given in the Appendices.

As regards the extra-canonical translation, the versions trans-

mitted within the ’Brug pa and Karma bKa’ brgyud schools
are — apart from minor differences — virtually the same. The
version transmitted within the branch of the ’Bri gung bKa’
brgyud, on the other hand, shows a greater degree of difference.
There are indications that this version has undergone a separate
editorial process, which — even though occasionally offering
preferable readings — has often resulted in readings that are in
fact worse than those of B and G. Interestingly, there is also no
mention of the translators in the colophon of the version trans-
mitted in T, which may be taken as a further support for the the-
ory that this particular version has been subsequently redacted.
Similar observations regarding the relationship between the
translation in the ’Bri gung chos mdzod and the other extra-ca-
nonical versions can be made in the case of Indrabhūti’s JS.

2.3 Previous Scholarship

The beginning point of the ‘modern’ scholarship of the AS is
marked by the publication of the first edition of the text by Malati
J. Shendge in 1964. Despite the fact that the AS, unlike other
texts it is closely related to, was first edited and translated rela-
tively late, it (as well as its authoress Lakṣmīṅkarā) has since then
received some remarkable attention within academia.
In her publication of 1964, Shendge not only published the San-
skrit edition of the AS based on Baroda MS ‘13124’ (MS B
above), but she also accompanied it with the first edition of the
canonical Tibetan translation, along with an annotated English
translation of the text. In the introduction to her study, Shendge

14
draws attention to Nāgārjuna’s Pañcakrama (PK), which she
rightly introduces as a scripture of particular interest for the
understanding of this sādhana and of concern for the contex-
tualization of the AS.22 Considering both the limited material
available at that time as well as the early stage of her career at
which Shendge conducted her work, her study deserves recog-
nition for its quality and the many important contextual refer-
ences given, making it a cornerstone in the modern history of
the study of Lakṣmīṅkarā’s work.23
In 1987, the AS was edited for a second time by Samdhong
Rinpoche and Vrajvallabh Dwivedi within the context of the
so-called Guhyādi-Aṣṭasiddhi-Saṅgraha. In this publication
further witnesses of the Sanskrit text have been employed, and
both the Sanskrit text as well as the canonical Tibetan transla-
tion of the AS were improved significantly in several places.
Apart from the textual improvements of the AS,24 another qual-
ity of this publication — and perhaps the most important one
— is to be seen in the fact that the AS has been made available
together with seven other, closely related works both in their
Sanskrit originals as well as their Tibetan canonical transla-
tions. By this, the study of these works, of their contents, tex-
tual background and intertextual relations has been facilitated
significantly.

22
This includes the mentioning of the importance of the Grub pa sde bdun for the study
of the AS, perhaps the earliest mention in academia. Another quality of Shendge’s study
consists of the many references to parallel passages in the GS, JS, PVS and HT which
she supplies in the notes to her critical edition, and in further references in her annotated
translation.
23
Relying on Shendge’s work, and in some sense continuing it, the AS was again “edi-
ted” and translated in 1995 by Ramprasad Mishra under the title ‘Advayasiddhi, the
Tāntric View of Lakṣmīṅkarā’. In his publication, Mishra first focuses on the authoress
of the AS by paraphrasing the influences and historical data of her life. Following it, Mi-
shra prints the Sanskrit text, unfortunately without a critical apparatus, with no recourse
to the various sources that had come to light since the work of Shendge, and without
improving the editio princeps. His plain, but yet rather “inventive” English translation is
accompanied by a lengthy section discussing the “principals” of the AS in which Mishra
touches upon a variety of topics, providing a general introduction to the mantranaya.
Other than stated in the Preface to this publication (p. 9), this work unfortunately offers
very little to the ‘need of the researcher’.
24
Despite the many improvements of this new edition of the AS, two points of criti-
cism may be mentioned here briefly. The first one concerns the sometimes misleading
and partly incorrect reports of the variants in the MSS and xylographs. Regarding the
‘Baroda’ MS, the Sarnath edition seems to be relying exclusively on the editio princeps
by Shendge. The second point of criticism regards the lack of references to parallels in
other works.

15
The first study to look at the AS in its broader context, i.e. in
relation to the so-called Grub pa sde bdun and thus using the
Guhyādi-Aṣṭasiddhi-Saṅgraha’s potential, was conducted in
the course of the doctoral research of Adam Charles Krug, the
results of which were submitted to the University of California
in 2018. Therein, not only various important doctrinal topics
are addressed, but also various verses of the AS are translated,
put it into context, and compared with other texts contained
in the Grub pa sde bdun.25 Krug’s study, as well as his other
contributions (2018a and 2020), certainly deserve recognition.
Concerning the ground that had been covered, it is only natural
that the individual texts, and in relation to the AS most notably
the JS and the GS, could not be studied in due detail. The close
study of each of these texts, and, following that, the study of
their relation, remains a task yet to be fulfilled.26
Despite the important scholarly contributions that have been
made to establish the textual basis for the further research of
Lakṣmīṅkarā’s AS, the comprehensive study of the life and
works of the female tantric master Lakṣmīṅkarā, and her re-
lation to the authors of the texts together with which the AS
is transmitted in the Sanskrit and Tibetan sources, remains a
notable gap in our knowledge. Despite the fact that the hagiog-
raphic sources of Lakṣmīṅkarā by Abhayadatta, Tāranātha and
Gos lo tsā ba gZhon nu dpal have been translated and studied
in some detail (see note 2), the problem of the identification
of the various authors known as Lakṣmī remains, although it
has received some noteworthy scholarly attention in the past
decades. Dragomir Dimitrov (2000) briefly addresses this
problem and identifies five different historical figures, offer-
ing a comprehensive overview of the various Lakṣmīs based
on their writings and the available hagio- and historiograph-
ical information. The contributions of Ulrich Timme Kragh
from 2010 and 2011 and the topics raised therein offer further

25
In Krug 2018b, the following verses of the AS are translated: 1–2 (pp. 125, 142–142),
4 (p. 177), 8 (p. 177), 9–12 (pp. 126, 239–330), 14 (p. 158), 15–17 (pp. 126–127, 177), 26
(p. 177) and 36 (p. 128).
26
Regarding this, see also Gerloff and Schott 2020, wherein the need of a thorough
reassessment of the JS, one of the two texts most closely connected to the AS, is voiced.
Also contained in this article is the translation of another passage from the ’Bri gung
chos mdzod from which it becomes clear that the AS is not only closely related to, but in
fact part of an overarching textual structure connected to, and expressing different facets
of, the mahāmudrā doctrine.

16
important observations for the study of the life and works of
Lakṣmīṅkarā. In his article of 2010, Kragh points out the im-
portance of the SSP for the study of Lakṣmīṅkarā’s life and her
relation to Indrabhūti. Kragh shows that the work must have
been translated in the middle or late eleventh century, making it
a rather early account of translated life stories within the Tibet-
an textual landscape. Other potentially important sources to be
studied in relation to Lakṣmīṅkarā as the authoress of the AS
are the hagiographical accounts (rnam thar) of dGe slong ma
dPal mo and the so-called smyung gnas tradition. Although it is
almost certain that these have to be different persons, the simi-
larities of their accounts make dGe slong ma dPal mo’s smyung
gnas tradition an indispensable subject in the future study of
Lakṣmīṅkarā’s life.27 A comprehensive study comparing these
accounts remains a desideratum.
As of today, and without demeaning any of the aforementioned
important contributions, it remains indeed possible to build
upon the “earlier” academic exploration of tantric sources by
carefully re-evaluating them with the aid of the resources and
means that are now, almost 60 years after Shendge conducted
her first research of the AS, at our disposal.

27
See also notes 2 and 335. Regarding this, two important articles have been publi-
shed by Ivette Vargas-O’Bryan (2001) and Sangseraima Ujeed (2016). The very title of
Ujeed’s article ‘Dge-slong-ma dpal-mo, the Princess, the Mahasiddha, the Nun and the
Lineage Holder’ already implies the problems of possibly conflated information, persons
and narratives. This situation, in which various persons and traditions seem to be in-
terwoven by, at least, some Tibetan traditions, can be demonstrated exemplarily also via
the lo rgyus of the AS as transmitted in ’Bri gung chos mdzod (see Appendices). The fact
that Lakṣmīṅkarā, the female tantric master from Oḍḍiyāna, upholds a strong position
against smyung ba and other rituals in the AS (cf. AS 3, 14, 24 etc.) makes it rather infea-
sible to assume that she could be the same person as dGe slong ma dPal mo, the founder
of a smyung gnas tradition practiced in Tibet.

17
3. Critical Editions and Annotated Translation

Conventions
In approaching the task of translating the AS, we have taken re-
course to the fundamentals of what we perceive to be the classi-
cal approach towards translation in the field of Buddhist Stud-
ies. That is to say, the translation is based on our edition, i.e. an
honest attempt to reconstruct the Sanskrit and Tibetan originals
of our version of the text with preference given to the Sanskrit,
the original language of the composition. In the edition itself,
presented with a positive critical apparatus, we have based our
decisions on the premise to choose those readings that, when
this seems possible, can explain later developmental stages of
the text (generic principle) and that reflect what we believe to
be in line with the “original flavour” of the text.28 That is to say,
we have, in some instances, consciously refrained from stan-
dardising or correcting the text towards a smoother or, formally
speaking, more correct version, when we felt certain “incon-
sistencies” or possibly problematic formations to be part of the
original composition. However, this edition — as is the case for
any edition — is nothing but a hypothesis. Every improvement
is very much welcomed by the editors.
In the translation, we have aimed at consistency and objectiv-
ity, trying to neither overinterpret nor simplify the text, but to
translate it in an objective manner as close to the grammar, style

28
In the apparatus, first the lemma is given, followed by the siglum of the source(s) of
this reading or by an editorial mark such as em. in the case of emendations. Separated
by a right-hand bracket, the variants are supplied, each followed by the siglum of its
source(s). In the course of editing the Sanskrit and Tibetan texts, we have silently stan-
dardized minor orthographical variants such as geminations after a repha (ryya, rvva
etc.), de-geminations (tva as in sattva, etc.), homogenic nasals and the like, and we have
applied the standard rules of sandhi where this was deemed appropriate. In the Tibetan
edition, we have silently corrected minor inconsistencies of the use of ba and pa. The
punctuation too has been standardized silently, and verse numbers have been supplied
for referential purposes. These are purely editorial. Names of persons, places and works
are printed with an upper-case initial; text titles are rendered in italics. The Sanskrit text
is printed in Roman script according to the International Alphabet of Sanskrit Translite-
ration (IAST). The Tibetan is transliterated in accordance with the Wylie system.

18
and vocabulary of the original as was possible without overly
harming the readability of the English. We have annotated our
translation according to our understanding of the text with both
philological as well as contextual, and sometimes doctrinal
notes, addressing major doubts and problems, and supplying
information necessary for the correct understanding of the con-
tents of this work. References regarding citations and parallels
to other primary sources, whenever these could be identified,
are supplied in the final note at the end of each Sanskrit verse.
Occasionally, translations of such parallel passages have been
supplied when these seem to contain further useful informa-
tion. These, together with all other annotations, are given in a
single note at the end of each translated verse and follow the
sequence of the pādas, i.e. quarters of each verse, to which they
refer. General thoughts about the doctrines and content follow
those that are of a more “technical” nature, such as philological
remarks and the like.

19
Conspectus Siglorum

Sanskrit
B Advayasiddhi: Oriental Institute Baroda MS 13124: fol.
35r4–37r6 (69–73)
Ba Shendge’s edition of the Baroda MS from 1964
BaMS Shendge’s report of readings of the Baroda MS given in
her notes from 1964. In cases where no variant reading
is given, Ba and BaMS are supposedly the same.
K1 NAK 5-45 (= NGMPP A 0134-02 and A 0915-03):
fol. 15r2–v12; this MS is also described in Isaacson and
Sferra 2015: 127ff.
K2 NAK 4-71 (split into two parts: NGMPP A 0112-5 and
A 0137-4): fol. 20r9–21v4
Sa Sarnath edition
Saka Sarnath’s report of readings of the “ka” MS (= K2)
Sakha Sarnath’s report of readings of the “kha” MS
(Śāntarakṣita Library MS ‘21549’)
Saga Sarnath’s report of readings of the “ga” MS (= K1)
Sagha Sarnath’s report of readings of the “gha” MS
(Śāntarakṣita Library MS ‘21537’)
Samu Sarnath’s report of readings in the editio princeps

AS Advayasiddhi of Lakṣmīṅkarā
CVP Cittaviśuddhiprakaraṇa of Āryadeva (ed. Patel 1949)
GS Guhyasiddhi (ed. Rinpoche and Dwivedi 1987)
GST Guhyasamājatantra (ed. Matsunaga 1978)
HT Hevajratantra (ed. Snellgrove 1959)
JS Jñānasiddhi of Indrabhūti (ed. Rinpoche and Dwivedi
1987)
MĀ Muktāvalī, a pañjikā by Ratnākaraśānti on the Hevajra­
tantra (ed. Tripathi and Negi 2001)
PVS Prajñopāyaviniścayasiddhi (ed. Rinpoche and Dwivedi
1987)
SP Svādhiṣṭhāna(krama)prabheda of Āryadeva
(ed. Pandey 1990)
SS Subhāṣitasaṃgraha (ed. Bendall 1905); MSS: 1) NAK
3-652 (NGMPP A 1057-20) and 2) Takaoka CH 319
SSP *Sahajasiddhipaddhati of Lakṣmīṅkarā (Tōh. 226)

20
Tibetan
B “gNyis su med par grub pa’i sgrub thabs”. In: ‘Brug
lugs chos mdzod chen mo. Khams sgar gsung rab
nyams gso rgyan spel khang (Drukpa Kagyu Heritage
Project): Kathmandu. 200?. Vol. 35, 269–276. (TBRC:
W23779)
G “gNyis su med par grub pa’i sgrub thabs”. In: Kar-
ma pa sku phreng rim byon gyi gsung ‘bum phyogs
bsgrigs rnam par rgyal ba dpal zhwa dmar ba’i chos
sde. dPal brtsegs bod yig dpe rnying zhib ’jug khang:
Lha Sa. 2013. Vol. 36, 230–235.
T “gNyis su med pa grub pa”. In: ’Bri gung bka’ brgyud
chos mdzod chen mo. ’Bri gung mthil dgon: Lha sa.
2004. Vol. 1, 186–1

21
22
 Editions
and
Translation

23
oṁ29 namaḥ śrīvajrasattvāya ||
’jam dpal gzhon nur gyur pa la phyag ’tshal lo ||

Verse one

prakṛtiprabhāsvaraṃ nāthaṃ30
sarvajñaṃ31 tribhavodbhavam32 |
praṇamya śirasā vajram
īpsitārthaphalapradam ||

mgon po33 rang bzhin ’od gsal ba ||

kun mkhyen srid pa gsum ’byung ba34 ||
’dod pa’i don gyi35 ’bras bu stsol ||
rdo rje can la spyi bos ’dud ||

29
oṁ B K 2 Ba Sa ] om. K1 Sa ka Sa kha
30
°aṃ 𝝨 ] °a° K 2
31
°aṃ 𝝨 ] °a° K 2
32
tri° 𝝨 ] vi° Sa ka Sa kha
33
mgon po T ] mgon po’i B G
34
’byung ba B G ] ’byung zhing T
35
don gyi B G ] don yis T

24
Oṁ — Obeisance to the Glorious Vajrasattva!36

Verse one

Having bowed down with [my] head to the vajra, the protector
who is luminous by nature, the omniscient one who arises in
the three existences, and who bestows the fruit of the desired
aims [of sentient beings];37

36
Tib. reads “Obeisance to Mañjuśrī Kumārabhūta”.
37
Pāda a of the Sanskrit is hypermetrical, reading nine instead of eight syllables.
Thus, the pāda needs to be recited slightly faster than a common line in anuṣṭubh. We
take tribhava as referring to the Three Spheres (traidhātu, khams gsum), i.e. the Three
Realms (kāma-, rūpa- and ārūpyadhātu) as expressed in the canonical translation. An
alternative interpretation, perhaps less likely, is to understand tribhava in the sense of
tryadhvan, i.e. past, present and future. Alternatively, one may understand the compound
tribhavodbhavam in an instrumental sense, i.e. he by whom is given rise to the three
existences (cp. Shendge 1964: 24; Mishra 1993: 31). In the third pāda, the Tibetan tran-
slates vajriṇam (accusative of vajrin), i.e. “he who possesses the vajra”, here most likely
referring to Vajrasattva.

25
Verse two

deśakālatithivāra-
nakṣatrair38 maṇḍalair vinā |
vakṣye ’haṃ39 vajrasattvasya
saṃkṣepāt40 sādhanaṃ param ||41

yul dang dus dang tshes grangs gza’42 ||

rgyu skar43 dkyil ’khor med par44 ni ||
rdo rje sems dpa’i45 sgrub pa’i thabs46 ||
mdor bsdus par47 ni bshad par bya ||

38
°air 𝝨 ] °ai K 2
39
vakṣye ’haṃ Ba Sa ] vakṣaha B Ba MS K 2; vakṣeha K1
40
saṃkṣepāt 𝝨 ] saṃkṣapāt K 2
41
Cp. GS 4.61: deśakālatithivāranakṣatraṃ bāhyamaṇḍalam | ebhir vinā pravakṣyāmi
sādhanaṃ vajradhāriṇaḥ ||.
42
gza’ B G ] bzang T
43
rgyu skar G T ] sgyu skar B
44
med par B T ] med pa G
45
sems dpa’i em. (cp. D P)] sems dpa’ B G T
46
sgrub pa’i thabs B G ] bsgrub pa’i thabs T
47
bsdus par ni B G ] bsdus pa ni T

26
 Verse two

I will explain in a condensed manner the supreme

sādhana of Vajrasattva, without maṇḍalas, place, time,
a succession of lunar days, or lunar constellations.48

48
Note that the first pāda of the Sanskrit is unmetrical. The same reading, however, is
used in GS 4.61a. The exact meaning of tithi-vāra is slightly unclear. The translation “the
succession of lunar days” is inspired by the use of vāra as a kind of final maker: “the time
fixed or appointed for anything [...] a person’s turn [...] often, esp. with numerals, = times
[...]” (MW s.v. vāra). Alternatively, tithi-vāra might be taken as a dvandva (cp. Shendge
1964: 25; Mishra 1993: 31; Krug 2018: 125, 142) which also seems to be implied by the
Tibetan tshes grangs [dang] gza’ (BG; bzang T). When taking vāra as a separate word
it might be translated as “week-day,” or more broadly as “occasion” or “circumstance”
etc. within an astrological context. It should be further noted that the Tibetan rendering
of vāra as gza’ (or bzang for bzung ?) — which more commonly might be a rendering of
graha, i.e. “planet” — and the exact force of this choice, also remain unclear. In the last
pāda, unlike the canonical translation, the extra-canonical versions do not render param.

27
Verse three

niyamavratopavāsair
akṣaroccāraṇabhāvanaiḥ49 |
atattvayogī50 na sidhyet51
kalpakoṭiśatair api ||52

sdom dang brtul53 zhugs smyung ba54 dang ||

yi ge brjod dang bsgom pa yis || [B 271]
bskal pa bye ba rnams su yang || [G 231]
rnal ’byor de nyid55 med mi ’grub ||

49
°bhāvanaiḥ Ba Sa ] °bhāvaṇaiḥ B K1 K 2 Ba MS
50
°yogī B K1p.c. K 2 Ba Sa ] °yogo K1a.c. Saga Sagha
51
si(d)dhyet B K1 K 2 Sa ] siddhayet Ba
52
A citation of this verse is contained in the anonymous Yuktipradīpa (see note below).
Cp. JS 1.86: vratopavāsaniyamair devatārūpabhāvanaiḥ | nānābhuja-samāyuktaiḥ
sidhyate na hi sādhanam ||.
53
brtul B G ] rtul T
54
smyung ba B G ] smra ba T
55
de nyid B G ] om. T (In T we find a kākapada “x” written below the mi, perhaps an
insertion mark. There might have been a corresponding marginal note on the right-hand
bottom of the folio which, however, due to the quality of the scans, is faint and not legible.)

28
 Verse three

Even through hundreds of millions of aeons, the one who

has no union with reality (atattvayogin) will not become
accomplished through tantric vows, observances, fasting,
recitation of syllables and meditations.56

56
A citation of this verse is contained in the anonymous Yuktipradīpa. For a descrip-
tion of this short work, cf. Szántó 2015: 760. Negatives of an ‘old’ palm-leaf MS of this
text which was discovered by Giuseppe Tucci in the Sa skya Monastery in Tibet are
preserved in Giuseppe Tucci’s Collection (envelope 7/Cc; see Sferra 2000: 409). The
readings in this verse remain problematic. In the first pāda, one would rather expect
to read vratopavāsaniyamair for metrical reasons. A similar formulation is used in JS
1.86a (see note above). In the second pāda, one could consider emending the reading
to akṣaroccārabhāvanaiḥ to prevent the hypermetricism. Such usages can be found
in other tantric works such as, e.g., in the Svacchandabhairavatantra (5.55). Yet, the
MS of the Yuktipradīpa to some degree confirms the hypermetrical reading, attesting
akṣaṇoccāraṇabhāvanaiḥ. The third pāda too — as it stands — is metrically bad. For a
‘smoother’ ra-vipulā there should be a caesura after the fourth syllable. We have taken
the word atattvayogī as a nañ-tatpuruṣa compound referring to the practitioner on the
level of the utpannakrama. The term tattvayoga (literally: “union with reality”) can
refer to different technical procedures in the various Buddhist tantric systems (Guhya-
samāja, Hevajra and Kālacakra). In this (AS 10, 25) and other closely related texts (GS
chapters 6–8 PVS, chapter 5 et al.), the term is specifically used in the contexts of the
yogic practices of ritual worship and sexual union synonymously with expressions such
as tattvabhāvanā, tattvacaryā etc. A similar formulation, and perhaps an allusion to
this verse or its underlying source, can be found in Kelikuliśa’s commentary on the HT,
the Trivajraratnāvalīmālikā, reading nātattvayogī sidhyatīti vacanāt. A palm-leaf MS
of this work is likewise preserved in Giuseppe Tucci’s Collection (folder 43; see Sferra
2000: 411).

29
Verse four

viḍva[K2 20v] jrodakabījādyair57

nāsikābhyantarodbhavaiḥ58 |
pūjayet satataṃ mantrī
svātmānaṃ59 tattvabhāvanaiḥ60 ||61

bshang dang rdo rje chu sa bon62 ||

sna yi63 nang nas64 byung ba dang ||
de nyid bsgom pas65 bdag nyid ni ||
sngags pas rtag tu mchod par bya ||

57
viḍvajrodaka° Sa ] vidvajodaka° B; vidvajrodaka° K1 K 2 Ba • °bījādyair K1 Ba Sa ]
°bījādyai° B K 2 Ba MS
58
°ābhyantarodbhavaiḥ Ba Sa ] °ābhyantarodbhavaḥ B K1; °ātyantarodbhavaḥ K 2
59
svātmāṇaṃ em. (cf. SS) ] ātmāṇaṃ 𝝨
60
tattvabhāvanaiḥ Ba Sa ] tasvabhavanaiḥ B K1 K 2; tatsvabhavanaiḥ Sa kha Sagha
61
With reference to Mahālakṣmīpāda (!), this verse is quoted in SS (ed. Bendall 1905:
40–41): Mahālakṣmī-pādair apy uktaṃ | dvivajrodakabījādyair nāsikābhyantarodbhava-
iḥ | pūjayet satataṃ mantrī svātmānam tattvabhāvanaiḥ ||; cp. GS 6.72cd–73ab which
reads svadehe vajrabhāvanaiḥ in pāda d.
62
bshang dang rdo rje chu sa bon BG] dri dang rdo rje chu la sogs T
63
sna yi B G ] sna’i T
64
nang nas B G ] nang na T
65
bsgom pas B G ] bsgom ba T

30
 Verse four

The mantrin should always worship himself by [frequent]

cultivations of reality, with excrement, vajra-water (urine),
seed (bīja) and so on, and with what arises within the nose.66

66
This verse is quoted in the SS. A close parallel is also found in Padmavajra’s GS
(6.72cd–73ab) which reads svadehe vajrabhāvanaiḥ in pāda d (“[he should make offe-
rings] to his own body through vajra-meditations”). It is noteworthy that the author of
the SS attributes this verse to Mahālakṣmīpāda (!), and not Padmavajra, the author of the
GS, which evidently was also known to him. In the first pāda, the so-called Five Nectars
(pañcāmṛta, bdud rtsi lnga) are addressed. Although in Buddhist tantric texts the term
bīja is most frequently used in the sense of “seed-syllable”, we have taken it here as a sy-
nonym of śukra (semen). As regards the second pāda, the exact meaning and syntactical
function of the content is somewhat uncertain. It is also possible, although less probable,
to take it as an adjective qualifying the content of the first pāda, a choice that has been
made by Krug, who also translated this verse (2018b: 177). Both Shendge and Mishra
have adopted a similar interpretation as is presented here. In pāda d, we have decided
to adopt the reading svātmānaṃ as it is preserved in the SS. It seems that the sva° had
dropped out at some point in the course of the transmission of the AS, and that it was
mistakenly re-inserted at the wrong position as a substitute of the missing °tva° in ta(t)
tvabhāvanaiḥ. Whether or not this reading is original is impossible to say, yet it prevents
the hiatus between mantrī and ātmānaṃ. Such a hiatus, however, occurs at one further
place in the AS, namely in verse 27ab.

31
Verse five

jananīṃ67 bhaginīś68 caiva

duhitṛbhāgineyikān69 |
prajñopāyavidhānena
pūjayed yogavi[B 35v]t sadā ||70

rang ma sring mo71 nyid dang ni ||

bu mo dang ni tsha mo rnams ||
shes rab thabs kyi72 cho ga yis73 ||
rnal ’byor shes pas rtag tu mchod ||

67
jananīṃ K1p.c. Sa ] jananī° B K1a.c. K 2 Ba Sa ka Sa kha Saga Sa mu
68
bhaginīś 𝝨 ] bhaginīṃś K1p.c.
69
°bhāgineyikān K1p.c. Ba Sa ] °bhāganeyikān B K1a.c. K2 Saka Sakha Saga; °bhāganeyikā BaMS
70
Cp. PVS 5.25: janayitrīṃ svasāraṃ ca svaputrīṃ bhāgineyikām | kāmayan tattvayo-
gena laghu sidhyeta sādhakaḥ ||.
71
rang ma sring mo BG] rang la sring ma T
72
thabs kyi B G ] thabs kyis T
73
cho ga yis T ] cho ga yi B; cho ga ni G

32
 Verse five

He who knows the yoga should always worship [his own]

mother, sisters, daughters and nieces, indeed, by the method
of wisdom and means.74

74
In pāda b, the Sanskrit has a metrical flaw, breaking the rule that either the second or
third syllable, or both, of each pāda should be guru (cf. Piṅgala’s Chandaḥsūtra 5.10: na
prathamāt snau (Śāstrī 1872: 102). A similar idea is found, for instance, in the PVS 5.25.

33
Verse six

ekāṅgavikalāṃ75 hīnāṃ
śilpinīṃ76 śvapacīṃ77 tathā |
yoṣitāṃ pūjayen nityaṃ
jñānavajraprabhāvanaiḥ ||78

tatremāni bījapadāni79 bhavanti | oṁ āḥ hūṁ 80 |

yan lag gcig81 ni ma tshang dman ||

bzo mo de bzhin sme sha can ||
ye shes rdo rje de bsgom pas82 ||
bud med rtag tu mchod par bya ||

de la ’di yi sa bon ni oṁ āḥ hūṁ ngo83 |

75
°vikalāṃ K1p.c. Ba Sa ] °vikalā° B K1a.c. K 2 Ba MS
76
śilpinīṃ K1p.c. Ba Sa ] śilpinī° B K1a.c. K 2 Ba MS; śilpināṃ Saga
77
śvapacīṃ Sa(em.) ] °svayacittakān B K1a.c. K 2; śvapacin K1p.c; śvapacikāṃ Ba(em.);
°śvapacika° Ba MS; svacikāṃ Sa ka Saga Sa mu
78
Cp. JS 1.80: sarvāṅgakutsitāyāṃ vā na kuryād avamānanam | striyaṃ sarvakulotpan-
nāṃ pūjayed vajradhāriṇam ||; cp. also Āryadeva’s CVP 106 and SS (p. 39): ekāṅgavik-
alāṃ hīnāṃ garhitām antyajām api | yoṣitaṃ pūjayen nityaṃ jñānavajraprabhāvanaiḥ ||.
The term jñānavajra is moreover found in chapter 15 of the JS where the famous mantra
containing this term is cited: oṁ sarvatathāgatajñānavajrasvabhāvātmako ’ham ||.
79
tatremāni bījapadāni B K1 K 2 Ba MS Sa ] tattvasyemāni mantrabījapadāni Ba(em.) Sa mu
80
hūṁ B K 2 Sa ] hūṃ K1 Ba Sa mu; hūḥ Sa ka
81
gcig BG] cig T
82
de bsgom pas BG] mthu’ yis ni T
83
de la ’di yi sa bon ni oṁ āḥ hūṁ ngo B G ] om. T (cp. SaTib: de kho na nyid kyi sngags
kyi ’bras bu ni ’di yin no)

34
 Verse six

He should always worship women by the cultivation of the

wisdom-vajra, women with low social status, those missing a
limb, artisans and, likewise, outcastes.84

Regarding it, the following are the seed-syllables:

oṁ āḥ hūṁ

84
śvapacikā, in pāda b, literally means “a female who cooks [and eats] dogs”. Here
it might be taken as synonymous for anybody with a very low social status or entirely
outside of public and social life. °prabhāvanaiḥ, in pāda d, is rendered by T as mthu’ yis
ni, which would rather correspond to °prabhāvaiḥ (“by the power of …”). B and G read
de bsgom pas, and the canonical sources translated bsgom pa che. The latter two, though
B and G do not render the upasarga pra°, tend to support (pra)bhāvana. The use of de,
however, remains slightly conspicuous and one could consider emending the reading to
rab bsgom pas or to adopt the canonical wording in which the che might be understood
as a rendering of pra°.

35
Verse seven

yena yena hi badhyante85

jantavo raudrakarmaṇā |
sopāyena86 tu tenaiva
mucyante bhavabandhanāt ||87

skye bo rnams ni mi bzad las ||

gang dang gang [B 272] gis ’ching ’gyur ba ||
thabs dang bcas pa de nyid kyis88 ||
srid pa’i ’ching ba las grol ’gyur |||

85
badhyante 𝝨 ] baddhyante K1p.c. Sa
86
sopāyena K1 Ba Sa(sil.) ] sopāyana B K 2 Ba MS
87
This famous verse is also found (literal) in GS 6.86cd–87ab, Advayavivaraṇapr-
ajñopāyaviniścayasiddhi (ed. Rinpoche and Dwivedi 1987: 217), Yoginīsaṃcāratant-
ra 17.20, HT II.ii.50, Sampuṭatantra 6.4.35, Amṛtakaṇikā (ed. Lal 1994: 68), Citta-
viśuddhiprakaraṇa 6 and SS (ed. Bendall 1905: 38) wherein the stanza is attributed to
Āryadeva’s Cittaviśuddhiprakaraṇa. Cp. also JS 1.15 and HT I.ix.19: yena tu yena tu
badhyate lokas tena tu tena tu bandhanaṃ muñcet | loko muhyati vetti na tattvan tattva-
vivarjataḥ siddhiṃ na lapsyet ||. Et al.
88
de nyid kyis B G ] de nyid kyi T

36
 Verse seven

By whichever terrifying karman beings are bound,

by precisely that, together with the means, however,
they are liberated from the chain of becoming.

37
Verse eight

samayān aharahaḥ kuryād

divyān pañcakulodbhavān |
pūjayec89 ca pradīpādyaiḥ
sakṣīrair viśvasaṃbhavaiḥ90 ||91

dam pa’i rigs lnga las ’byung bas92 ||

dam tshig nyin dang mtshan du bya ||
mar me la sogs ’o mar93 bcas ||
sna tshogs las ’byung94 rtag tu mchod ||8||

89
pūjayec K1 Ba Sa(sil.) ] pūjayac B K 2 Ba MS
90
viśva° 𝝨 ] bilva° B
91
Cp. GS 6.67cd–68ab (cd: literal), 68ab: anyaiś cāpi viśeṣādyaiḥ samayaiḥ pūjayed
dṛḍham. Cp. also JS 1.13: naraśvahayagodīpaṃ kariṇo gardabhasya ca | bhakṣayet tatt-
vasiddhyarthaṃ sarvasaṅkalpavarjitaḥ || and 1.79: mahāpradīpasaṃyuktaiḥ samayānyair
jugupsitaiḥ | pañcāmṛtaṃ sadā bhakṣyaṃ pūjyāś caiva tathāgatāḥ ||.
92
’byung bas em.] ’byung ba’i BG; ’byung ba yis T (hypermetrical)
93
’o mar B G ] ’dod mar T
94
las ’byung em. ] las byung 𝝨

38
 Verse eight

Day after day, he should make the pledges arising from the
five divine families. And, he should perform the worship with
the ‘lamps’ and so on arising from various things, together
with milk.95

95
Pāda a, in the Sanskrit, is hypermetrical, i.e. it has one metrical instance too many.
Further, the Tibetan reads “day and night” instead of “day by day”. Regarding the word
divyān in pāda b, we have followed the interpretation of the Tibetan translators who
have taken it to qualify the following pañcakula° in the sense of a sāpekṣasamāsa. It is
also possible to take it as an adjective qualifying the initial samayān. In pāda c, ‘lamps’
is short for the so-called “five lamps” (pañcapradīpa) which refer to the “five kinds of
meat,” i.e. the flesh of a human, dog, horse, cow and elephant. Indrabhūti, so it seems,
offers an alternative for the meat of an elephant, namely that of a donkey or ass (cp. JS
1.13). The °ādya in pradīpādyaiḥ can be taken as referring to the five kinds of nectar
(pañcāmṛta) and/or other pledge substances. See also JS 1.79 where the sets of five sub-
stances of flesh and nectar are also mentioned. The canonical translation renders this
part as gsal ba’i tshogs kyi(s). It is possible that the translators might have read pra-
dīpāḍhyair in their exemplar. Note also that MS B reads bilva (a plant (Aegle Marmelos)
traditionally used in Indian ritual, medicine and cuisine) instead of viśva. This reading,
however, is not supported by the Tibetan nor by any other sources known to us, and it
might well be just a misreading of the ligature śva which, orthographically, shows some
similarity to the ligature lva. Pāda d of the Sanskrit, moreover, reads exactly like pāda
17b, where it, perhaps owing to a different understanding of the context by the Tibetan
translators, has been translated differently.

39
Verse nine

[Ba p.17, Sa p.162]

protphullanayano mantrī
nityaṃ prahasitānanaḥ |
cittam āropya saṃbodhau
bhāvayej96 jñānasāgaram ||97

rtag tu mig ni gdangs ba dang ||

rtag tu rab tu dgod pa’i98 bzhin ||
sems ni yang dag99 byang chub la ||
bzhag ste100 ye shes rgya mtsho bsgom ||

96
bhāvayej K1p.c. Ba Sa(sil.) ] bhāvaye B K1a.c. K 2 Ba MS
97
Cp. GS 6.68cd–69ab (cd: literal), reading protphullanayano bhutvā nityaṃ ca suga-
tāśayaḥ in the first half.
98
dgod pa’i B G ] rgod ba T
99
yang dag B G ] dge pa T
100
bzhag ste B G ] bzhag la T

40
 Verse nine

The mantrin, with eyes wide open and always laughing, hav-
ing fixed the mind on complete awakening, should cultivate
the ocean of wisdom (jñānasāgara).

41
Verse ten

yāvat101 sthiracalā bhāvāḥ102

santy atra tribhavālaye |
sarve103 te tattvayogena
draṣṭavyā vajradhṛg yathā104 ||105

ji srid brtan dang [T 187] g.yo ba’i dngos106 ||

srid gsum gnas ’dir yod pa rnams ||
de nyid sbyor bas de dag kun ||
rdo rje ’dzin pa ji bzhin blta107 ||

101
yāvat K1p.c. Sa ] yāvaṃtaḥ Ba; yāvanta B K1a.c. K 2 Ba MS; yāvantaḥ Sa ka Sa mu
102
bhāvāḥ K1p.c. ] bhāvā B K1a.c. K 2 Ba MS Sa; °bhāvāḥ Ba
103
sarve K1p.c. Ba Sa ] sarv(v)a B K1a.c. K 2 Ba MS
104
yathā B K1 K 2 Sa ] tathā Ba(sil.) Sa mu
105
This verse is quoted in SS (p. 41): † yāvantaḥ sthiracalā bhāvāḥ † santy atra tribha-
vālaye | sarve te tattvayogena draṣṭavyā vajradhṛg yathā ||. Cp. GS 6.75ab: tat sarvaṃ
tattvayogena draṣṭavyaṃ yuktiveditam |.
106
dngos B G ] ngos T
107
ji bzhin blta B G ] ci bzhin ldang T

42
 Verse ten

As long as things, stable and moving, remain here in the

abode of the three existences, all these should be seen through
tattvayoga, just like the Vajra-holder.

43
Verse eleven

paravādinaś ca ye kecil
liṅgabhedair108 vyavasthitāḥ |
te ’py atra nāvamantavyā109
vajrasattvavikurvitaiḥ110 ||111

pha rol smra ba su yang rung ||

mtshan ma’i dbye bas gnas pa rnams ||
rdo rje sems dpa’i rnam ’phrul yis112 ||
de la113 ’dir ni brnyas mi bya114 ||

108
°bhedair K1 Ba Sa(sil.) ] °bheder B Ba MS; °bhaveda° K 2a.c.; °bha° K 2p.c.
109
°mantavyā Ba Sa(sil.) ] °gantavyā B K1 K 2 Ba MS
110
°vikurvitaiḥ Sa(em.) ] °vikurvvate B K1a.c. K 2 Ba MS Sa ka; °vikurvvataḥ K1p.c.; °vikurvite
Ba Sa kha
111
This verse too is cited in the SS (p. 41) as well as in Candrakīrti’s Pradīpoddyotana
(ed. Cakravarti 1984: 206), both reading vajrasattvavikurvitam in pāda d.
112
rnam ’phrul yis em.] rnam ’phrul yin BT; rnal ’phrul yin G
113
de la B G ] de las T
114
brnyas mi bya B G ] smra mi bya T

44
 Verse eleven

Moreover, any who adhere to other doctrines are

distinguished [from us just] by differences in defining marks.
They too should not be disrespected here by those who have
assumed [the state of] Vajrasattva.115

115
The Sanskrit in pāda a is hypermetrical. Note further that other interpretations of
liṅgabheda, in pāda b, are indeed possible. The author might have used the expression
deliberately as a pun on the Śaivas. In general, the ‘distinguishing marks’ should be
understood here as referring to any kind of signs and/or characteristics by which a practi-
tioner/follower of a particular system or doctrine is to be recognised. This can include
objects of worship, clothes, rites, etc. and, by extension, also doctrinal and philosophical
viewpoints. Note further that the Tibetan, in pāda c, reads rnam ’phrul yin (before emen-
dation), which corresponds to °vikurvitāḥ, and thus refers to paravādinaḥ. This nominal
sentence would translate as: “Any who adhere to other doctrines [and] remain with dif-
ferences in defining marks, having assumed the state of Vajrasattva, they should not be
disrespected here.” We have decided to follow the emendation in Sa, since this reading
can more easily explain the variants in the surviving MSS, and since it is logically sound.
An emendation to °vikurvitāḥ, following the Tibetan, would mean (as evident in the pre-
vious translation) that the paravādins have already assumed the state of Vajrasattva, by
which they would not be paravādins any longer. Note that the attestations of this stanza
in the SS and in Candrakīrti’s Pradīpoddyotana read °vikurvitam. As for the expression
vikurvita, cp. JS 1.17, 8.9, GS 2.18–19, SP 8, 53 et al.

45
Verse twelve

sarvān samarasīkṛtya
bhāvān nairātmyaniḥsṛtān |
bhāvayet satataṃ mantrī
dehaṃ116 prakṛtinirmalam ||

dngos rnams bdag med la gnas [G 232] te ||

thams cad ro gcig117 nyid du bya ||
lus ni rang bzhin dri med par ||
sngags pas rtag tu bsgom par bya ||

116
dehaṃ Ba Sa ] deha° B K1 K 2 Ba MS Sa ka Sa kha Saga Sagha
117
ro gcig B G ] ro cig T

46
 Verse twelve

Having made everything into the same flavour, [all] things

issued from selflessness, the mantrin should constantly culti-
vate [his own] body as stainless by nature.118

118
The Tibetan switches pādas and and b. In pāda a, the Tibetan sounds as if the tran-
slators read a form sthā (Tib. gnas) rather than niḥsṛ. In pāda b, the Tibetan reads ro gcig
(Skt. ekarasī°) for samarasī° which would rather correspond to ro mnyam. However,
these two can be treated as synonyms.

47
Verse thirteen

gandhamālyādibhir119 vastrair120
dhūpanaivedyakais121 tathā |
gītavādyais tathā nṛtyaiḥ122
sopāyair123 bhajate vibhuḥ ||124

dri dang phreng sogs gos rnams dang ||

spos dang lha bshos la sogs dang ||
glu dang rol mo gar rnams kyis125 ||
thabs dang bcas pas gtso bo mchod126 ||

119
°bhir K1p.c. Ba(sil.) Sa(sil.) ] °bhiḥ B K1a.c. K 2
120
°vastrair K1p.c. Ba(sil.) Sa(sil.) ] °vastrai B K1a.c. K 2
121
°kais K1p.c. Ba Sa(sil.) ] °kes K1a.c. K 2 B; °ke Ba MS
122
nṛtyaiḥ K1p.c. Sa ] nṛtya° B K1a.c. K 2 Ba Sa mu; nṛtyaṃ Sa ka
123
sopāyair K1p.c. Ba(sil.) Sa(sil.) ] sopāyaiḥ B K1a.c. K 2
124
Cp. JS 1.70: sugandhigandhalepādyair api liptaṃ yathāvidhi | ātmānaṃ pūjayed yogī
sarvabuddham anusmaran ||.
125
rnams kyis B G ] rnams kyi T
126
bcas pas gtso bo mchod B G ] bcas par gtso bor bzhugs T

48
 Verse thirteen

With perfumes, garlands and the like, with garments, incense

and food for gods, with songs, music and likewise dance — to-
gether with [all these] means — the sovereign treats himself.127

127
Note that the Tibetan BG reads gtso bo mchod, which sounds as if the translators
read vibhum rather than vibhuḥ. Considering both the context and the parallel in JS 1.70,
the reading vibhuḥ is preferable. For unknown T reads thabs dang bcas pas gtso bor
bzhugs in the final pāda.

49
Verse fourteen

na kaṣṭakalpanāṃ kuryān
nopa[B 36a]vāsaṃ128 na ca kriyām |
snānaṃ śaucaṃ129 na caivātra
grāmyadharmavivarjanam130 ||131

dka’ ba’i rtog pa mi bya zhing ||

smyung ba ma yin bya ba min132 ||
’dir ni gtsang sbra khrus mi dgos ||
grong ba’i chos ni rnam spang min ||

128
°vāsaṃ Ba Sa ] °vāsa B K1a.c. Ba MS; °vāse(°) K1p.c. Sagha
129
snānaṃ śaucaṃ Ba Sa ] śnānaṃ saucaṃ B K1a.c. K 2; snānaṃ saucaṃ K1p.c.
130
grāmya° K1p.c. Sa(em.) ] grāma° B K1a.c. K 2 Ba
131
Cp. CVP 58: na kaṣṭakalpanāṃ kuryān nopavāsena ca kriyām | snānaṃ śaucaṃ na
caivātra grāmadharmaṃ vivarjayet ||; cp. also HT II.iii.41: khānaṃ pāṇaṃ yathāprāptaṃ
gamyāgamyaṃ na varjayet | snānaṃ śaucaṃ na kurvīta grāmyadharmaṃ na varjayet ||;
JS 16.9: nopavāsarato yogī na snānaśuddhikalpanaḥ | nānākalpaviviktātmā bhaved yogī
varātmakaḥ ||. Similar ideas, i.e. those expressing “distance from ritual and religious
worship”, are found in various works related to the svādhiṣṭhānakrama and are also
expressed within the Dohās.
132
ma yin bya ba min em. ] yang ni mi bya’o B G; mas yin bya ba min T

50
 Verse fourteen

He should neither perform painful practice (kaṣṭakalpanā),

nor fasting (upavāsa) or ritual activities (kriyā), nor [should
he perform] ritual bathing [or] cleansing. And in this con-
text, indeed, [he should not] abandon the villagers’ duties
(grāmyadharma).133

133
In pāda b, the reading nopavāsena (ed. Patel 1949) finds support in MS K1p.c.. No-
netheless, we have rejected this variant in line with the previous editions of Śāstrī (1898)
and Yamada (1936). To our knowledge, all surviving Sanskrit MSS of the CVP support
the reading chosen here. The majority of the Tibetan attestations of this verse, both in
the AS and in the CVP, on the other hand support only one negation in this line and thus
tendentiously seem to be closer to the reading given in Patel. Only the Tibetan transla-
tion of the CVP as transmitted in G (the same collection also used here for the Tibetan
edition of the AS) has a double negation, rendering this line as smyung ba med cing bya
ba med (Vol. 36: 372). Consequently, we have decided to emend the reading to ma yin
bya ba min, basically following the Tibetan version T. Regarding the interpretation of
kaṣṭakalpanā, we are following Patel (1949: 93). In all likelihood, the term grāmyad-
harma is here to be understood as a euphemism for sexual intercourse. Regarding it,
see Amarakośa II.8.57 wherein the term is defined as follows: vyavāyo grāmyadharmo
maithunaṃ nidhuvanaṃ ratam ||.

51
Verse fifteen

na cāpi vandayed devān

kāṣṭhapāṣāṇamṛṇmayān134 |
pūjām asyaiva kāyasya
kuryān135 nityaṃ136 samāhitaḥ137 ||138

shing [B 273] dang rdo ba sa’i rang bzhin139 ||

lha rnams la yang phyag mi bya ||
rtag tu mnyam par gzhag pa yis ||
lus ’di nyid la140 mchod par bya ||

134
°pāṣāṇamṛṇ° K1p.c. Ba(sil.) Sa(sil.) ] °pāṣānamṛn° B K 2; °pāṣānmṛn° K1a.c.
135
kury(y)ān 𝝨 ] kūryyāt K 2
136
nityaṃ B K1 K 2 Sa ] nitya° Ba(sil.) Sa mu
137
samāhitaḥ B K 2 Sa ] samāhitamḥ K1; °samāhitaḥ Ba; samāhitaṃ Sa kha Saga
138
The first half of this verse is cited in SS (p. 41); the second part is missing because of
a lacuna in the surviving witnesses of the SS (lacuna also in NGMPP A 1057/20 fol. 13v8
and Takaoka CH319 fol. 20r1). The surviving part reads as printed here. Cf. GS 6.41cd –
42ab (abc: literal) in which pāda d reads kuryād vai vandanaṃ guroḥ |; cp. HT I.iii.44ab:
na vandayed imān devān kāṣṭhapāṣāṇamṛṇmayān | satataṃ devatāmūrtyā sthātavyaṃ
yoginā yataḥ ||, and SP 49: devān na vandayed evaṃ bhikṣūṃś cāpi na vandayet | athavā
vandayet sarvān svādhiṣṭhānakrameṇa tu ||.
139
shing dang rdo ba sa’i rang bzhin BG ] shing rdo sa’i rang bzhin gyis T (hypometrical)
140
’di nyid la em. ] ’di la ni B G; ’di nyid ma T

52
 Verse fifteen

And, moreover, he should not venerate gods made of wood,

stone or clay. Remaining in meditative equipoise (samāhita)
he should always perform the worship of his body alone.141

The GS reads: “... he should indeed perform the veneration of the teacher.” HT reads:
141

“He shall not venerate these gods consisting of wood, stone or clay, since constant abi-
ding with the form of the deity should be done by the yogin.” The term samāhita here,
and in its other occurrence in verse 21, has been translated with “meditative equipoise”.
This term, however, does have various implications and should be understood in the sen-
se of an undistracted, i.e. effortlessly concentrated, or focused, mind-state that is brought
about through meditation and that similarly may denote a final phase meditative state in
which the practitioner remains.

53
Verse sixteen

makṣikāchardisaṃmiśrair142
viṇmūtrādyaiś143 ca bhāvitaiḥ |
pañcapradīpasaṃyuktaiḥ
pūjayed vajradhāriṇam ||144

sbrang ma skyugs pa ’dres pa145 dang ||

dri chu legs par bsgos pa dang146 ||
sgron ma lnga dang bcas pa yis147 ||
rdo rje ’dzin par mchod par bya ||

142
makṣikā° 𝝨 ] macchikṣikā° K 2a.c. • °saṃmiśrair K1 Ba(sil.) Sa(sil.) ] °sanṃiśrai B K 2
143
viṇmūtrā° Sa ] viṭmūtrā° B K1 K 2; viṣamūtrā° Ba(sil.) Sa mu
144
Cp. GS 6.27 (ab: literal), pādas cd read: °saṃmiśram [read °saṃmiśrair] ātmānaṃ
pūjayet sadā |; JS 1.84: makṣikāchardisaṃmiśraiḥ purīṣādivimiśritaiḥ | pūjayed vajrasat-
tvātmā nirvikalpena cetasā ||.
145
skyugs pa ’dres pa B G ] skyug pa bsres pa T
146
bsgos pa dang B G ] bsres pa yis T
147
sgron ma lnga dang bcas pa yis B G ] om. T

54
 Verse sixteen

He should worship the Vajra-holder with feces and urine and

so forth (i.e. the five nectars) that have been prepared with
a mixture of honey (lit. ‘bee-vomit’), and furnished with the
‘ five lamps’.148

148
Note that the Tibetan, in pāda b, does not render the °ādi, but instead seems to have
read a form such as saṃbhāvitaiḥ. In pāda d, the GS (6.27) reads “He should always
worship himself with ...”.

55
Verse seventeen

abalāsvayaṃbhukusumaiḥ149
sakṣīrair150 viśvasaṃbhavaiḥ |
pūjayed devatās151 tena152
dehasthās153 tattvabhāvanaiḥ154 ||155

stobs med me tog rang byung ba156 ||

byang chub sems ni157 ’o mar bcas ||
de nyid bsgom pas lus gnas pa’i ||
lha rnams rtag tu mchod par bya ||

149
abalā° 𝝨 ] abalā Ba • °svayaṃbhu° Ba(sil.) Sa(sil.) ] °svayas tu B K1 K 2
150
sakṣīrair Ba (°ḥ) Sa(sil.) ] sakṣīra° B K1 K 2 Ba MS
151
devatās em. (cf. SS and Tib.) ] devatāṃ 𝝨
152
tena K1 K 2 Ba(em.) Sa ] tena na B Ba MS (dittography)
153
dehasthās em. (cf. SS and Tib.) ] dehasthān B K1 K 2; dehasthāṃ Ba Sa
154
tattva° 𝝨 ] °ta° K1a.c.
155
This verse too is cited in SS (p. 41). Because of a lacuna in the surviving MSS, the
majority of the first pāda is missing: – – – – – – – saṃ bodhicittasamanvitaṃ [|] pūjayed
devatās tena dehasthās tattvabhāvanaiḥ ||; cp. JS 1.81: ajātapuṣpasaṃyuktāṃ bodhicitta-
samanvitām | bhakṣayet tattvasiddhyarthaṃ nirvikalpavidhisthitaḥ ||; cp. also GS 8.46:
kausumāmṛtam āsādya bodhicittasamanvitam | khadhātudivyatejaskaṃ pūjayet tena
vajriṇam ||. Regarding the expression tattvabhāvanā, it may be noted here that chapter
four of the PVS is labeled “tattvabhāvanā” (PVS p. 80: iti prajñopāyaviniścayasiddhau
tattvabhāvanā nāma caturthaḥ paricchedaḥ).
156
rang byung ba B G ] rab ’byung ba T
157
byang chub sems ni T ] byang chub sems dpa’ BG

56
 Verse seventeen

He should [always] worship the deities abiding in [his] body

by the cultivation(s) of reality through this [procedure of wor-
ship], with the self-arisen flower of women (menstrual blood)
together with milk (semen) from which there is the arising of
everything.158

158
Pāda a of the Sanskrit is hypermetrical. In pāda b, instead of rendering sakṣīrair
viśvasaṃbhavaiḥ, the Tibetan attests byang chub sems ni/dpa’ o mar bcas’, i.e. “bodhi-
citta together with milk” (with ni, the reading of T) or “the milk [of] a Bodhisattva” (with
dpa’, the reading of BG). A reading with bodhicitta finds support in the surviving parts
of the quote in the SS. There, however, sakṣīra is not attested. On account of the textual
witnesses, it is difficult to judge which reading was the sample for the Tibetan translation
and which might have been the original reading. Besides the interpretation we have cho-
sen above, namely to understand kṣīra as bodhicitta and to take viśvasaṃbhavaiḥ as an
adjective qualifying the other instrumentals, there are at least a couple of other options.
It is not impossible to interpret the compound viśvasaṃbhavaiḥ as an upacāra in the
sense of bodhicittād viśvasyotpattiḥ, to be understood here as semen (śukra, i.e. “relative
bodhicitta”), the counterpart of the “self-arisen flower of women”, i.e. menstrual blood
(rakta or rajas). The other option would be to take viśva as a synonym of dharmodaya/ā,
the “source of the entire universe”. In both of these alternatives kṣīra, would be understo-
od in its literal sense as “milk”. In pāda c, the Tibetan does not render tena, but instead
reads rtag tu in pāda d, which would suggest either satatam, sadā or nityam (“always”),
of which only the last is metrically correct. Nonetheless, it is rather difficult to explain
how an easily comprehensible reading would become corrupted to a much more difficult
one. In addition to this, the quote in the SS also supports tena. We have thus decided to
keep the reading as it is preserved in the surviving Sanskrit witnesses, and interpreted it
in the sense of tena prakāreṇa. It is also possible that the rtag tu got into the text partly
because of a memory of the somewhat similar line in verse 4cd.

57
Verse eighteen

parasvaharaṇaṃ159 kuryāt
paradārā[K2 21r]niṣevaṇam160 |
vaktavyaṃ ca ṃṛṣāvākyaṃ
sarvabuddhāṃś161 ca ghātayet ||162

gzhan gyi nor ni brku bya zhing163 ||

gzhan gyi bud med bsten par bya ||
rdzun ni rtag tu smra bya zhing164 ||
sangs rgyas thams cad gsad par bya ||

159
°haraṇaṃ 𝝨 ] °raṇaṃ K1a.c.
160
°niṣevaṇam Ba Sa(sil.) ] °niṣevanaḥ B K 2 Ba MS; °niṣevanam K1
161
°buddhāṃś K1 Ba Sa(sil.) ] °buddhāś B K 2; °buddhaś Ba MS
162
This verse too is quoted in SS (p. 41), reading parasvaharaṇaṃ kāryaṃ paradārāni-
ṣevaṇam | vaktavyaṃ cānṛtaṃ nityaṃ sarvabuddhāṃs ca ghātayet ||; cp. JS 1.14: ghātayet
tribhavodbhūtān paravittāni hārayet | kāmayet paradārān vai mṛṣāvādam udīrayet || and
GS 1.15. Similar statements can, for instance, also be found in the Guhyasamājatantra
(16.61–62: prāṇinaś ca tvayā ghātyā vaktavyaṃ ca mṛṣāvacaḥ | adattaṃ ca tvayā grāhyaṃ
sevanaṃ yoṣitām api || anena vajramārgeṇa vajrasattvān pracodayet | eṣo hi sarvabud-
dhānāṃ samayaḥ paramaśāśvataḥ ||) and Hevajratantra (HT II.iii.29: prāṇinaś ca tvayā
ghātyā vaktavyaṃ ca mṛṣāvacaḥ | adattaṃ ca tvayā grāhyaṃ sevanaṃ parayoṣitaḥ ||).
163
brku bya zhing B G ] rku bya zhing T
164
smra bya zhing B G ] smra bar bya T

58
 Verse eighteen

He shall take others’ wealth and enjoy others’ women.

Also, wrong statements are to be told.
And he should kill all the buddhas.165

165
Note that the Tibetan rdzun ni rtag tu smra bya zhing in pāda c seems to confirm
the reading vaktavyaṃ cānṛtaṃ nityaṃ as it is preserved in the SS. It is possible that
this reading in fact is the original one. The context of this verse is that of the so-called
“Holy Madman” (unmattavrata, smyon pa’i brtul zhugs). It should not go unmentioned
that such statements are not meant to be taken literally, but rather quite the opposite,
and that these teachings (when indeed put into practice and displayed publicly as a spi-
ritual means) are restricted to very specific contexts and limited to a rather specific set
of practitioners (see Davidson 2002: 223–224). This seems to be implied as well in § 3
of the ’Bri gung chos mdzod (see Appendices: Additional passage in T (189,2–190,3)).
And this also becomes obvious when reading, for instance, the verse from the HT (II.
iii.30) which directly follows the verse given above as a parallel: ekacittaṃ prāṇivad-
haṃ [proktaṃ] prāṇa[ṃ]* cittaṃ yato mataṃ | lokān uttarāyiṣyāmi mṛṣāvādañ ca
śabditaṃ | yoṣicchukram adattañ ca paradārāḥ svābhasundarī || (*HT reads [proktaṃ]
prāṇa, HT MĀ reads prāṇa°). “Killing living beings is taught to be the single[-pointed]
mind, since prāṇa is considered the mind. And, ‘I will save the sentient beings’ is
called telling lies. What has not been given is the woman’s sexual fluids, and another’s
wife is [nothing but] the Sundarī (Nairātmyā) with the splendor of oneself.” Regarding
this passage, see also MĀ (p. 162f.). Another passage worth consulting in this regard is
found in Nāropā’s Sekoddeśaṭīkā which provides a rather elaborate explanation on this
topic (cf. Sferra 2006: 93ff.)

59
Verse nineteen

[Ba p.19]

śailamṛṇmayacaityādīn166
na kuryāt pustake167 ratim |
na maṇḍalāni [K1 15v] svapne ’pi168
kāyavākcittakarmaṇā169 ||170

rdo dang sa yi171 mchod rten sogs ||

glegs bam la sogs dga’ mi bya172 ||
rmi lam du yang dkyil ’khor sogs173 ||
lus ngag yid kyi las mi bya’o174 ||

166
°mṛṇ° Ba(sil.) Sa(sil.) ] °mṛn° B K1 K 2 • °caityā° K1p.c. Ba Sa(sil.) ] °caitrā° B K1a.c. K 2 Ba MS
167
pustake B K1p.c. Ba Sa(sil.) ] puṣṭake K1a.c. K 2
168
svapne (’)pi K1 K 2 Ba Sa(sil.) ] svapneti B Ba MS
169
°citta° 𝝨 ] °cirtta° K1a.c.
170
A closely related parallel is found in GS 6.42cd–43ab: śailamṛṇmayastūpādīn na
kuryāt pustake ratim | maṇḍalāni na svapne ’pi kuryān mudrāṃ na hastayoḥ ||. Cp.
also SP 48: caitya karma na kurvīta na ca pustakavācanam | karotu vācayec cāpi svād-
hiṣṭhānakrameṇa tu ||.
171
sa yi B G ] sa’i T (hypometrical)
172
dga’ mi bya T ] dgar mi bya B G
173
dkyil ’khor sogs B G ] dkyil ’khor dang T
174
lus ngag yid kyi las mi bya’o conj. ] lus kyi rnams ni mi bya’o B G T

60
 Verse nineteen

He should not — with actions of body, speech and mind —

build caityas consisting of stone and clay, take delight in
scripture, [and], not even in dreams, [create] maṇḍalas.175

175
The idea of “non-engagement and absence from ritual” is found in various sources.
See also a passage cited in the SS (p. 58): tathā ca mudrābandho na maṇḍalaṃ na cai-
tyaṃ na ca pustakavācanaṃ na kāyakleśaṃ na paṭakāṣṭhapāṣāṇapratimāṃ praṇamati
na Śrāvaka-Pratyekabuddhaṃ na tithinakṣatramuhūrtakālāpekṣaṇaṃ karoti |. Simi-
larly, this literary “theme” is addressed in various of the major Dohākośas, such as in
Kāṇha’s Dohākośa: 28–30; Tillopā’s Dohākośa 19–22; and essentially the first third of
Saraha’s Dohākośa. This is further described in a paragraph entitled “Cultural Critique”
in Jackson 2004. In this context, similar to the previous verse, such statements are not
necessarily to be taken literally. See, for instance, HT I.x.43: na mantrajāpo na tapo
na homo na māṇḍaleyaṃ na ca maṇḍalaṃ ca | sa mantrajāpaḥ sa tapaḥ sa homas tan
māṇḍaleyaṃ tan maṇḍalaṃ ca || (“No recitation of mantras, no austerities, no oblation,
and nothing related to a maṇḍala. That is the recitation of mantras, that is the austerities
and oblation, and that is the maṇḍala and what relates to it.”). GS 6.40cd–43 further
provides some interesting context for this as well as the preceding verse: tataḥ siṃhavad
vicared guhyasiddhipratiṣṭhitaḥ ||40|| unmattaveśadhṛg bhūtvā sarvākāravivarjitaḥ | na
cāpi vandayed devīṃ kāṣṭhapāṣāṇamṛṇmayām ||41|| pūjām asyaiva kāyasya kuryād vai
vandanaṃ guroḥ | śailamṛṇmayastūpādīn na kuryāt pustake ratim ||42|| maṇḍalāni na
svapne ’pi kuryān mudrāṃ na hastayoḥ | vandanaṃ naiva kurvīta triyānapathavartinām
||43|| (“Then, after having become one who has ‘mad’ behaviour, who is completely wi-
thout any ‘aspects’, firmly established in the ‘secret accomplishment’, he shall abide/
behave like a lion. He also shall not venerate a goddess made of wood, stone or clay. He
should perform worship of his body alone and the veneration of [his] teacher, indeed.
He should not build stūpas consisting of stone and clay [nor] take delight in books. He
should not create maṇḍalas, even in a dream, nor mudrās with both [his] hands. Not at
all he shall perform the veneration of those on the paths of the three yānas.”). In pādas
b and c, the Tibetan does not render the plurals, but supplies ādi (sogs) instead. In pāda
d, the Tibetan does not render °vākcitta°, but reads lus kyi rnams ni mi bya’o (“he should
perform any physical [activities]”) instead. It is noteworthy that the Tibetan translation in
this case is surprisingly inaccurate, perhaps being a mistake owed to a bad transmission.
Considering the rather straightforward content of this pāda, it would have been easy to
render it more accurately, such as conjectured above.

61
Verse twenty

[Sa p.163]

jugupsāṃ176 naiva kurvīta

sarvavastuṣu177 mantravit |
vajrasattvaḥ svayaṃ tatra178
sākṣādrūpeṇa saṃsthitaḥ179 ||180

dngos po kun la sngags shes pas ||

smad pa181 dag tu mi bya’o ||
rdo rje sems dpa’182 rang bzhin der183 ||
mngon sum gzugs kyis184 yang dag [G 233] bzhugs ||

176
jugupsāṃ K1 Sa(sil.) ] jugupsā 𝝨
177
°vastuṣu Ba Sa(sil.) ] °vastu B K1 K 2 Ba MS
178
vajrasattvaḥ svayaṃ tatra K1p.c. Ba Sa(sil.) ] vajrasatve svayaṃ tantra° Ba.c.; vajrasatva-
svayaṃ tantra° Bp.c. Ba MS; vajrasatvasvayaṃ tatra K1a.c. K 2
179
°saṃsthitaḥ K1p.c. Ba Sa(sil.) ] °saṃsthitāḥ B K1a.c. K 2 Ba MS
180
Cp. CVP 130cd (= Catuḥśatakā 127ef): svayaṃ vajradharo rājā sākṣādrūpeṇa saṃst-
hitaḥ ||.
181
smad pa B G ] smon pa T
182
sems dpa’ T ] sems dpa’i B G
183
der B G ] te T
184
gzugs kyis B G ] gzugs kyi T

62
 Verse twenty

Concerning all things, he who knows the mantra[naya] should

never develop disgust. Vajrasattva himself remains therein
with a directly perceptible nature (sākṣādrūpeṇa).185

185
Cp. JS and GS wherein the notion of not developing “disgust” (jugupsā) is particu-
larly emphasised in various places.

63
Verse twenty-one

gamyāgamyavikalpaṃ tu
bhakṣyābhakṣyaṃ tathaiva ca |
peyāpeyaṃ tathā mantrī
kuryān186 naiva samāhitaḥ ||187

bgrod dang bgrod min rnam rtog dang188 ||

de bzhin du ni bza’ mi bza’ ||
sngags pas189 de bzhin btung mi btung ||
mnyam par gzhag pas brtag mi bya190 ||

186
kuryān K1 Ba(sil.) Sa(sil.) ] kuryyā B K 2
187
Cp. JS 1.18: bhakṣyābhakṣyavinirmuktaḥ peyāpeyavivarjitaḥ | gamyāgamya-vim-
uktātmā bhaved yogī samāhitaḥ ||, HT I.vi.21: bhakṣyābhakṣyavicāraṃ tu peyāpeyaṃ
tathaiva ca | gamyāgamyaṃ tathā mantrī vikalpaṃ naiva kārayet || et al.
188
rnam rtog dang B G ] rnam rtog pa T
189
sngags pas B G ] bsngag pas T
190
brtag mi bya B G ] b...g mi bya T (T has a small lacuna where the ligature rta is
supposed to be.)

64
 Verse twenty-one

The mantrin, remaining in meditative equipoise, should

indeed not make concepts about whom to approach and not to
approach, what to eat or not eat, and likewise what to drink
and not to drink.191

191
The Tibetan freely renders the form kuryāt here as brtag mi bya. In this verse, which
finds close parallels in several others tantric works (e.g. JS 1.18 and HT I.iv.29 et al.),
the expression gamyāgamya is to be understood with a sexual connotation in the sense of
“whom to approach and not to approach for sexual intercourse.”

65
Verse twenty-two

vairocanasamudbhūtān192
sarvaprāṇyaṅgasaṃbhavān193 |
prāṇakān gṛhya194 tattvajño
bhakṣayet195 siddhihetunā ||196

rnam par snang mdzad197 las byung ba’i ||

yan lag las byung198 srog [B 274] chags rnams ||
srog chags de nyid shes pas blang199 ||
dngos grub [T 188] rgyur ni200 bza ’bar bya ||

192
°samudbhūtān Ba(sil.) Sa ] °samudbhūtāt B K1 K 2 Sagha
193
°saṃbhavān Ba Sa ] °sambhavāt B K1 K 2 Sagha; °saṃbhavat Ba MS
194
gṛhya 𝝨 ] guhya° Ba(sil.) Sa mu
195
bhakṣayet 𝝨 ] bhakṣaye K 2
196
Cp. JS 1.78: sarvasattvāṅgasaṃbhūtān prāṇino ’pi viḍudbhavān | saṃgṛhya ca
mahāyogī bhakṣayet tattvasiddhaye || and GS 6.71cd–6.72ab, reading siddhikāṅkṣayā
instead.
197
snang mdzad B G ] dag pa T
198
yan lag las byung B G ] srin bu la sogs T
199
srog chags de nyid shes pas blang B G ] de nyid shes pas blang byas la T
200
rgyur ni B G ] bsgrub phyir T

66
 Verse twenty-two

Having seized living beings (worms etc.) that have arisen from
vairocana (feces), [or] those arising from the limbs of any
living being, he who knows reality should consume [them] as
a cause of accomplishment.201

201
Note that the Tibetan versions B and G render the Sanskrit forms °samudbhūtān
and °saṃbhavān in pādas a and b both with byung ba as if the Sanskrit read °bhūtān
twice and without the upasargas. Moreover, the Tibetan translators, by using the geni-
tive byung ba’i, seem to have understood the compound in pāda a as an adjective to the
compound in pāda b. In pāda b, furthermore, the word order does not seem to reflect that
of the Sanskrit, and we find the Tibetan translation difficult to account for. One rather
would have expected the reversed word order srog chags rnams yan lag las byung (or bet-
ter: srog chags kun gyi yan lag ‘byung). The redacted version T differs significantly in
this half, reading rnam par dag pa instead of rnam par snang mdzad in pāda a and srin
bu la sogs instead of yan lag las byung in pāda b. Regrettably, we are not able to account
for these differences. The latter might have been a marginal note or interlinear gloss on
srog chags rnams which found its way into the text in the course of the redaction. The
form gṛhya, in pāda c of the Sanskrit, despite the fact that the grammatically correct
form is gṛhītvā, has been accepted here for metrical purposes. This form is also attested
in the GS. In the last pāda, the GS reads: “... consume them with the desire (kāṅkṣā) for
accomplishment”.

67
Verse twenty-three

sarvavarṇa[B 36v]samudbhūtā
jugupsyā202 naiva yoṣitaḥ203 |
saiva bhagavatī prajñā
saṃvṛtyā204 rūpam āśritā205 ||206

rigs kun las ni byung ba yi207 ||

bud med rnams ni smad mi bya ||
de nyid bcom ldan208 yum shes rab ||
kun rdzob gzugs la209 brten pa210 yin ||

202
jugupsyā Ba(nt.) Sa(em.) ] jugupsā B K1 K 2 Ba; jugupyā Saa.c.
203
yoṣitaḥ Ba(sil.) Sa(sil.) ] coṣitaḥ B K1 K 2
204
saṃvṛtyā B K1 K 2 Sa ] savṛtyā Ba (misprint?)
205
rūpam āśritā B K1 Ba Sa ] rūpeṇa māśritā K 2; rūpaṃ samāśritā Sa ka
206
Cp. JS 1.82: caṇḍālakulasambhūtāṃ ḍombikāṃ vā viśeṣataḥ | jugupsitakulot-pan-
nāṃ sevayan siddhim āpnuyāt ||.
207
byung ba yi B G ] byung ba yis T
208
bcom ldan B T ] bcom pa ldan G (hypermetrical)
209
gzugs la B G ] sku la T
210
brten pa yin T ] bsten pa yin B G

68
 Verse twenty-three

Women, born in any class, are never to be disrespected. She

alone is the bhagavatī, the wisdom consort (prajñā), who has
taken up a physical form [only] by convention.211

211
 In pāda b, the emendation jugupsyā in Sa has already been suggested in Ba, and it is
further supported here by the extra-canonical Tibetan translation smad mi bya. Further,
even though it is neither supported by the Sanskrit MSS, nor the canonical and extra-ca-
nonical translations, which both mark the plural, one could consider emending yoṣitaḥ
to yoṣitā, understanding it not as an instrumental singular of yoṣit, but as a nominative
singular of the less common ‘extended’ stem yoṣitā in order to avoid the somewhat di-
sturbing change of number in the two halves of the verse. Also, it is noteworthy that T
translates rūpam, in pāda d, here as sku (the honorific form of kāya “body”), which is
reserved only for buddha-forms or other deities and, although not as literal as gzugs, is
fully acceptable here in this context.

69
Verse twenty-four

na tithir212 na ca nakṣatraṃ
nopavāso vidhīyate |
advayajñānayuktasya
siddhir bhavati213 saugatī ||214

tshes grangs dang ni215 rgyu skar216 med217 ||

smyung ba218 dag kyang ma yin te ||
gnyis med219 ye shes dang ldan pas ||
bde bar gshegs pa ’grub par ’gyur ||

212
tithir K1 Ba Sa(sil.) ] tithi B K 2 Ba MS
213
bhavati 𝝨 ] bharvati K 2
214
Also found in the GS 1.66 (literal).
215
dang ni T ] gza’ dang B G
216
rgyu skar G T ] sgyu skar B (cp. v.2)
217
med B G ] min T
218
smyung ba B G ] bsnyung ba T
219
gnyis med B G ] gnyis de T

70
 Verse twenty-four

Neither a lunar day or constellation, nor fasting is prescribed.

For him who is yoked with non-dual gnosis, there is the ac-
complishment relating to a Sugata (buddhahood).220

220
In pāda b, the Tibetan sounds as if the translators had read nopavāso ’pi vidyate. The
ergative in pāda c has been accepted. It is also possible to emend the reading to dang
ldan pa’i, which would be a bit closer to the Sanskrit °yuktasya, yet less comprehensible
for a Tibetan reader.

71
Verse twenty-five

bahunātra kim uktena

yad bhaved 221 upalabdhikam |
tat sarvaṃ tattvayogena
draṣṭavyaṃ yuktivedinā222 ||223

’dir ni mang du smos ci dgos ||

gang zhig224 nye bar dmigs ’gyur ba ||
de kun de nyid225 sbyor ba yis ||
rigs pa’i226 shes pas227 blta bar bya ||

221
bhaved K1 Sa(sil.) ] bhavad B K 2 Ba; bhavaty Ba(nt.)
222
yuktivedinā em. (cp. Tib. and GS) ] ta(t)tvavedinā 𝝨
223
Cf. GS 6.74cd–75ab, reading yuktiveditam instead of tattvavedinā; cp. HeTa I.viii.37
which reads: hīnamadhyamotkṛṣṭāny evānyāni yāni tāni ca | sarvāṇy etāni samānīti
draṣṭavyaṃ tattvabhāvanaiḥ ||, and Sahajasiddhi 2.1 (ed. Rinpoche and Dwivedi 1987):
hīnamadhyamotkṛṣṭāni anyāni yāni tāni ca | sarvāṇi tāni samāni draṣṭavyāni tattvabhāv-
anaiḥ ||.
224
gang zhig B G ] gang gis T
225
de nyid T ] de yi B G
226
rigs pa’i em. ] rig pa’i B G T (cp. GS: yuktiveditam)
227
shes pas T ] shes pa B G

72
 Verse twenty-five

What is the use of talking a lot in this regard?! All that which
may be perceived is to be seen through tattvayoga by the one
who knows yukti.228

228
We have decided to emend the Sanskrit text against what is preserved in all MSS
(tattvavedinā), based on the reading found in the GS (yuktiveditam). This emendation
finds further support in the canonical and extra-canonical Tibetan translations, both
of which are reading rig pa, which we are taking here for rigs pa, i.e. yukti. The cor-
ruption of the Sanskrit text might have been caused by the preceding tattvayogena,
either due to an eye-skip, through a misplaced insertion of a previously dropped out
tattva, or by similar formulations in related texts. As for the term yukti, it is often used
in the sense of “reasoning” opposed to scripture (āgama). Here, however, it is possible
that the authors, i.e. Lakṣmīṅkarā and Padmavajra, were consciously playing with the
verbal root yuj (“to yoke”) from which both yukti as well as yoga are derived. In the
context of this verse, yukti thus might also be understood in the sense of “application,
practice, or means of union”.

73
Verse twenty-six

hastyaśvakharagāvoṣṭra-
pradīpaṃ229 śvānasaṃbhavam230 |
mahāpradīpasaṃmiśraṃ231
bhakṣayed yogavit sadā ||232

glang po rta bong glang rnga mo233 ||

khyi las byung ba’i sgron ma ni ||
sgron ma che234 dang bsres pa ni ||
rnal ’byor shes pas rtag tu bza’ ||

229
°aśva° K1p.c. Ba(sil.) Sa(sil.) ] °asva° B K1a.c. K 2
230
°am 𝝨 ] °an K 2
231
°pradīpa° 𝝨 ] °pradīpā° Ba.c. • °saṃmiśraṃ 𝝨 ] °saṃnmiśraṃ K1a.c.
232
Cp. GS 6.31: hastyaśvakharagāvoṣṭrapradīpaṃ manujodbhavam | bhakṣayet satataṃ
mantrī na cānnaṃ tu vibhakṣayet ||.
233
glang rnga mo B G ] ba rnga mong T
234
sgron ma che B G ] sgron ma phye T

74
 Verse twenty-six

The one who knows yoga should always eat the ‘lamps’[, i.e.
the flesh] of elephant, horse, donkey, cow and camel, [and/or]
that which arises from a dog, mixed together with the ‘great
lamp’ (human meat).235

235
Usually, one expects the five kinds of meat to be only those of man, horse, cow, dog
and elephant. Here, however, Lakṣmīṅkarā seems to offer further alternatives, sugge-
sting dog-meat mixed with human flesh as the ultima ratio, in as much as these ingre-
dients should be relatively easily available everywhere. Regarding this, see also the note
on verse eight, and compare GS 6.31 and JS 1.13, 78–79. Perhaps also worth mentioning
is the striking reminiscence of the list of animals in pāda a of the first five of the eight
animals that are held in the skull-bowls in the right hands of the 16-armed form of He-
vajra. Regarding this, cp. HT II.v.24 et al.

75
Verse twenty-seven

[Ba p.21]

na cādhyāsaktiṃ236 kurvīta
ekasminn api yogavit |
samatācittayogena
bhāvanīyo237 bhavārṇavaḥ238 ||

rnal ’byor shes pas239 gcig la240 yang ||

lhag par chags par mi bya zhing241 ||
mnyam pa’i sems kyi sbyor ba yis ||
srid pa’i rgya mtshor242 bsgom par bya243 ||

236
°saktiṃ K1p.c. Ba ] °śakti B Sa ka; °śaktiṃ K1a.c. K 2; °sakti Ba MS Sa
237
bhāvanīyo Ba Sa ] bhāvanīyā B K1 K 2 Ba MS Sa kha Saga
238
bhavārṇavaḥ B K 2 Ba Sa ] bhavārṇave K1 Sa kha Saga Sagha
239
rnal ’byor shes pas em.] rnal ’byor shes pa B G; rnal par shes pas T
240
gcig la em. (cp. ekasmin)] ci la B G; rtsi la T
241
In between pādas b and c, G reads mnyam pa bza’ || rnal ‘byor shes pa ci la yang ||
lhag par chags par mi bya zhing ||, evidently repeating pādas ab and the beginning of c,
possibly due to an eye-skip or another mistake.
242
srid pa’i rgya mtshor B G ] bde ba chen por T
243
bsgom par bya T ] goms par bya B G

76
 Verse twenty-seven

And he who knows yoga should not produce strong attach-

ment even towards a single one/thing. The ocean of existence
is to be contemplated on by the ‘yoga of the mind of equanim-
ity’.244

In pāda d of the Tibetan, the extra-canonical version T reads bde ba chen por bsgom
244

par bya (“he should cultivate great bliss …”) instead of srid pa’i rgya mtshor (B and G).
Even though a reading mahāsukhaḥ would be possible in the Sanskrit of pāda d, it cannot
be explained on account of any of the sources currently available to us and must therefore
be regarded as one of the various examples in which T — perhaps due to the redactions
that happened in the course of its transmission — offers surprising and unique readings,
the origins of which remain uncertain.

77
Verse twenty-eight

utpattisthitinirodhaṃ ca
āsaṃpṛktaṃ245 pṛthagjanaiḥ |
tasya bhāve246 tu saṃsāro
nānyatra pralayodbhavaḥ247 ||

skye dang gnas dang ’gag pa la ||

so so’i skye bo ’jug byed pa248 ||
de med na ni ’khor ba ru ||
gzhan dag tu ni ’jug mi ’gyur ||

245
ca āsaṃpṛktaṃ Ba(sil.) ] cāsanpṛttam B K1a.c. K 2; cāsaṃpṛktaṃ K1p.c. Sa(sil.)
246
bhāve 𝝨 ] bhāva K 2
247
pralayodbhavaḥ 𝝨 ] pralayīdbhavaḥ K 2
248
byed pa T ] byed na B G

78
 Verse twenty-eight

Arising, abiding and cessation are related to ordinary people.

In the existence of these, however, is saṃsāra. Elsewhere
there is no perishing and arising.249

249
The reading in the first pāda of the Sanskrit verse as it is printed here, is hyper-
metrical. We decided to follow the reading in Ba wherein the sandhi between ca and
āsaṃpṛktaṃ is not applied, in order to preserve the final cadence of an uneven pāda.
The application of the sandhi, as done in the MSS and Sa, solves the hypermetricism
yet leads to an unmetrical pāda which, from our point of view, is less preferable. As
regards the Tibetan, pāda c reads de med na ni ’khor ba ru, which rather sounds as if
the translators had read tasyābhāve tu saṃsāre. The canonical translation, on the other
hand, translates something like tasyābhāve na saṃsāro, which, in the context of this
verse, seems preferable. The extra-canonical version can be understood more or less in
line with the Sanskrit when pāda c is read as an individual sentence and so so’i skye bo is
again supplied as subject in the last pāda (“When saṃsāra is not existing, [ordinary pe-
ople] do not engage into anything else [any longer].”). One also could consider adopting
the canonical translation for the third pāda of the verse, reading de dag med na ‘khor ba
med (“When these are not existing, saṃsāra does not exist, [and people] do not engage
in anything else [any longer].”). With regard to the last pāda, it has to be noted that
both the extra-canonical and canonical versions differ significantly from the text as it is
transmitted in Sanskrit. The inaccurate rendition of the last Sanskrit pāda, despite the
fact that its translation should have been relatively easy (for instance: gzhan du ’ jig/’gag
dang ’byung ba med), could indicate that the Tibetan was based on a somewhat different
version of the Sanskrit text. The form ’ jug mi ’gyur in the extra-canonical version (in
combination with the Sanskrit prefix pra°) might indicate a form derived from the roots
vṛt or viś. The canonical version, on the other hand, translates skye ba dang ni ’byung ba
med, and is, on account of the Sanskrit witnesses currently available, even more challen-
ging to explain.

79
Verse twenty-nine

dinas tu250 bhagavān vajrī

naktaṃ251 prajñā vidhīyate252 |
evaṃ tu bhāvayed yogī
laghu253 siddhim avāpnuyāt ||254

nyin mo bcom ldan rdo rje can ||

mtshan mo shes rab ces su bshad255 [G 234] ||
de ltar sgom pa’i256 rnal ’byor pas257 ||
dngos grub myur du ’thob par ’gyur 258 ||

250
dinas tu em. ] dinam tu B; dinan tu K1 K 2; dinaṃ tu Ba Sa
251
naktaṃ B K1p.c. Ba Sa(sil.) ] nakṣam K1a.c. K 2
252
vidhīyate B K1 K 2 Ba Saa.c. ] ’bhidhīyate Sa(em.)
253
laghu corr. ] laghu° 𝝨
254
Cf. HT I.viii.23 and Ṣaṭsāhasrikā Hevajratantraṭīkā 4.2 (ed. Shendge 2004: 23).
255
ces su bshad B G ] zhes su brjod T
256
sgom pa’i B G ] bsgom pa’i T
257
rnal ’byor pas B G ] rnal ’byor pa T
258
’thob par ’gyur B G ] thob par ’gyur T

80
 Verse twenty-nine

Day is the lord, the vajrin; night, on the other hand, is taught
to be prajñā (the consort). And in this way (evaṃ tu) the yogin
should meditate. He shall obtain accomplishment easily. 259

259
There are various ways of expressing the pair of male and female, such as “wisdom
and means,” (Skt. prajñopāya) or “sun and moon” and the like. Such a symbolism of a
pair which expresses two equal parts of a harmonious principal, or unity, is omnipresent
in tantric Buddhist literature and strongly interwoven with various practices drawing
upon the sexual connotation of these pairs, such as śukra and rakta in verse 17. Cf. HT
I.vii.23: tatraivaṃ mantavyam | dinas tu bhagavān vajrī naktaṃ prajñā ca bhaṇyate || and
Ṣaṭsāhasrikāhevajratantraṭīkā 4.2 (ed. Shendge 2004: 23): dinaṃ tu bhagavān vajrī na-
ktaṃ prajñā prakīrtitā | yathādityaḥ śivaḥ proktaḥ candraḥ śaktiḥ tathā smṛtā ||. The final
pāda of this stanza is found in many sources, such as HT I.viii.22b, II.v.61b, GST 15.14d,
Saṃpuṭatantra ch. 4 et al. In pāda c of the Tibetan, it is possible that the translators had
read bhāvayan, while it is likewise possible that the genitive was used to avoid two finite
verbs.

81
Verse thirty

yat tad avyaktarūpaṃ tu

sarvasattveṣu saṃsthitam |
guruvaktrāt paraṃ tattvaṃ
prāpyate nātra saṃśayaḥ ||260

gang de mi gsal gzugs su ni ||

sems can [B 275] rnams la rnam gnas pa261 ||
bla ma’i zhal nas262 de nyid mchog ||
rnyed ’gyur263 ’di la the tshom med ||

260
With reference to the Mahālakṣmīsādhana, this verse is quoted in SS (pp. 8–9):
Mahālakṣmī-Sādhane ’py uktaṃ | yat tad avyaktarūpaṃ tu sarvadehe vyavasthitaṃ |
guruvaktrāt paraṃ tattvaṃ prāpyate nātra saṃśayaḥ ||.
261
gnas pa T ] gnas pas B G
262
zhal nas B G ] kha nas T
263
rnyed ’gyur B G ] brnyed ’gyur T

82
 Verse thirty

That unmanifested form, however, which is firmly located

within all sentient beings, is the supreme reality obtained from
the mouth of the guru — there is no doubt about it.264

This verse is cited in the SS (pp. 8–9), reading sarvadehe vyavasthitaṃ in pāda b. On
264

the one hand, the Tibetan translation supports °sattveṣu (sems can rnams la) as attested
in the MSS, yet without rendering the sarva°. Regarding this, compare the verse below
wherein the compound sarvasattveṣu occurs again and is translated as sems can kun la.
On the other hand, the Tibetan confirms the reading vyavasthitaṃ (rnam gnas pa) found
in the SS. Though not as clearly as in verses 17 and 18, the Tibetan again partially sup-
ports a reading found in the SS. Perhaps the translators had read something like sattveṣu
ca/tu/api vyavasthitam.

83
Verse thirty-one

[Sa p.164]

apratiṣṭhitanirvāṇaṃ265
nirnimittaṃ266 nirālayam |
vyāpakaṃ sarvasattveṣu
saṃbodhiḥ267 paramaṃ padam ||268

mi gnas pa yi269 mya ngan ’das ||

mtshan ma med cing gnas pa med270 ||
sems can kun la khyab byed pa ||
byang chub gnas mchog yin zhes bshad ||

265
apratiṣṭhita° B K1 Sa ] āpratiṣṭhita° Ba(sil.) Sa mu; apratiṣṭita° K 2
266
nirnimittaṃ Ba Sa ] nirmittaṃ B; nimittaṃ K1 K 2; nirmitaṃ Ba MS (haplography)
267
saṃbodhiḥ Ba(sil.) Sa(sil.) ] saṃbodhi B K1 K 2
268
Cp. GS 3.7cd: vyāpakaṃ niḥsvabhāvaṃ ca tad bodhiḥ paramaṃ padam ||.
269
gnas pa yi B G ] gnas pa’i T (hypometrical)
270
gnas pa med B G ] mya ngan ’das T (maybe an eye-skip)

84
 Verse thirty-one

The not-permanently-fixed nirvāṇa (apratiṣṭhitanirvāṇa) is

without characteristic marks and without a place; pervading
all sentient beings, it is supreme awakening (saṃbodhi), the
highest stage.271

271
The term apratiṣṭhitanirvāṇa is translated according to Edgerton’s BHSD (s.v. apra-
tiṣṭhita). Note that the Tibetan reads an additional zhes bshad in the final pāda, which
would usually correspond to something like ity uktam in Sanskrit, and does not render
the prefix sam°- of saṃbodhiḥ. It is not certain whether this indicates that the translators
had a different reading in their exemplar, or whether they took some liberty in their
translation.

85
Verse thirty-two

evaṃ matvā272 tu vai [B 37r] yogī

yo ’bhyased buddhimān273 sadā |
sa sidhyati274 na saṃdeho
mandapuṇyo ’pi mānavaḥ ||275

de nyid shes pa’i rnal ’byor pas276 ||

blo dang ldan pas rtag bslabs na277 ||
mi de bsod nams mi ldan yang ||
’grub par ’gyur bar the tshom med ||

272
matvā K1 Ba(sil.) Sa(sil.) ] satvā B K 2
273
’bhyased buddhimān K1p.c. Sa ] bhāsed buvimān B K 2 (“Bindefehler”); bhyased bud-
dhimān K1a.c.; bhāvayeb duddhimān Ba (sic!); bhāsayedbuviman Ba MS; bhāvayed bud-
dhimān Sa mu
274
sidhyati B K1a.c. K 2 ] siddhyati K1p.c. Ba(sil.) Sa(sil.)
275
Cf. HT I.viii.41/42 and Ratnākaraśānti’s Hevajrasahajasadyoga 9 (ed. Isaacson
2002: 464), both reading susamāhitaḥ instead of buddhimān sadā; also cited in the
Sārārthapañjikā, the commentary on Tillipāda’s Dohākośa (ed. Bagchi 1938: 45) and
the Viṣamapadabh(p)añjikā, the commentary on Saraha’s Dohākośa (ed. Bagchi 1938:
154).
276
rnal ’byor pas B G ] rnal ’byor pa T
277
bslabs na B G ] bslab nas T

86
 Verse thirty-two

And, having thought in that way, he, the wise yogin who would
frequently practice indeed, attains accomplishment, no doubt,
even when he is a person with little merit. 278

Note that the Tibetan translation differs in several places. In pāda a, the Sanskrit
278

reads evaṃ matvā. The Tibetan, on the other hand, sounds rather as if translating tatt-
vajñaḥ (de ñid shes pa’i; cp. AS 22c) without rendering the particles tu and vai. Here,
one could consider emending to de ltar shes pa’i, the reading attested in the canonical
translation. The particle na in B and G at the end of the second pāda could be an in-
dication that the translators had read ’bhyasad, i.e. a present participle, instead of the
optative form ’bhyased. Further, the relative pronoun yo is not reflected in the Tibetan.
Last but not least, the translation bsod nams mi ldan yang rather sounds like a translation
of apuṇyo ’pi than of mandapuṇyo ’pi.

87
Verse thirty-three

ācāryāt parataraṃ nāsti

trailokye279 sacarācare |
yasya prasādāt prāpyante
siddhayo ’nekadhā280 budhaiḥ ||281

khams gsum rgyu dang mi rgyu kun ||

slob dpon las ni mchog gzhan med ||
gang gi282 drin gyis mkhas pa rnams283 ||
dngos grub dpag med ’thob par ’gyur ||

279
trailokye K1 Sa(sil.) ] trailokya° B K 2 Ba
280
’nekad hā 𝝨 ] ’nakadhā K 2
281
This verse too is quoted in SS (p. 9), reading ācāryād gurutaro1 nāsti trailokye2 sa-
carācare | yasya3 prasādāt prāpyante siddhayo [’]nekadhā budhaiḥ ||; in his notes, Bendall
remarks the following: (1) Scan as trisyllable; (2) °kya Ms; (3) yatpr° contra metrum Ms.
282
gang gi T ] gang zhig B G
283
mkhas pa rnams B G ] thob ’gyur pa T (maybe an eye-skip)

88
 Verse thirty-three

Among what belongs the three worlds, whether moving or

unmoving, there is nothing higher than the ācārya. Thanks to
his kindness, manifold accomplishments are attained by the
wise ones.284

284
This verse is quoted in the SS which attests gurutaro instead of parataraṃ. Both
readings are hypermetrical and need to be read a bit faster than usual. Which of the two
readings is more original is difficult to ascertain. The wordplay in the SS (guru = teacher
/ heavy) is certainly attractive. Yet, the extra-canonical version of the Tibetan translation
reads mchog gzhan, which could be taken in support of parataraṃ. The reading in the
SS can be translated as “there is nobody heavier, i.e. more authoritative, than the teacher
...”. An emendation to paramaṃ would solve the metrical irregularity in the first pāda.
The third pāda, moreover, forms a ma-vipulā.

89
Verse thirty-four

vajrasattvaḥ sa vai [K2 21v] jñeyaḥ285

sarvabuddhair286 namaskṛtaḥ |
ācāryaḥ287 paramo288 devaḥ
pūjanīyaḥ prayatnataḥ289 ||290

rdo rje sems dpa’ de shes bya291 ||

sangs rgyas kun gyis292 phyag byas pa’i293 ||
slob dpon lha’i294 mchog yin pas ||
rab tu ’bad pas mchod par bya295 ||

285
vajrasattvaḥ sa vai jñeyaḥ K1p.c. Ba Sa ] vajrasatvasarv(v)ajñeyaḥ B K1a.c. K 2; vajrasat-
tvasarvajñaḥ yaḥ Ba MS; vajrasattvaḥ sarvajñeyaḥ Sagha
286
°buddhair B K1 Ba Sa (MSS °ḥ) ] °buddhaḥ K 2
287
ācāryaḥ K1p.c. Sa ] ācāryya° B K1a.c. K 2; ācāryāḥ Ba(sil.) Sa ka Sa mu
288
paramo K1 Ba Sa(sil.) ] paramā B K 2 Ba MS
289
prayatnataḥ K1 Ba(sil.) Sa(sil.) ] prayattataḥ B K 2
290
Cp. JS 1.28: yasmān na tatsamo hy asti pūjanīyo mahāmuniḥ | tasmāt sarva-prayatne-
na pūjayed guruvaraṃ vratī ||; cp. also JS 1.57.
291
de shes bya em.] de zhes bya B G; der shes byas T
292
kun gyis B G ] thams cad T
293
phyag byas pa’i B G ] phyag byas pa T
294
lha’i T ] lha mi’i B G
295
’bad pas mchod par bya B G ] ’bad de mchod bya’o T

90
 Verse thirty-four

He indeed, the ācārya, is to be known as Vajrasattva, ven-

erated by all the buddhas, to be worshiped with effort as the
highest god.296

296
In pāda c, the Tibetan versions B and G render paramo devaḥ with lha mi’i mchog
yin pas, which would rather correspond to devanarayor paramo, i.e. “since the teacher
is the highest among gods and men ...”. Further, one could consider emending yin pas
(BGT) to yin pa(r) to better reflect the reading in the Sanskrit.

91
Verse thirty-five

sa eva tathatārūpī
lokānugrahahetunā |
rūpam āśritya saṃvṛtyā
saṃsthito297 yogapīṭhake ||

de nyid de bzhin nyid gzugs las ||

sems can rnams [T 189] ni rjes gzung phyir298 ||
kun rdzob gzugs la brten nas ni ||
rnal ’byor gzhi la gnas par ’gyur299 ||

297
saṃsthito K1 Ba Sa(sil.) ] saṃsthitā B K 2; saṃsthita Ba MS
298
rjes gzung phyir B G ] rjes bzung phyir T
299
gzhi la gnas par ’gyur B G ] gnas la gnas par ’grub T

92
 Verse thirty-five

Only he has the nature of suchness for the sake of benefiting

the people. Having taken a form [just] by convention, he is
situated at the seat of yoga (yogapīṭha).300

The Tibetan version T renders saṃsthito yogapīṭhake with the reading rnal ’byor
300

gnas la gnas par ’grub. Even though both readings rnal ’byor gnas la for yogapīṭhake,
and gnas par ’grub for saṃsthito might be acceptable, we have to suspect that this rea-
ding might again be the result of a redaction process that does not necessarily reflect the
original Sanskrit very literally.

93
Verse thirty-six

[Ba p.23]

mṛtyur eṣa vikalpo ’yaṃ301

na bhāvaḥ302 sarvavastuṣu |
hanyate svavikalpena
pṛthagjanavijṛmbhitaiḥ303 ||304

’chi ba zhes ’di305 rnam rtog ste ||

dngos rnams kun la yod ma yin ||
so so’i skye bo rnam ’phrul pa’i306 ||
rang gi rnam par rtog pa bcom ||

301
vikalpo ’yaṃ Ba Sa ] vikalpāya B; vikalpo yam K1 Sa ka Sa kha Saga Sagha; vikalpāyam
K 2; vikalpaya Ba MS
302
na bhāvaḥ Ba(sil.) Sa ] na bhāva° B; abhāva° K1a.c. K 2; abhāvaḥ K1p.c. Sa ka Sa kha Saga Sagha
303
pṛthag° 𝝨 ] pṛthakag° K 2 (“Trennfehler”) • °vijṛmbhitaiḥ K1p.c. Ba(sil.) Sa(sil.) ] °vi-
jṛmbhiteḥ B K 2; °vi(gha)mbhiteḥ K1a.c.
304
A similar version is quoted in Vāgīśvarakīrti’s Mṛtyuvañcanopadeśa 4.81 (ed. Sch-
neider 2010), reading tad āha | mṛtyur nāma vikalpo ’yam abhāve sarvavastunaḥ | hanyate
svavikalpena pṛthagjanavijṛmbhitair iti ||.
305
zhes ’di T ] zhes bya B G
306
rnam ’phrul pa’i T ] rnams ’khrul pas B G

94
Verse thirty-six

This [so-called] death is [just] a concept (vikalpa). There is

no true existence (bhāva) in anything. It/one is destroyed by
the concept of a self (svavikalpa) [and] what is manifoldly
displayed by ordinary beings.307

307
A very similar version of this stanza is quoted in Vāgīśvarakīrti’s Mṛtyuvañcanop-
adeśa 4.81 (ed. Schneider 2010), reading tad āha | mṛtyur nāma vikalpo ’yam abhāve
sarvavastunaḥ | hanyate svavikalpena pṛthagjanavijṛmbhitair iti ||. Note that the Tibetan
translators of the AS also might have read nāma instead of eṣa, a reading that is also
found in Lūyīpād a’s Śrīcakrasaṃvarābhisamaya 43i (ed. Sakurai 1998). There are basi-
cally two ways of interpreting this stanza, so it seems, and Lakṣmīṅkarā might have been
consciously playing with the double meaning (śleṣa) of the word mṛtyu when composing
this final verse. On the one level, the sentence can be interpreted as done above. In
this interpretation the two halves are taken as two separate sentences with two different
subjects, and sva° is understood in the sense of svātmabhāva°. This is more or less the
interpretation found in Krug (2018: 128). On the other level, perhaps the actual intent of
the verse, this stanza also can be taken to mean “This dying [in saṃsāra] is a concept
(vikalpa). There is no true existence (bhāva) in anything. It is destroyed by means of
one’s own concept together with what is displayed by ordinary beings[, such as rāga,
moha and so forth].” In this interpretation, the single subject of the verse is death which
equals vikalpa. The term svavikalpena is not to be understood in its common sense of
“wrong concepts”, as opposed to the first interpretation in which svavikalpena means
something like svātmabhāvavikalpena, but rather in the sense of upāyena, i.e. “using
concepts as a means”. Furthermore, the concept of killing (maraṇa) in the sense of a
“ritual killing [of wrong conceptions]”, in which death than equals liberation (maraṇam
ityādi maraṇam mokṣah), is not too uncommon in Buddhist tantric literature such as,
for instance, in HT I.v.21: pitari praptaṃ yat saukhyaṃ tat sukhaṃ bhujyate svayaṃ |
maraṇaṃ yena sukhena tat sukhaṃ dhyānam ucyate ||, which is commented on in the
Yogaratnamālā (ed. Snellgrove 1959: 119) as follows: maraṇam ityādi maraṇam mokṣah
| tatraiva cittacetasikāvidyālakṣaṇānām astaṅgamāt tad eva maraṇam | yena sukheneti
sakalakleśopakleśakṣayalakṣaṇena mahāsukheneti ||. The extra-canonical Tibetan ver-
sion, basically supporting the second interpretation, can be translated as follows: “This
called death is a misconception, it does not exist in anything. It destroys one’s concept of
a self, which is displayed by ordinary people.” The canonical version, again significantly
deviating from the extra-canonical one and the Sanskrit, reads ’chi bdag de ni rtag pa
che || thams cad ma lus stong pa nyid || rang gi rtog pa mi snang la || gzhan dag de la dga’
bar byed ||. This, tentatively and rather freely, could be translated as follows: “Death is
a great concept, everything without exception is emptiness. When the concept of a self
is not appearing, everything else is made blissful in it[, i.e in death].” Though it remains
rather speculative how to interpret the verse exactly, it might be taken as referring to the
idea of liberation at the moment of death.

95
Colophons

Sanskrit (without application of all sandhis): śrīmadoḍḍiyāna-

vinirgatā308 mahāyogapīṭhāgatā akhilayogatantratattvagar-
bhā309 śrīlakṣmīmukhakamalād viniḥsṛtā310 svādhiṣṭhānakra-
modayā advayasiddhir311 nāma sādhanopāyikā312 samāpteti ||

Tibetan: gNyis med grub pa zhes bya ba’i sgrub pa’i thabs
O rgyan313 zhes bya ba314 rnal ’byor gyi gnas mchog [B 276]
chen por dngos grub [G 235] brnyes pa’i Lha mo dPal chen mo
zhes bya bas mdzad pa rdzogs so315 || rgya gar gyi mkhan po
Śraddhākaravarman dang | lo tsā ba Rin chen bZang pos bsgyur
ba’o316 || ||

308
°oḍḍiyāna° K1p.c. Saga Sagha ] °oḍiyāna° B K1a.c. K 2 Ba Sa
309
akhila° B K1 K 2 Ba ] ’khila° Sa(sil.) • °tattvagarbhā Sa ] °tatvagarbha° B K 2; °garbha°
K1 Sa kha; °tattvagarbhā- Ba(sil.)
310
viniḥsṛtā K1p.c. Ba(sil.) Sa(sil.) ] viniḥśṛtā B K1a.c. K 2
311
advayasiddhir K1 Ba Sa(sil.) ] advayasiddhi° B K 2 Ba MS
312
sādhanopāyikā Ba Sa ] sādhanoyogikā B K1a.c. K 2 Ba MS; sādhanopayogikā K1p.c. Sa ka
Sa kha Saga Sagha
313
gNyis med grub pa zhes bya ba’i sgrub pa’i thabs O rgyan B G ] gNyis su med pa’i
grub pa zhes bya ba’i bsgrub pa’i thabs rdzogs so || U rgyan T
314
zhes bya ba BG] zhes bya T
315
rnal ’byor gyi gnas mchog chen por dngos grub brnyes pa’i Lha mo dPal chen mo
zhes bya bas mdzad pa rdzogs so B G ] rnal ’byor gyis gnas chen po ngos grub brnyes pa’i
Lha mo dPal chen mo las byung ba zhes ba rdzogs so || śubhaṁ T
316
rgya gar … bsgyur ba‘o B G ] om. T (Note that the spelling in B and G is shrad.d.ha
ka ra warma.)

96
Sanskrit: The sādhanopāyikā317 called Advayasiddhi which has
its arising from the ‘stage of self-empowerment’ (svādhiṣṭhā-
na-krama) is finished. 318 Having come from the Glorious Oḍḍi-
yāna [and] having arrived at the Great Seat of Yoga (mahāyo-
gapīṭha), it — which has as its central part the reality of all
yogatantras — was issued from the lotus-mouth of Śrīlakṣmī.

Tibetan: The sādhana called Advayasiddhi — “The Non-dual

Accomplishment” — composed by [her] called the Great and
Glorious Goddess who attained accomplishment at the Su-
premely Great Seat of Yoga known as O rgyan is finished. It
has been translated by the Indian scholar Śraddhākaravarman
and the [Tibetan] Lo tsā ba Rin chen bZang po.

317
The reading sādhanopāyikā follows the previous editions Ba Sa, basically confir-
med by the Tibetan translation. The Sanskrit MSS, however, in fact attest sādhano(pa)
yogikā.
318
The precise meaning of the expression svādhiṣṭhānakramodayā remains uncertain.
This uncertainty is owed to the ambiguity of the referent of the term svādhiṣṭhānakrama
and its syntactic relationship with the final member of the compound (°udaya). On the
one hand, it is not entirely certain whether the first part of the compound refers to the te-
chnical term svādhiṣṭhānakrama in the context of the Pañcakrama, to a particular work
titled svādhiṣṭhānakrama, or to a particular work having the svādhiṣṭhānakrama as its
main concern. On the other hand, it is possible to take the compound as a genitive, ablati-
ve or perhaps even as an instrumental bahuvrīhi. In any case, it is worth mentioning that
the AS contains notable parallels to the Svādhiṣṭhāna(krama)prabheda of Āryadeva (cp.
SP 6–8, 14–15, 35–39, 46–50) which — as is also the case for the translations of the AS
and JS — was translated into Tibetan by Śraddhākaravarman and Rin chen bZang po un-
der the title bDag byin gyis brlab pa’i rim pa rnam par dbye ba (Tōh. 1805). Among the
various texts concerned with the svādhiṣṭhānakrama, attention may also be drawn to the
following two works, a closer study of which might reveal further information relevant
to Lakṣmīṅkarā and the AS: 1) the *Sahajasamvarasvādhiṣṭhāna (dPal lhan cig skyes
pa bde ba’i mchog bdag byin gyis brlab pa), Tōh. 1459, by Indrabhūti the Middle (’bring
po) and the 2) Oḍḍiyānasvādhiṣṭhānakramavajrayoginīsādhana, IASWR MBB-II-240
(cf. English 2002: 378). A commentary on the Pañcakrama by a Śrīlakṣmī, apparently
a Kashmiri scholar (kha che mkhas pa chen po dPal ldan Lakṣmī), is preserved in the
bsTan ’gyur under the title Rim pa lnga’i ’grel pa’i don gsal bar byed pa (Tōh. 1842).
This work, however, does not seem to stem from the same Lakṣmī as the authoress of the
AS.

97
Bibliography
Bagchi, Prabodh Candra. 1938. Dohākoṣa (Apabramśa Texts
of the Sahajayāna School). Calcutta: Calcutta University Press.
Bendall, Cecil. 1905. Subhāṣita-Saṃgraha — An Anthology of
Extracts from Buddhist Works Compiled by an Unknown Au-
thor, to Illustrate the Doctrines of Scholastic and Mystic (Tān-
trik) Buddhism. Leipzig: Harrassowitz.
Chakravarti, Chintaharan. 1984. Guhyasamājatantrapradīpo
d-dyotanaṭīkāṣaṭkoṭivyākhya .̄ Patna: Kashi Prasad Jayaswal
Research Institute.
Davidson, Ronald M. 2002. Indían Esoteric Buddhism, A So-
cial History Of The Tantric Movement. New York: Columbia
University Press.
Dimitrov, Dragomir. 2000. ‘Lakṣmī — On the Identity of Some
Indo-Tibetan Scholars of the 9th–13th Centuries’. In: Zentral-Asia-
tische Studien (Vol. 30), 9–26. Wiesbaden: Harrassowitz Verlag.
Donaldson, Thomas Eugene. 2001. Iconography of the Bud-
dhist Sculpture of Orissa. Vol.1 (Text). New Delhi: Indira
Gandhi National Centre for the Arts. Abhinav Publications.
Dowman, Keith. 1985. Masters of Mahāmudrā. Songs and His-
tories of the Eighty-Four Buddhist Siddhas. Albany: State Uni-
versity of New York Press.
Edgerton, Franklin. 1953. Buddhist Hybrid Sanskrit Grammar
and Dictionary. New Haven: Yale University Press.
English, Elizabeth. 2002. Vajrayoginī: Her Visualizations, Rit-
uals, & Forms. A Study of the Cult of Vajrayoginī in India.
Boston: Wisdom Publications.
Gerloff, Torsten. 2020. Saroruhavajra’s Hevajra-Lineage:
A  Close Study of the Surviving Sanskrit and Tibetan Works.
Part I: Saroruhavajra’s Works & Part II: Commentarial Works.
Hamburg: Department of Indian and Tibetan Studies, Universi-
tät Hamburg (Indian and Tibetan Studies Series 7).
Gerloff, Torsten and Julian Schott. 2020. ‘Towards a Reassess-
ment of Indrabhūti’s Jñānasiddhi’. In: Buddhist Studies Re-
view, 37.2, pp. 241–260.

98
Grünwedel, Albert. 1914. Bka’ babs bdun ldan. Täranätha’s
Edelsteinmine, das Buch von den Vermittlern der Sieben Ins-
pirationen. Aus dem Tibetischen übersetzt von Albert Grünwe-
del. I–II. Petrograd: Academie Imperiale des Sciences (Biblio-
theca Buddhica 18).
Grünwedel, Albert. 1916. Die Geschichten der vierundachtzig
Zauberer (Mahäsiddhas). Leipzig: BG Teubner (Baessler-Ar-
chiv, Bd. 5, Nr. 4/5).
Harding, Sarah (trnsl.). 2002. Jamgön Kongtrul. Creation and
Completion – Essential Points of Tantric Meditation (trans-
lated and introduced by Sarah Harding with commentary by
Khenchen Thrangu Rinpoche). Boston: Wisdom Publications.
Isaacson, Harunaga. 2002. ‘Ratnākaraśānti’s Hevajrasaha-
ja-sadyoga (Studies in Ratnākaraśānti’s tantric works I)’. In:
Raffaele Torella (ed.). Le Parole e i Marmi: studi in onore di
Raniero Gnoli nel suo 70° compleanno, Vol.1, 457–487. Roma:
Istituto Italiano per l’Africa e l’Oriente (Serie Orientale Roma
XCII).
Isaacson, Harunaga and Francesco Sferra. 2015. The Sekanird-
eśa of Maitreyanātha (Advayavajra) with the Sekanirdeśapañ-
jika ̄ of Rāmapāla: Critical Edition of the Sanskrit and Tibetan
Texts with English Translation and Reproductions of the MSS.
Napoli: Universita ́ Degli Studi Di Napoli “L’Orientale” (Man-
uscripta Buddhica 2).
Jackson, Roger R. and Matthew T. Kapstein (eds.). 2011. ‘Ma-
hāmudra ̄ and the bKa’-brgyud Tradition’ In: PIATS 2006:
Tibetan Studies: Proceedings of the Eleventh Seminar of the
International Association for Tibetan Studies, Königswinter
2006 (Vol. 25), Königswinter: International Association for Ti-
betan Studies.
Kragh, Ulrich Timme. 2010. ‘On the Making of the Tibetan
Translation of Lakṣmi ›̄ s *Sahajasiddhipaddhati: ‘Bro Lotsa ̄
ba Shes rab Grags and his Translation Endeavors. (Materials
for the Study of the Female Tantric Master Lakṣmi ̄ of Uḍḍiyā-
na, part 1)’. In: Indo-Iranian Journal (Vol. 53.), 195–232. Lei-
den: Koninklijke Brill NV.
Kragh, Ulrich Timme. 2011. ‘Appropriation and Assertion of
the Female Self, Materials for the Study of the Female Tantric
Master Lakṣmī of Uḍḍiyāna’. In: Journal of Feminist Studies in
Religion (27.2), 85–108.

99
Krug, Adam Charles. 2018a. ‘Tantric Epistemology and the Prob-
lem of Ineffability in the Seven Siddhi Texts’. In: Manel Herat
(ed.). Buddhism and Linguistics: Theory and Philosophy, 149–184.
Cham: Palgrave Macmillan (Springer International Publishing AG).
Krug, Adam Charles. 2018b. The Seven Siddhi Texts: The Oḍi-
yāna Mahāmudra ̄ Lineage in its Indic and Tibetan Contexts.
Santa Barbara: University of California (Dissertation).
Krug, Adam Charles. 2020. ‘The Seven Siddhi Texts (Grub pa
sde bdun): Remarks on the Corpus and Its Employment in Sa
skya-Bka’ brgyud Mahāmudrā Polemical Literature’. In: Jack-
son, Roger R. and Klaus-Dieter Mathes (eds.). Mahāmudrā in
India and Tibet, 90–122. Leiden, Boston: Koninklijke Brill NV.
Malalasekera, G. P. et al. (eds.). 1997. Encyclopaedia of Bud-
dhism Volume 1, 1961–1965. Sri Lanka: Government of Ceylon
Press (reprint).
Matsunaga, Yukei. 1978. The Guhyasamāja Tantra: A New
Critical Edition. Osaka: Toho Shuppan.
Mishra, Ramprasad. 1995. Advayasiddhi, the Tāntric View of
Lakṣmīṅkarā. Delhi: Kant Publications (Oriental Publishers).
Monier-Williams, Sir Monier. 1997. A Sanskrit-English Dictio-
nary — Etymologically and Philologically Arranged with spe-
cial reference to Cognate Indo-European Languages. Delhi:
Motilal Banarsidass (First Edition Published 1899 by Oxford
University Press).
Pandey, Janardan. 1990. ‘Svādhiṣṭhānaprabheda’. In: Dhīḥ,
Volume 10, 20–24.
Patel, Prabhubhai Bhikhabhai. 1949. Cittavisuddhiprakarana
of Aryadeva. India: Visva-Bharati.
Rinpoche, S. and V. Dwivedi. 1987. Guhyādi-Aṣṭasid-
dhi-saṅgraha. Sarnath, Varanasi: Central Institute of Higher
Tibetan Studies (Rare Buddhist Text Series 1).
Rinpoche, S. and V. Dwivedi. 1992. Mahāmāyātantram — Rat-
nākaraśāntiviracitaguṇavatīṭīkāsayutam. Sarnath, Varanasi:
Central Institute of Higher Tibetan Studies (Rare Buddhist Text
Series 10).
Robinson, James B. 1979. Buddha’s Lions — The Lives of the
eighty-four Siddhas. Berkeley: Dharma Publishing.
Roerich, George N. 1995. The Blue Annals. New Delhi: Motilal
Banarsidass Publishers PVT. LTD. (reprint).

100
Śāstri ,̄ Haraprasād. 1898. ‘The discovery of a work by Āryadēva
in Sanskrit’. In: Journal of the Asiatic Society of Bengal. Cal-
cutta: Baptist Mission Press, Vol. LXVII, 175–185.
Śāstri ,̄ Viśvanātha. 1872. ‘Piṅgalacchandaḥsūtram — Ch-
handah ̣ Sūtra of Pingāla Āchārya with the Commentary of
Halāyudha’. In: Asiatic Society of Bengal. Bibliotheca Indica:
A Collection of Oriental Works (New Series, No. 258, Fascicu-
lus II). Calcutta: Ganeśa Press.
Sakurai, Munenobu. 1998. ‘Cakrasaṃvarābhisamaya-no gen-
tenkenkyū – bondun-kōtei-tekusuto’ (A Critical Study of the
Cakrasaṃvarābhisamaya). In: Chisan-gakuhō (Journal of
Chisan Studies), Vol. 47, 1–32.
Sandhak. 198?. Krung go’i bod kyi shes rig zhib ‘ jug lte gnas
su nyar ba’i ta la’i lo ma’i bstan bcos (spyin shog ‘dril ma’i
par) kyi dkar chag mdor gsal (中国藏学研究中心收 藏的梵文
贝叶经 (缩微胶卷) 目录).
Schneider, Johannes. 2010. Vāgīśvarakīrtis Mṛtyuvañcan­
opadeśa, eine buddhistische Lehrschrift zur Abwehr des Todes.
Wien: Verlag der Österreichischen Akademie der Wissenschaf-
ten (Beiträge zur Kultur- und Geistesgeschichte Asiens 66).
Sferra, Francesco. 2000. ‘Sanskrit Manuscripts and Photos of
Sanskrit Manuscripts in Giuseppe Tucci’s Collection. A Prelim-
inary Report’. In: Balcerowicz, Piotr and Marek Mejor (eds.).
On the Understanding of Other Cultures — Proceedings of the
International Conference on Sanskrit and related studies to
commemorate the centenary of the birth of Stanisław Schayer
(1899-1941). Warszawa: Instytut Orientalistyczny, Uniwersytet
Warszawski (Studia indologiczne, Tom 7), 397–447. 
Sferra, Francesco and Stefania Merzagora. 2006. The Sekod-
deśaṭīkā by Nāropā (Paramārthasaṃgraha). Roma: Istituto
Italiano per L’Africa e L’Oriente.
Shendge, Malati J. 1964. ‘Advayasiddhi’. In: Journal of the Ori-
ental Institute (The M.S. University Oriental Series 8), 1–30.
Shendge, Malati J. 2004. Ṣaṭsāhasrikā-hevajra-ṭīkā — A Criti-
cal Edition. Delhi: Pratibha Prakashan.
Snellgrove, Dwight Lowell. 1959. The Hevajra Tantra —
A Critical Study. (Part I — Introduction and Translation; Part
II — Sanskrit and Tibetan Texts. London, New York, Toronto:
Oxford University Press.
Szántó, Péter-Dániel. 2015. ‘Tantric Prakaraṇas’. In: Brill’s En-

101
cyclopedia of Buddhism, Volume 1, 756–761. Leiden: Konin-
klijke Brill NV.
Templeman, David. 1989. Taranātha’s bKa’ babs bdun ldan —
The Seven Instruction Lineages by Jo.nang Taranātha. Dha-
ramsala: Library of Tibetan Works and Archives.
Tribe, Anthony. 2016. Tantric Buddhist Practice in India:
Vilāsavajra’s commentary on the Mañjuśrī-nāmasaṃgīti. A
critical edition and annotated translation of chapters 1–5 with
introductions. London, New York: Routledge.
Tripathi, Ram Shankar and Thakur Sain Negi (eds.). 2001. He-
vajratantram With Muktāvalī Pañjikā of Mahāpaṇḍitācārya
Ratnākaraśānti. Sarnath, Varanasi: Central Institute of Higher
Tibetan Studies (Bibliotheca Indo-Tibetica Series XLVIII).
Ui, Hakuju et al. (eds). 1934. The Tibetan Tripiṭaka — Tōhoku
University Catalogue of the sDe-dGe Edition of the Tibetan
Sacred Canon (bKa’-gyur and bsTan-’gyur) — A Complete Cat-
alogue of the Tibetan Buddhist Canons. Sendai: Tōhoku Uni-
versity.
Ujeed, Sangseraima. 2016. ‘Dge-slong-ma dpal-mo, the Prin-
cess, the Mahasiddha, the Nun and the Lineage Holder – as
Presented in the thob-yig of Za-ya Paṇḍita Blo-bzang ’phrin
las (1642–1715)’. In: Journal of the Oxford Centre for Buddhist
Studies (JOCBS) 2016(5), 128–166.
Vargas-O’Bryan, Ivette. 2001. ‘The Life of dGe slong ma dPal
mo: The Experience of a Leper, Founder of a Fasting Ritual and
Transmitter of Buddhist Teachings on Suffering and Renunci-
ation in Tibetan Religious History’. In: Journal of the Interna-
tional Association of Buddhist Studies (JIABS) Vol. 24, No. 2,
157–158.
Yamada, Ryūjyo. 1936. ‘Cittaviśuddhiprakaraṇa Sanskrit edi-
tion, Japanese translation, Tibetan translation’ (Japanese; 心
障清浄論 梵文・邦訳・蔵文). In: Shinshōshōjyōron (Bonbun,
Hōyaku, Zōbun), Tōhoku Imperial University, Bunka Vol. 3–8,
883–949.

102
Appendices

Additional Passage in T (184,5–186,1) — Lo rgyus of

Lakṣmīṅkarā’s Advayasiddhi

grub pa’i gnas su grub pa’i rigs su ’khrungs ||

grub mchog [184,6] rnams la grub skor gdams pa gsan
||
grub pa thob nas gNyis med grub pa gsungs ||
dPal ldan dPal mo’i gsung ’di mnyan par gyis ||

de la ’di’i lo rgyus ni | nub phyogs U rgyan gyi yul du rgyal po

Indrabodhi’i319 sras mor sku ’khrungs | lo bcu gnyis [185,1] la rab
tu byung | rigs pa’i gnas lnga la sbyangs par mkhas par gyur |
lo bcu lnga la mdze nad drag po cig byung | jo bo khang sar pa
ni mjal ba la byon pas | dkor gnyer gyis skye ba dman pa’i steng
du mdze nad byung ’dug | zer nas mjal kha bkag | sgo rgyab nas
[185,2] g.yol ba btag pas ’phags pa thugs rje chen po khri kha
na320 dkor gnyer gyis321 mthong bas | khyod mdze mo nus pa can
cig ’dug pas ngo mtshar che gsung nas nang du khrid | de nas
dpal gyis dngos grub la thugs rje chen po’i grub skor322 zhus nas
| lo bcu gnyis nas grub pa thob par lung [185,3] bstan ||

Li khar shing ’phel gyi lha khang gi323 sgo rtsa na sdod pas |
dkor gnyer gyis324 khyod kyis sa cha ’di ma rul zer nas sdod du
mi ster ’jag spyil yang ’phangs | lha khang gi lho ngos na sdod
pas lo gsum na lung bstan yang byung mdze nad yang phan |
’dug gang dang lha khang gi sgo phyogs su thams cad [185,4]
dpal mos325 bstan nas ’dug |

319
In this and the following lo rgyus, we have standardized the forms i n.tra b.ho di, i
n.tra b.ho d.hi(e) and i n.tra b.hu ti to Indrabodhi and Indrabhūti respectively. Since the
first spelling is an attested alternative for Indrabhūti (see Kragh 2010: 197 note 6 et al.),
we have accepted these variants and kept them without standardising them.
320
’phags pa thugs rje chen po khri kha na conj.] thugs rje che pa khri kha na ’phags pa
Ta.c.; thugs rje chen pa khri kha nas ’phags pa Tp.c.
321
dkor gnyer gyis em.] kor gnyer gyi T
322
grub skor em.] bsgrub skor T (cp. 184,6)
323
lha khang gi corr.] lha gang gi T
324
dkor gnyer gyis corr.] kor gnyer gyis T
325
lha khang gi sgo phyogs su thams cad dpal mos conj.] lha gang gis sgo thams cad dpal
mo phyogs su T

103
lo bcu gnyis song ba’i sa ga zla ba’i tshes gcig la326 sgrol ma dkar
sngon gzigs | tshes brgyad la don zhags lha lnga gzigs | bco lnga la
bcu cig zhal gzigs | ’phags pa grub thob327 Thugs kyi rDo rje dang In-
drabodhi la gsang sngags lam rnams [185,5] zhus nas | lho dpal gyi ri
la328 bsgoms pas phyag rgya chen po mchog gi329 dngos grub brnyes |

mthar phyin pa’i don du grub thob Phag tshang ba dang lhan cig
tu330 spyod pa mdzad pa’i bkur ba bsal ba’i don du kha sar pa ṇa’i
dus ston la | khong rang gi dbu331 gri gug gis bcad nas mkhar [185,6]
sil gyi rtse bskyon | steng gi nam mkha’ la332 ’phags nas sprin gseb
du gar mdzad | sa la babs nas bden pa brjod pas dbu dang lus snga
ma bzhin du gyur |

lCam gyis333 spyan ras gzigs phyag stong spyan stong du sprul
nas phyag re re la mchod rdzas mi ’dra ba re re thogs nas ming po
I[186,1]ndrabodhi la phul ba dang | rgyal pos bde mchog phyag
stong dang ldan par sprul nas blangs | rdo rje sems dpa’ khyed
gnyis sku re’i don du de ltar sbyar334 rung ma zer skad | de Ye shes
grub pa la brten nas gNyis med grub pa mdzad pa ni ’di ltar || ||

Translation — Lo rgyus of Lakṣmīṅkarā’s Advayasiddhi

Born into the lineage of accomplished ones in a place of
accomplishment,
As regards the[se] supremely accomplished ones, listen to the
instructions of the ‘accomplishment cycle’335!
Having attained accomplishment, she taught the
Advayasiddhi;
Listen to this teaching of Glorious Lakṣmī (dPal ldan dpal mo)!

326
gcig la corr.] cig la T
327
’phags pa grub thob em.] ’phags pa thob T
328
dpal gyi ri la corr. ] dpal gyis ri la T
329
mchog gi em. ] mchog gis T
330
tu corr.] du T
331
rang gi dbu corr.] rang gis dbu T
332
nam mkha’ la corr.] namkha’ la T
333
lcam gyis em.] lcam gyi T
334
sbyar conj.] byar T
335
This phrase might be taken as referring to the Grub pa sde bdun.

104
In this regard, the traditional account (lo rgyus) of it [i.e.
Lakṣmīṅkarā’s Advayasiddhi] is as follows:336 In the western di-
rection of the land of Oḍḍiyāna (u rgyan), [Lakṣmīṅkarā] was born
as the sister (sras mo) of King Indrabodhi [the Second]. At twelve
years of age (lo bcu gnyis la), she renounced worldly life (rab tu
byung) [and] became skillful in the training of the Five Sciences
(rigs pa’i gnas lnga). At fifteen years of age, a severe leprosy (mdze
nad drag po) befell [her]. As she came for an audience, the Master
of Khang sar (jo bo khang sar pa) refused [her], the caretaker of
the temple saying ‘leprosy had come upon [you,] vile woman (skye
ba dman pa)’. Because the caretaker of the temple saw the No-
ble Compassionate One on a throne337 as he was examining [her]
being kept away outside of the door,338 he let her in, saying ‘how
wondrous, you are a leper lady with some capability!’. Then, af-
ter the glorious one had requested the ‘Accomplishment Cycle of
the Greatly Compassionate One’ (thugs rje chen po’i bsgrub skor)
for [her own] accomplishment, she was given the prophecy (lung
bstan) of attaining accomplishment after twelve years.
As she was standing at the gate of the temple of Li khar shing
’phel,339 the caretaker of the temple spoke ‘Don’t spoil this
piece of land!’ and did not grant [her] a place to stay, but urged
[her] to a thatch hut [outside the temple complex, where she
could stay]. As she was staying in the southern side of the tem-
ple, within three years both the prophecy came about and the
leprosy got better. Teaching all people (thams cad) at the side
of the temple gate where she was dwelling, Lakṣmī (dpal mo)
remained [at the temple site].340

336
It is noteworthy that this account, though being given here in the context of the Ad-
vayasiddhi and the Grub pa sde bdun, strongly resembles the rnam thar of dGe slong ma
dPal mo (Skt. Bhikṣuṇī Lakṣmī), whose association with the smyung gnas tradition(s)
seems widely accepted. The life story of dGe slong ma dPal mo in the context of smyung
gnas has been studied, for instance, in Vargas-O’Brian 2001 and Ujeed 2016. Dowman
(1985: 375), Dimitrov (2000: 9), Vargas-O’Brian (2001: 163–164, note 19) and Ujeed
(2016: 152) all agree on the point that there was more than one Lakṣmī. The similarity of
the various accounts impedes a precise identification.
337
Here we have conjecturally emended the reading transmitted in T and taken ’phags
pa thugs rje chen po as an epithet of Avalokiteśvara. Both readings in the xylograph,
before and after correction, do not make any sense to us.
338
We have kept the slightly less smooth reading g.yol ba. Here one could also consider
emending to gsol ba.
339
In his translation of the Deb ther sngon po (Blue Annals), Roerich identifies Li khar
shing ’phel as Puṇḍravardhana (Malda and Dinajpur; see Roerich 1995: 1008).
340
The syntax and grammar of this sentence are quite doubtful, and thus the meaning

105
On the first day of the fourth month after twelve years had passed,
she had a vision of White Tārā. On the eighth day, she saw five
deities of Avalokiteśvara (don zhags lha lnga).341 On the fifteenth
[day], she saw the eleven-faced [form of Avalokiteśvara]. After
she had requested the paths of mantra from the noble siddhas
Cittavajra and Indrabodhi, she attained the accomplishment
of the Supreme Great Seal through meditation practice on the
southern [mountain] Śrīparvata (lho dpal gyi ri).342
In order to bring to completion the siddha Phag tshang ba343 and in
order to make clear the worship that is performed in the practice of
the innate, on the occasion of the festival of Khasarpaṇa [Avalokiteś-
vara] she cut off her head with a chopper and stuck it on top of a
skull-staff. Having ascended into the sky above, she performed a
dance among the clouds. When she descended to the ground, her
head and body were as before [again] in order to proclaim the truth.
The lady [Lakṣmīṅkarā], having emanated Avalokiteśvara with
a thousand arms and a thousand eyes, holding various differ-

remains somewhat unclear. There are, besides our attempt, also other possibilities of
correcting and interpreting this sentence. The use and position of thams cad and the fact
that neither the grammatical subject nor the object of this sentence are clearly marked,
makes it particularly difficult to propose a solid reconstruction. Another possibility is to
shift the shad and read phan ’dug | khang dang lha khang gi (…).
341
Here, probably, short for don yod zhags pa lha lnga ba ri lugs. Surrounding the
two-armed white form are the four deities Amoghapāśa, Hayagrīva, Ekajaṭā and Bhṛkuṭī,
clockwise from East to North (cf. TBRC ID T595).
342
Both of the above-mentioned siddhas, Indrabodhi (Bla ma dGe slong gNyid log pa)
and Cittavajra (dPal ldan Thugs rDo rje) are also mentioned in the SSP (fol. 12b6 ff.),
followed by a short section in which Lakṣmī gives an account of her relation to them (SSP
sDe dge Vol. 52: fol. 13b7–14a4). It is further worth mentioning that, apart from the places
U rgyan and dPal gyi ri, we were not able to trace any of the other elements described in
T in our glance at the SSP. A closer study of the SSP and the accounts preserved in the
smyung gnas tradition of dGe slong ma dPal mo is highly recommended. The transmis-
sion lineage involving Indrabhūti and Cittavajra, though described there differently, is
also found in Dowman who, in his historiography section, states that “Lakṣmī received
the Guhyasamajā from Cittavajra (also described as her brother), and taught it to In-
drabhūti.” (1985: 375).
343
The siddha Phag tshang ba is mentioned in a short rnam thar of Anaṅgavajra (gNyis
med rol pa’i slob dpon yan lag med pa’i rdo rje’i rnam thar), where he is referred to as
a siddha of inferior social status (rigs ngan) in U rgyan. Regarding this, see bKa’ brgyud
rin po che gser gyi ’phreng ba’i gdan rabs bla ma brgyud pa’i rnam thar, in The col-
lected works (gsung ‘bum) of rGyal dbang Kun dga’ Rin chen (TBRC: W23892), Vol.1,
fol. 36,5ff. See also rDo rje’i rnal ‘byor gyi ‘khrid yig mthong ba don ldan gyi lhan thabs
‘od brgya ‘bar ba, in Jo nang rje btsun tA ra nA tha’i gsung ’bum (TBRC: W1PD45495),
dPe bsdur ma, Vol. 7, 249. The syntactical position of grub thob Phag tshang ba is slight-
ly confusing. We have taken it in the sense of grub thob Phag tshang ba mthar phyin pa’i
don du.

106
ent offerings in each hand, offered [these] to her brother In-
drabodhi. And [her brother] the king, assuming the form of
the thousand-armed Saṃvara, received [the offerings] saying
(skad) ‘United in this way,344 one cannot possibly tell apart
(rung ma zer) the content of the individual forms of the two,
Vajrasattva and You!’.
Thus is the Advayasiddhi, a work based on the Jñānasiddhi.

Additional Passage in T (189,2–190,1)345

—Description of the AS
§1a: de ltar gNyis su med par grub pa la | gnyis su med pa ni
thabs dang [189,3] shes rab bo || thabs ni rnam par rtog pa dang
bral ba’i ye shes so || shes rab ni rang bzhin gyis ’od gsal ba’o ||
gnyis su med pa ni zung du ’jug pa’i rdo rje yin la | sems dpa’
ni gzugs kyi skur bzhengs pa’o || de ni rdo rjes sems dpa’o || de
bsgrub pa’i thabs las [189,4] bsgrub gzhi ni rang346 sems can
thams cad las thun mong du gnas pa’i de kho na nyid do || de la
bla ma’i drin gyis mtshon par byas nas | rtag tu347 mnyam par
gzhag pa ni bsgrub pa’o || bya ba dang bya ba ma yin pa gang
yang rung ste | de’i grongs su348 ’gro ba ni mchod pa’o || de lta
bu’i rnal [189,5] ’byor gyi349 ngor mi rung ma ni gang yang med
pa’o || (b)zung ’dzin gyis rnam par rtog pa rgyun du byas pa’i
spang blang (l)dog ’jug rnams ni ’dir mi dgos pa’o || rtsal mthar
phyin pa ni grub pa’o ||
§1b: ’on kyang de la rnam par rtog pa’i lam dang rnam par rtog
pa dang bral ba’i lam mo || ji srid [189,6] du350 rnam par rtog pa
gtso bor gyur pa’i lam de >||< srid du ni yid kyi mos pa tsam las

344
We have conjecturally emended the reading de ltar byar to the phonetically identical
de ltar sbyar, which seems to be the smallest possible emendation yielding good sense.
The reading as it is transmitted in T contradicts what one would expect at the end of a lo
rgyus. In our understanding, something that is united becomes one, i.e. non-dual (Skt.
advaya), which seems to be a pun on the title of the AS.
345
Uncertain readings are given within parentheses. Shads to be disregarded are enclo-
sed in between angled brackets.
346
rdo rje em.] rdo rjes T
347
rtag tu corr.] rtag du T
348
grongs su T ] read grangs su?
349
rnal ’byor gyi em.] rnal ’byor gyis T
350
ji srid du corr.] ci srid du T

107
dngos su spyod pa ni ma gnang ngo || ji srid351 rnam par mi rtog
pa gtso bor gyur pa de nas ni >||< dngos su spyod pa yang gnang
ba yin no || bya ba dang mi bya ba dang dgos pa dang mi dgos
pa sogs kyi352 rnam [190,1] grangs ni gzhung nas bstan to || de ni
gNyis med grub pa’i don bsdud pa’o ||

Additional Passage in T (190,1–190,3) — Lo rgyus of the

*Sahajasiddhi
§2: de ltar rJe btsun dPal mo’i sku drin las ||
gNyis med grub pa’i legs bshad nyan zin gyis ||
da ni dPal ldan dBang po’i Blo353 de la ||
Lhan cig skyes grub nyan phyir la ’ur ’dod ||354
de la bla ma dGe slong gNyid [190,2] log pa’am || sa’i bdag po
dBang po’i Blo Indrabhūti chung ba dang | La ba’i na bza’ can
dang | Dur khrod pa rnams mtshan gyis rnam grangs yin la |
des mdzad pa’i Lhan cig skyes grub kyi gzhung dang | de’i ’grel
pa lCam dPal mo nyid kyis mdzad pa yod pa | de’i tshul [190,3]
yang ji lta bu355 yin snyam na | ’di ltar Indrabhodhis mdzad pa’i
Lhan cig skyes grub bzhugs sho ||

351
ji srid corr.] ci srid T
352
sogs kyi em.] sogs kyis T
353
dpal ldan dbang po’i blo em.] dpal ldan dpo’i blo T
354
The designation Venerable Goddess (rje btsun lha mo) refers to Lakṣmīṅkarā and
might be used interchangeably with lha mo dpal chen mo (Skt. *mahāśrīdevī) as found
in the colophon. The expression dbang po’i blo must, also in view of the following pas-
sage and the verse introducing the *Sahajasiddhi, be referring to Indrabhūti. A tentative
translation of the verse runs as follows: “In this way, having heard these beautiful verses
of the Advayasiddhi thanks to the kindness of the Glorious Lakṣmī, now, as regards
him, the glorious Indrabuddhi [i.e. Indrabhūti], there is the sudden wish to listen to [his]
*Sahajasiddhi.”.
355
ji lta bu corr.] ci lta bu T

108
Semi-critical Edition of the Canonical Tibetan
Translation
Sigla:
D “gNyis su med par grub pa’i sgrub thabs.” In: sDe dge
bsTan ’gyur, Vol. 51, 122–125 (Tōhoku 2220)
P “gNyis su med par grub pa’i sgrub thabs.” In: Peking
bsTan ‘gyur, Vol. 68, 64a–65b (Ōtani 3064)356
Ba Shendge’s reading of the Tibetan translation from 1964
Sa Sarnath edition (ed. Rinpoche and Dwivedi 1987)

rgya gar skad du || Advayasiddhi-sādhana nāma ||

bod skad du || gNyis su med par grub pa’i sgrub [P 64b] thabs
zhes bya ba |
’jam dpal gzhon nur gyur pa la phyag ’tshal lo ||

In Shendge, it is remarked that “I have used the Tib. text of Peking photographic edn.
356

in Tokyo 1957 and collated it with Sde dge edn. Rgyud, hgrel, wi: F 60b–62a. There are
no major deviances in the two translations. The translation in some places is literal and in
many others free. The translator’s aim seems to be not so much to give literal translation
but an idea of the total meaning.” (1964: 12–13).

109
mgon po’i rang bzhin gsal ba ni ||
khams gsum las byung thams cad mkhyen ||
rdo rje [D 61a] che la mgos phyag ’tshal ||
dgos pa’i don gyi ’bras bu ’byin ||1||
yul dus chos dang gza’ skar dang ||
dkyil ’khor brtag pa mi dgos pa’i ||
rdo rje sems dpa’i sgrub thabs bshad ||
bsdus te bsgrub pa357 sgrub pa’i mchog ||2||
dam bca’358 dka’ thub smyung ba dang ||
snying po bsgom pa359 smra ba la ||
’di nyid med par gyur na ni ||
bskal pa ’bum phrag brgyar mi ’gyur ||3||
che chung sa bon la sogs tshogs ||
sna yi nang nas ’byung ba dang ||
sngags pa rtag tu mchod pa byed ||
bdag nyid de nyid bsgoms te360 mchod ||4||
ma dang sring mo bu mo dang ||
de bzhin tsha mo nyid dang ni ||
shes rab361 thabs su bya ba ste ||
rnal ’byor pas ni rtag tu mchod ||5||
yan lag gcig med rigs ngan dang ||
bzo bo362 khyi sha can sogs kyi ||
bu mo rtag tu mchod par bya ||
ye shes rdo rje bsgom pa che ||6||
de kho na nyid kyi sngags kyi ’bras bu ni ’di yin no || oṁ āḥ
hūṁ |
’di ltar sems can ’ching ba ni ||
drag po’i las kyis363 sems can rnams ||
de ltar thabs dang ldan pa yis364 ||

357
bsgrub pa D Ba] sgrub pa P Sa
358
dam bca’ D Sa] dam bcas P Ba
359
bsgom pa P Sa] sgom pa D Ba
360
bsgoms te P Ba] bsgom ste D Sa
361
shes rab D P Ba] shes rabs Sa
362
bzo bo D Sa] gzo bo P Ba
363
las kyis D Sa] las kyi P Ba
364
de ltar thabs dang ldan pa yis Sa(sil.)] thabs kyis D; thabs kyi P Ba

110
’khor ba dag las ’grol bar byed365 ||7||
dam tshig nyin mtshan rtag tu ni ||
rigs lnga las byung bzang po dang ||
gsal ba’i tshogs kyis366 mchod pa byed ||
’o ma ’bab bas sna tshogs ’byung ||8||
sngags pa rab dga’i mig gis ni ||
rtag tu mig gi mdangs phyung ste ||
yang dag sems ni brtan po yis ||
ye shes rgya mtsho skam par byed ||9||
ji ltar g.yo dang mi g.yo ba ||
khams gsum thams cad ma lus pa ||
thams cad de nyid rnal ’byor ni ||
rdo rje ’dzin pa nyid du mthong ||10||
gzhan du kha cig367 smra ba dang ||
mtshan ma’i bye brag ’dug pa la ||
de la brnyas par mi bya ste ||
rdo rje sems dpa’i rdzu ’phrul nyid ||11||
thams cad mnyam pa nyid gnas te ||
ma lus dngos po med pa ni ||
sngags pas rtag tu sgom par byed ||
phung po’i rang bzhin dag pa nyid [P 65a] ||12||
dri dang gos dang me tog sogs ||
spos dang lha bshos de bzhin du ||
glu dang rol mo gar sogs kyis368 ||
thabs kyi bdag po dga’ bar ’gro ||13||
dka’ thub ngan pas gzir mi bya ||
smyung ba yis369 [D 61b] ni las mi bya ||
khrus dang gtsang sbra mi bya370 ste ||
blo gros chung ngu’i spyod mi bya ||14||
lugs ma la sogs phyag mi bya ||
sa shing rdo sogs de bzhin no ||

365
’grol bar byed Sa(sil.)] grol D P Ba
366
tshogs kyis P Ba] tshogs kyi D Sa
367
kha cig D P Ba ] kha gcig Sa(sil.)
368
sogs kyis D Sa] sogs kyi P Ba
369
smyung ba yis P Ba] smyung ba yi D Sa
370
mi bya D Sa] mi bya ba P Ba (hypermetrical)

111
bdag gi lus la rgyun du mchod ||
rtag tu sems gcig kho nas mchod ||15||
sbrang bu skyugs pa bsres pa yis371 ||
che chung tshogs kyis372 sbyang bar bya ||
gsal ba lnga dang ldan par byed ||
rdo rje ’dzin pa mchod par bya373 ||16||
bud med rang byung me tog ni ||
byang chub sems dang ldan par bya ||
rtag tu lha mo mchod par bya ||
phung po de nyid374 bsgom par bya375 ||17||
gzhan gyi nor dag ’phrog par376 bya ||
gzhan gyi bud med bsten par bya ||
brdzun gyi tshig377 kyang smra bar bya ||
sangs rgyas thams cad gsad par bya378 ||18||
rdo sogs mchod rten mi bya ste ||
glegs bam dag kyang bklag mi bya ||
rmi lam tsam du dkyil ’khor ni ||
lus kyis379 bya bar mi bya’o ||19||
thams cad dag ni kun la yang ||
rtog pa dag ni mi bya ste ||
rdo rje sems dpa’i bdag nyid du ||
dngos po ma lus gyur par bsam ||20||
sngags pa ’gro dang ’dug pa dang ||
de bzhin za dang mi za dang ||
’thung dang mi ’thung ba dag la ||
sems ni rtog par mi bya’o ||21||
rnam par snang mdzad byung ba la ||
sems can thams cad yan lag tu ||

371
skyugs pa bsres pa yis D Sa] bsre ba’i sgyugs pa yis P Ba
372
tshogs kyis D Sa] tshogs kyi P Ba
373
mchod par bya D Sa] mchod par byed P Ba
374
phung po de nyid P Ba] phung po’i de nyid D Sa
375
bsgom par bya D Ba Sa] bsgom par bya ba P (hypermetrical)
376
’phrog par D Sa] dbrog par P Ba
377
brdzun gyi tshig D Sa] rdzun gyi tshig P Ba
378
gsad par bya P Sa Ba] gsung bar bya D
379
lus kyis D Sa] lus kyi P Ba

112
mkhas pas sems can kun blangs te380 ||
dngos grub rgyur381 ni bza’ bar bya ||22||
thams cad rigs nas byung ba yi382 ||
bud med dag la brtag mi bya ||
de nyid bcom ldan pha rol phyin ||
kun rdzob tsam gyis383 gzugs nyid do ||23||
tshes grangs skar ma mi dgos te ||
smyung bar yang ni gnas mi bya ||
gnyis med ye shes ldan pa yi384 ||
bde bar gshegs pa’i dngos grub ’gyur ||24||
mang po smras te ci zhig bya ||
ji lta ji ltar385 mthong ba dag ||
de dag thams cad rnal ’byor gyis ||
[P 65a] mthong zhing rigs pas386 shes par bya ||25||
ba lang387 rta bong rnga mo dang ||
ba dang khyi las byung ba las ||
gsal bar che ba bsres byas te ||
rnal ’byor mkhyen pas rtag tu bza’ ||26||
gcig la chags par mi bya ste ||
gang yang blang dor mi bya’o ||
rnal ’byor mnyam pa’i sems kyis ni ||
’khor ba’i rgya [D 62a] mtsho rgal par byed ||27||
’byung dang ’dug dang ’gro ba dang ||
kun gyis mang po spros pa dang ||
de dag med na ’khor ba med ||
skye ba dang ni ’byung ba med ||28||
bcom ldan rdo rje nyi mas ni ||
shes rab mtshan mos bshad pa ste ||
de ’dra’i rnal ’byor bsgom byed na388 ||

380
blangs te D P Sa] blangs ste Ba
381
rgyur D P Sa] gyur Ba
382
byung ba yi D, P, Sa] byung ba yis Ba
383
tsam gyis P Ba] tsam gyi D Sa
384
ldan pa yi P] ldan pa yis D Sa; ldan pa yin Ba
385
ji ltar D, P, Sa] ji lta Ba
386
rigs pas em.] rig pa 𝝨
387
ba lang P Sa] ba long D Ba
388
bsgom byed na P Ba Sa] sgom me byed na D (hypermetrical)

113
myur du dngos grub thob par byed ||29||
thams cad sems kyi389 chos nyid ni ||
dngos por ma grub brtag dka’ ba ||

de nyid slob dpon zhal nas ni ||

thob par ’gyur bar the tshom med ||30||
mi gnas mya ngan ’das pa yin ||
mtshan ma med pa gnas ni390 med ||
† (…) †391 ||
de shes byang chub mchog tu ’gro392 ||31||
de ltar shes pa’i rnal ’byor ni ||
blo can rtag tu bslab par bya393 ||
de nyid ’grub par the tshom med ||
bsod nams med pa’ang ’grub par ’gyur ||32||
khams gsum na ni slob dpon las ||
che bar gyur pa gzhan med na ||
de nyid kyi ni394 bka’ drin gyis ||
dngos grub395 mang po thob par ’gyur ||33||
de nyid rdo rje sems dpa’ nyid ||
sangs rgyas thams cad phyag kyang ’tshal ||
slob dpon mchod pa bla na med ||
ci tsam nus par mchod par bya ||34||
de ni396 de bzhin gshegs pa nyid ||
sems can rnams kyi don gyi phyir ||
gzugs su snang ba kun rdzob ste ||
rnal ’byor bdag nyid gdan la ’dug397 ||35||
’chi bdag de ni rtog pa398 che ||

389
sems kyi P Ba] sems kyis D Sa
390
gnas ni D P Sa] gnad ni Ba
391
†(…)† D P Ba] sems can kun la khyab pa yi Sa(conj.) (rgya dper)
392
de shes byang chub mchog tu ’gro D P Ba] mngon par byang chub dam pa mchog
Sa(conj.) (rgya dper)
393
bslab par bya D Sa] brlab par bya P Ba
394
de nyid kyi ni D Sa] de nyid kyis ni P Ba
395
dngos grub D P Sa] dngob grub Ba
396
de ni D P Ba] de nyid Sa
397
’dug D Sa] bzhugs P Ba
398
rtog pa Sa(em.)] rtag pa D P Ba

114
thams cad ma lus stong pa nyid ||
rang gi rtog pas399 mi snang la ||
gzhan dag de la dga’ bar byed ||36||

gNyis su med par grub pa’i sgrub thabs zhes bya ba | ’di ni U
rgyan gyi che ba rnal ’byor gdan gyi400 sgrub pa’i lha’i gnas nas
byung ba bcom ldan ’das ma Lakṣmī’i zhal snga nas byung ba
rdzogs so ||

399
rtog pas P Ba] rtog pa D Sa
400
gdan gyi P Ba Sa] gdan gyis D

115
Translating
the Karmapas’ works