Yoga and Religion∗

James Mallinson
jim@khecari.com

May ,

e growth of yoga in the west is relentless. e latest survey by the

American magazine Yoga Journal estimates the number of its practitioners
in America at . million and the value of the “yoga industry” there at $ .
billion. With so much at stake it is not surprising that disputes about the
true nature of yoga — and thus its ownership — are flaring up with increas-
ing regularity. urrent controversies include the alifornia ncinitas case,
in which hristian groups are fighting against the introduction of yoga into
school curricula, and the Hindu American oundation’s “Take ack Yoga”
campaign, in which it is asserted that the Hindu roots of yoga have been
airbrushed out of its modern manifestations. ven the White House has
weighed in, clearly stating its opinion that “Yoga has become a universal
language of spiritual exercise in the United States, crossing many lines of
religion and cultures”. e opinions of many of the groups involved in
these controversies have been reasonably well thought out, at least in com-
parison to some more outlandish accusations put forward by fringe groups
who see yoga as, for example, the work of Satan. Whatever the arguments
put forward, at their root are the questions of whether yoga is Hindu and
who is entitled to practise it.
I am not going to address these controversies directly, but rather I shall
try to answer the underlying questions in the context of my specialist area,
traditional hathayoga,
. and in particular hathayoga
. during its formative pe-
riod in the th- th centuries. Hathayoga
. is the variety of yoga practice
∗ is is the revised text of a lecture given at a seminar on Modern Yoga organised by the UK Hindu hristian

oundation and held at Heythrop ollege, London on March th .

http://www.yogajournal.com/press/press release/ accessed April th .
http://m.economictimes.com/news/news-by-industry/et-cetera/barack-obama-white-house-embraces-yoga-
amid-conservative-contortions/articleshow/ .cms accessed April th .
J : .
which places greatest emphasis on physical methods and from which mod-
ern yoga is, at least in part, derived. urthermore, as we shall see below,
hathayoga
. came to be accepted within the orthodox Indian tradition as a
key part of yoga more broadly defined. In the process of answering the
questions of whether hathayoga
. in its formative period was Hindu and who
was entitled to practise it I shall also briefly address another claim about yoga
which has recently had an airing thanks to high-profile sex scandals involv-
ing prominent yoga teachers, namely that such scandals are unsurprising
since yoga was originally a sex cult.
Scholars and commentators, such as the increasing number of yoga blog-
gers, have had difficulty addressing these issues because, perhaps surpris-
ingly, the history of hathayoga
. has been very poorly studied. Most pro-
nouncements on the subject are based on three Sanskrit texts which were
arbitarily selected in the late th century and have since been held to be
the hathayoga
. canon. ese texts, the Śivasamhitā, . Hathapradīpikā
. and
Gheran. dasa
. mhitā,
. provide only a limited view of the tradition. Over the last
two decades I and a handful of other scholars have been working on other
texts produced during hathayoga’s
. formative period and we are beginning to
get a better idea of the bigger picture. I shall now give a brief summary of
what I infer from that corpus to be the early history of hathayoga.
.
lose study of the teachings on physical yoga practice in the early hatha .
corpus shows it to be a combination of two yogic methods. e first can
be traced back to before the common era, in descriptions of the practices
of ascetics found in the two great Sanskrit epics, the Mahābhārata and
Rāmāyana, . and in the Pali uddhist canon (the uddha says that he tried
these techniques but abandoned them on realising that they were of no use).
We also have reports from the entourage of Alexander the Great, who en-
countered a group of ascetics at Taxila, now in Pakistani Punjab, at the end
of the fourth century . e ascetics of this period would hold difficult
postures, sometimes for years on end, practise extended breath retention and
use specific physical methods that were forerunners of techniques which ap-
pear in later hathayogic
. works as mudrās. eir aims were to still the mind
and to develop a store of ascetic energy within their bodies. is energy,
called tapas, could be used to win boons such as supernatural powers from
the gods, and to give blessings and curses, while the stilling of the mind
was associated with liberation from the wheel of rebirth. Tapas was closely
See e.g. William road’s / / New York Times article on Yoga and Sex (http://tinyurl.com/ ua yu
accessed April th ). I shall address this question in much greater depth in a paper entitled “Yoga and
Sex: What is the Purpose of Vajrolīmudrā” to be delivered at the University of Vienna’s Yoga in Transformation:
Historical and Contemporary Perspectives on a Global Phenomenon conference on September th .
connected with celibacy, and some of these ascetic methods were directly
aimed at ensuring the practitioner’s celibacy and enhancing its benefits.
e other tradition of yoga practice, which combined with this older
ascetic tradition to produce the physical yoga systems formalised between
and , was developed within the tantric schools that flourished
in India during the first millennium. or various reasons tantra has incor-
rectly come to be seen in the west as primarily about sex magic and other
antinomian esoteric practices, but these were always a relatively minor and
rarely practised part of its teachings which in fact ran the whole gamut of
what we would call religion. uring the th- th centuries there devel-
oped within certain tantric traditions various techniques of visualisation of
a feminine energy known as Kun. dalinī,
. the coiled goddess, rising up from
the base of the spine through a series of cakras up to union with a male deity
in the head.
is visualisation-based tantric yoga was combined with the physical
techniques of the older ascetic tradition to produce hathayoga.
. We see this
process happening in a somewhat clumsy fashion in earlier texts of the cor-
pus, such as the Vivekamārtan. da.
. Later works, in particular the Śivasamhitā,
.
do a more coherent job. e best known of the hatha . texts, the Ha thapra-
.
dīpikā, which is for the most part a compilation, is among the less coher-
ent works of the corpus. In his attempt at inclusivity, the text’s compiler,
Svātmārāma, includes a wide range of sometimes contradictory teachings.
A good example of this is found in one of the key techniques of hathayoga,
.
khecarīmudrā, in which the tongue is turned back and upwards into the cav-
ity above the palate. In its older ascetic manifestation it is associated with
stilling the mind and blocking the fall of bindu, a substance produced in the
head, which, if not sealed there by khecarīmudrā, drips down to be burnt up
in the fire in the stomach or shed as semen. e loss of bindu causes old age
and death its preservation stops the yogi from ageing and dying. Various
other yogic methods are also geared towards the preservation of bindu. e
headstand, for example, uses gravity to keep bindu in the head. In tantric
conceptions of the subtle body, on the other hand, the head is home to a
store not of bindu but of amrta,
. the nectar of immortality. e tantric khe-
carīmudrā causes Kun. dalinī
. to reach this store of nectar and then, rather
than keep it where it is, flood the body with it. e Hathapradīpikā
. teaches
both these khecarīmudrās.
ese references are detailed in my draft article on “Śāktism and Hat.hayoga”, which can be accessed via my
page on academia.edu.
e ultra-inclusivist Hat.hapradīpikā also includes teachings from non-hat.ha texts, such as the Amanaska,
which is positively scornful of hat.ha physical methods. As indicated by its title, it teaches a yoga in which the
 So who were the practitioners of these two types of yoga As I have
already noted, the originators of the older tradition were celibate ascetics.
eir tradition flourishes to this day pictures of them bathing at the tri-
ennial Kumbh Mela festival have just been beamed around the world. We
can trace them back via travellers’ reports and Mughal painting in the late
medieval period, through mentions in texts in the hathayoga
. corpus back to
the naked “gymnosophists” or śramanas . encountered by Alexander and the
uddha.
e tantric tradition is also alive today, although it is likely to have
changed since its inception rather more than the ascetic tradition. It is now
best represented by an order of celibate yogis known as the Nāths. e
Nāths’ roots lie in first-millennium tantric sects in which sexual rites were
practised, but by the time their hathayogic
. texts were compiled they had
turned their back on the so-called “left-hand” practices of tantra and be-
come celibate ascetics, as evinced by the final verse of one of their earliest
texts, the circa thirteenth-century Goraksaśataka:
.
We drink the dripping liquid called bindu, “the drop”, not wine
we eat the rejection of the objects of the five senses, not meat
we do not embrace a sweetheart [but] the Susumnā. nādī,
. her
body curved like kuśa grass if we have intercourse †. . . † it
takes place in a mind dissolved in the void, not in a vagina.
e original practitioners of hathayoga
. were thus all celibate ascetics it
was no sex cult. eir teachings would have been passed on orally, from
guru to disciple. e older tradition is more than years old, but all
descriptions of its practices are from outsiders until the composition of the
hatha
. corpus starting in the th or th centuries. In fact, of all this tra-
dition’s various ascetic practices, it is only the techniques of yoga that are
taught in Sanskrit texts. To this day there are no textual instructions on their
various methods of cultivating tapas, such as holding one’s arm in the air for
years on end or sitting surrounded by smouldering fires in the summer sun.
Teachings on such matters are still only passed on orally. e purpose of
the ascetic tradition’s hathayoga
. texts was to bring to a wider audience those
of their teachings on yoga which were suited to householders.
Similarly, the tantric traditions of the first millennium consisted of a
variety of exclusive sects, access to which was through initiation from a guru
and whose practices, in addition to sect-specific visualisations, included the
repetition of secret mantras exclusive to each tradition and the use of secret
functions of the mind are to be stopped, but, unlike those of hat.hayoga, its techniques are purely mental.
ritual diagrams called man. dalas.
. ut in the texts produced by the tantric
hathayoga
. tradition in the first few centuries of the second millennium 
these exclusive features are removed. e teachings are made open to all.
In the Dattātreyayogaśāstra, the earliest text to teach a hathayoga
. named
as such, and which is in the tradition of the naked ascetics encountered by
Alexander and found to this day, we read the following ( a- b):

“Whether a rahmin, an ascetic, a uddhist, a Jain, a Skull-

earer or a materialist, the wise one who is endowed with
faith and constantly devoted to the practice of [hatha]
. yoga
will attain complete success. Success arises for one who is de-
voted to the practice how can it arise for one who does not
practice”
Although some of the religious categories said in these verses to be able to
practise yoga are now somewhat obscure, they represent a broad swathe of
the religious traditions of India in the thirteenth century, when the text was
composed. urthermore, the inclusion of the word “rahmin” in addition
to ascetic (śramana). indicates that the teachings are for both ascetics and
householders, worldly people who have not completely devoted their lives
to religious ends, and we find similar statements in other texts of the early
hatha
. corpus.
e “materialist”, or ārvāka, mentioned in this verse is particularly in-
teresting. While little is known about this tradition — we have none of
their texts, only allusions, usually negative, in the works of others — they
correspond to what we might today call atheists. ey believed that there is
no life after death and no soul animating the material body. Yet yoga was
apparently for them too.
Sanskrit texts are written from a male (and generally rahmin) perspec-
tive and so there is little in texts of yoga that directly applies to women.
Yet we do get some fleeting references which make it clear that there were
women practising yoga. In addition, the text just quoted as saying that
people of all religious traditions can achieve success through yoga, adds that
this is true for the young, old and infirm.
e slightly later Hathapradīpikā,
. which is the most inclusivist of all the
hatha
. texts, borrows this verse about the young, old and infirm, and goes
.g. Amr. tasiddhi . , Śivasamhitā
. . .
.g. Dattātreyayogaśāstra a, Hat.hapradīpikā . - .
Dattātreyayogaśāstra .
Hat.hapradīpikā . .
out of its way not to include any teachings which might identify it exclu-
sively with any particular tradition. us there are no mentions of mantras,
man. dalas
. or initiations, nor even of cakras. e now almost universally ac-
cepted system of six-plus-one cakras was yet to achieve hegemony and any
mention of a particular cakra system would have betrayed allegiance to its
particular tantric tradition so cakras were omitted altogether.
e universalism of hathayoga
. meant that it could be adopted by any
religious tradition and adapted to its ends — which sometimes involved
denying its universality. us its techniques are found in Jain and uddhist
texts, while the Śivasamhitā,
. a late text of the early hathayoga
. corpus, teaches
it within a specific tantric tradition, complete with exclusive secret mantras.
Meanwhile, yoga was catching the attention of the increasing number of
Muslims in India. At first this came through interaction between Sufis and
ascetic yogis, but later on texts on yoga in a variety of non-Sanskritic lan-
guages, including Arabic, Persian, Turkish, and Urdu — such as the Persian
Bahr. al-Hayāt
. or Water of Life, an illustrated manuscript of which from
includes the earliest paintings of non-seated yogic āsanas — were
commissioned by nobles at Islamic courts, both in India and the wider Is-
lamic world, and continue to be used among some Sufi orders to this day.
So, it is clear that in hathayoga’s
. formative period its codifiers saw it as
something that could be practised by anybody. ut was it Hindu efore I
answer this question I should point out that the word Hindu is itself prob-
lematic. It was first used by outsiders to describe the inhabitants of the
region around the Indus river, so rather than a religious marker it was a ge-
ographic or ethnic term, and the same is true of its first occurrences in Indic
languages in the fourteenth century. Its use to denote religious affiliation
is not found until several centuries later. e argument over the origins of
Hinduism, the religion of the Hindus, has two extreme positions. One side
says that Hinduism is the sanātana dharma or “perennial religion” which has
existed since time immemorial the other says that it was invented by the
ritish as they sought to understand India in their own terms of reference.
e truth lies somewhere between the two. ertainly in the th-century
yoga text that I cited earlier, there is no notion of an overarching Hinduism
in opposition to uddhism and Jainism. Indian religion has always com-
prised a huge range of differing worldviews and despite modern claims to
the contrary there has never been a universally accepted monolithic Hin-
e manuscript is in the hester eatty Library, ublin. Ten of its twenty-three pictures of āsanas are to be
displayed in the Yoga: e Art of Transformation exhibition to be held at the Sackler Gallery of the Smithsonian
Institute, Washington , from October .
On the term Hindu and the concept of Hinduism, a useful starting point is L  .
duism. When such a fiction is propounded it is usually aligned with the
Vedic and rahmanical traditions. If we look at the early history of yoga,
however, we can see that it appeared outside of these traditions.
e word yoga in the sense of a soteriology, a means to salvation, first
appeared in rahmanical texts in the last centuries to describe the prac-
tices of the non-rahmin ascetics amongst whom the techniques of hatha- .
yoga were developed (who were themselves yet to call their practices yoga).
eir yoga is one of a variety of features now seen to characterise Hin-
duism but which originated outside of the Vedic and rahmanical tradi-
tions among the ascetics of the region known as Greater Magadha to the
east of Allahabad in northern India. Prominent among these features are
vegetarianism and institutionalised renunciation. ut perhaps those most
closely identified with Hinduism are the concepts of karma and rebirth,
which are not found within Vedic texts but were adopted fairly early on in
the development of rahmanical religion.
e best known and perhaps earliest indicator of yoga’s adoption by the
rahmanical orthodoxy is Patañjali’s circa fourth-century Yogasūtra, which
teaches a soteriological method based on stilling the mind through medita-
tion and breath control but none of the more physical techniques of hatha- .
yoga. Like the soteriological methods called yoga in earlier Sanskrit texts,
the teachings of the Yogasūtra originated in non-rahmanical traditions, in-
cluding uddhism. ven with the composition of the Yogasūtra, yoga was
still not accepted by the most orthodox rahmanical traditions, nor did its
proponents seek it to be. Yoga as philosophy and its associated metaphysi-
cal system known as Sāmkhya
. were said in the circa fifth-century Sāmkhya-
.
kārikās to be opposed to rahmanical Vedic religion because of the latter’s
practice of animal sacrifice, and in his Tantravārttika the sixth-century
orthodox Vedic ritualist Kumārila groups Yoga and Sāmkhya . with other
heresies such as uddhism and Jainism.
e proponents of rahmanical Hinduism’s perennial tradition often
base it on the six supposedly classical and mutually reinforcing darśanas or
philosophies, one of which is Yoga. ut Yoga does not appear in lists of the
 . f. John Henry G ( : ), an eighteenth-century traveller to India, who wrote
of the posture-practising ascetics encountered by Alexander’s entourage “ese Gymnosophists were undoubtedly
not ramins [sic], as has been erroneously advanced by many authors, but of that sect of men now called Gioghys.”
 .
N : .
I write Yoga with a capital ‘Y’ when referring to yoga as a darśana or philosophy.
e other heresies listed are Pāśupata and Pāñcarātra. See  : - for a translation and
discussion.
six darśanas until the twelfth-century Sarvasiddhāntasamgraha
. and then it
is only as a metaphysical system and meditational method — the physical
techniques of hathayoga
. were still shunned by the rahmanical orthodox
as they had been for over a millennium. In the Bhagavadgītā, which can be
dated to the first few centuries of the common era, Kr. s. na
. dismisses the tough
physical practices of the śramana . ascetics because they are not enjoined in
scripture and their purpose is to show off. It is only with the composition
of the Sanskrit texts of the early hathayoga
. corpus that the physical practices
of hathayoga
. started to enter the mainstream.
After hathayoga’s
. teachings had been universalised in the th- th cen-
turies, its practices slowly became a key part of orthodox formulations of
yoga, which was itself on its way to becoming, for the intellectual elite
at least, the dominant Indian soteriological method (amongst the popu-
lation as a whole it came to be superseded by bhakti, devotion). In the
centuries following its formative period the techniques of hathayoga
. were
woven into the previously meditation-based yogic methods that had already
been adopted by the mainstream orthodox rahmin tradition, to the extent
that long passages — sometimes entire texts — from the early hatha . corpus
were used to compile the so-called “Yoga Upanisads”
. in the th and th
centuries. ontrary to the modern distinction, propounded by Swami
Vivekananda and others, between a mental rājayoga and an inferior phys-
ical hathayoga,
. in medieval India hathayoga
. was often seen as integral to
the classical yoga of Patañjali. Yogic methods were also integrated with
Vedantic philosophy, whose Advaita or “non-dual” tradition was, mirroring
the trajectory of yoga, well on its way to becoming the dominant philoso-
phy of the emergent “Hinduism”. is combination of Advaita and Yoga
was then posited as a cornerstone of the sanātana dharma which the most
vocal proponents of Hinduism today claim to be its eternal essence.
In sum, although yoga and its hatha . or difficult physical methods did
become an accepted part of mainstream Hinduism by the late medieval pe-
riod, it was not always so and to claim that they are key to some perennial
Hindu tradition is, at best, to show a poor grasp of India’s religious history.
Perhaps more pertinent to modern practitioners of yoga than the ques-
tion of whether or not yoga is Hindu — particularly in the light of recent
research which has shown that modern yoga’s techniques are drawn from a
H : - .
Bhagavadgītā ( . )
  .
See e.g. Vidyāranya’s
. th-century Dīpikā on the Aparoks.ānubhūti (ad  he expresses a similar sentiment
in his Jīvanmuktiviveka, vv. . . - ) and the ninth chapter of the Hamsavilāsa
. discussed in V .
wide range of traditions, many of which are clearly not Hindu at all —
is the question of whether any of its practices are grounded in a system of
metaphysics which might preclude adherence to the tenets of other religious
systems. At first sight I would say yes: the texts assume a belief in the princi-
ples of reincarnation and its corollary, liberation from the wheel of rebirth,
with the practice of yoga sometimes said to be the fruit of good deeds done
in past lives and its aim to be liberation. ut, as I mentioned earlier, the first
text to teach hathayoga
. says that it will work even for atheists, who we know
did not believe in karma and rebirth. And the Śāradātilaka, a contempora-
neous tantric work which includes some hathayogic
. teachings, opens its final
chapter by listing four conflicting definitions of the metaphysics of yoga.
It privileges none of them, but simply launches into its teachings on yoga
practice, the implication being that whatever your metaphysical aim, yoga
will get you there.

Works Cited
Primary Sources
Amanaska, ed. Jason irch. ritical edition as part of thesis submitted for
doctorate of philosophy at the University of Oxford, .
Amrtasiddhi
. of Virūpāksanātha.
. Maharaja Mansingh Pustak Prakash, Jodh-
pur, Acc. Nos. and .
Goraksaśataka.
. Government Oriental Manuscripts Library, Madras, MS
No. R .
Gheran. dasa
. mhitā,
. ed. and tr. J. Mallinson. New York: YogaVidya.com.
.
Dattātreyayogaśāstra. Unpublished critical edition by James Mallinson based
on Dattātreyayogaśāstra, ed. rahmamitra Avasthī, Svāmī Keśavānanda Yoga
Samsthāna
.  Man Singh Pustak Prakash Nos.  Wai Prajñā Pāthaśālā
.
/ - ,  aroda Oriental Institute  Mysore Government Ori-
ental Manuscripts Library  anjavur Palace Library  .
Vivekamārtan. da. of Goraksadeva.
. Oriental Institute of aroda Library. Acc.
No. .
S .
or details see my draft article “Hat.hayoga and Philosophy” which can be found on my academia.edu page.
On this text, which is different from the better known text of the same name, see M , in which
it is analysed and translated.
is edition was read with Professor Alexis Sanderson, Jason irch, r Péter-ániel Szántó and r Andrea
Acri in Oxford in early , all of whom I thank for their valuable emendations and suggestions.
Śāradātilakam of Laksma
. nadeśikendra.
. Patala
. as edited in 
.
Śivasamhitā,
. ed. and tr. J. Mallinson. New York: YogaVidya.com. .
Hathapradīpikā
. of Svātmārāma, ed. Svāmī igambarjī and r Pītambar Jhā.
Lonavla: Kaivalyadhām S.M.Y.M. Samiti. .

Secondary Sources
 , hristian. . Les Nātha-Yogin et les Upanisads.
. Paris: iffusion
de occard.
 , Johannes. . Greater Magadha: Studies in the Culture of
Early India. (Handbook of Oriental Studies, Section Two, India, Vol. .)
Leiden and oston: rill.
------------------------. . “e rāhmanical
. ontribution to Yoga”, pp.
- in International Journal of Hindu Studies , .
 , Gudrun. . “e Saradatilakatantra on yoga: A new
edition and translation of chapter ”, pp. - in Bulletin of SOAS, ,
.
 , Vincent. . “Apocalypticism, Heresy and Philosophy,”
pp. - in World View and eory in Indian Philosophy, ed. Piotr al-
cerowicz. elhi: Munshiram Manoharlal.
G , John Henry. . A voyage to the East Indies. London : S. Hooper.
H , Wilhelm. . India and Europe. Albany: State University of
New York Press.
J , Andrea. . “e Malleability of Yoga: A Response to hristian
and Hindu Opponents of the Popularization of Yoga”, pp. - in Journal
of Hindu-Christian Studies.
L  , avid N. . “Who Invented Hinduism”, pp. - in
Comparative Studies in Society and History, Volume , Issue .
M , James. . “e Original Goraksaśataka,”
. pp. – in
.G. White, ed., Yoga in Practice. Princeton: Princeton University Press.
While considering the text of the first four upadeśas of the HP as established in this edition authoritative, I
do not accept the editors’ inclusion of a fifth upadeśa within the text. e fifth upadeśa is found in very few HP
manuscripts, verses from it are not found in citations of the HP and its subject matter, yogic methods of curing
diseases, is very different from that of the rest of the text.
-----------------. . “Śāktism and Hathayoga”.
. Paper presented at a con-
ference on the Śākta traditions in September and to be published in a
volume of the conference proceedings by Routledge. A draft can be down-
loaded from my academia.edu page.
N , Andrew. . “oxography and oundary-ormation in
Late Medieval India”, pp. - in World View and eory in Indian Phi-
losophy, ed. Piotr alcerowicz. Warsaw Indological Series Volume . elhi:
Manohar.
S , Mark. . e Yoga Body. New York: Oxford University
Press.