The Yogatārāvalī and the Hidden History of Yoga

T HOSE WHO HAVE ATTENDED AN

Ašţāńgayoga class in the tradition of Pattabhi Jois are probably familiar with at least one verse of the Yogatārāvalī, whether they know the name of the text or not. Pattabhi Jois would chant the first verse of the Yogatārāvalī at the beginning of a morning class, and many of his students continue to do this before their yoga practice:

vande gurūņāµ caraņāravinde sandarśitasvātmasukhāvabodhe | niģśreyase jāńgalikāyamāne saµsārahālāhalamohaśāntyai ||1|| "I pay homage to the gurus' lotus feet, which have revealed the knowledge of the bliss of one's own self. Unsurpassed, [these lotus-feet] act like toxicologists 2 for curing the delusion that is the poison [known as] transmigration."

The name Yogatārāvalī can be understood as: 'A String of Stars on Yoga.' Each verse of this short text is likened to a shining star, which illuminates the topic of yoga. Its first verse acknowledges the importance of the teacher, who transmits to the student the liberating knowledge that extinguishes the suffering of worldly life.

In the Yogatārāvalī, the means to liberation is the practice of Haţha and Rājayoga. Its Haţhayoga consists of physical techniques; in particular, the three locks (mūla, uďďīyana 3 and jālandhara). These locks are applied during breath retentions (kumbhaka) and, as a result of this practice, the Haţhayogin is able to immerse the mind in an internal resonance (nādānusandhāna).

In contrast to this, Rājayoga is simply the practice of samādhi, a profound state of meditation in which there is no mental activity, no breathing and no positional consciousness. That is to say, Rājayoga is something beyond the common experience of "I" and "mine". The Yogatārāvalī puts it this way:

ahaµmamatvādi vihāya sarvaµ śrīrājayoge sthiramānasānām | na drašţŗtā nāsti ca dŗśyabhāvaģ sā jŗmbhate kevalasaµvid eva ||16|| "Having left behind everything beginning with the states of 'I' and 'mine', those whose minds are steady in the sacred [state of ] Rājayoga are neither observers nor objects of observation. Only an isolated awareness prevails."

The main difference between Haţha and Rājayoga is that the latter is beyond all techniques. In other words, once the Rājayogin has achieved samādhi, other 1 I would like to thank Eddie Stern for encouraging me to write this article and for his valuable comments on it. Thanks also to Jacqueline Hargreaves, James Mallinson, Mark Singleton and Elizabeth De Michelis for their many helpful comments on early drafts of this article. The illustration of Śāmbhavī Mudrā and figures 1 and 2 are by Jacqueline Hargreaves. Figures 1 and 2 contain images by Febrian Anugrah. 2 The term jāńgalika is usually spelt jāńgulika in Sanskrit texts and it is often translated as 'snake doctor.' For example, in Sures Chandra Banerji's "A Companion to Sanskrit Literature" (1989, p. 427), jāńgulika is defined as a "snake doctor; dealer in antidotes of poison." Also, Monier-Williams dictionary defines jāńgulika as a "snake charmer." However, in the Yogatārāvalī's first verse, jāńgalika is being used metaphorically to describe gurus who can cure a poison called Hālāhala. Rather than a snake poison, Hālāhala probably refers to either the mythological poison produced at the churning of the ocean and swallowed by Śiva (thereby causing the blueness of his neck) or some plant-based poison. Therefore, it appears that the guru is being likened to a doctor who specializes in the general treatment of poisons (i.e., a toxicologist) rather than a snake doctor or charmer. This is somewhat supported by a Sanskrit commentary on the Yogatārāvalī called the Yogabhāvaprakāśikā, which glosses jāńgalika as a višavaidya (i.e., 'a poison specialist' or 'toxicologist'), and so does the Amarakośa (1.8.510). 3 There are two different spellings of this bandha in the Yogatārāvalī: uďďīyana and uďyāņa. The difference is due to the metre of the verses. In manuscripts of the Haţhapradīpikā and Dattātreyayogaśāstra, this bandha is usually spelt uďďīyāna or uďďiyāna. It is likely that Haţha and Rājayoga began as separate traditions. Much of the Yogatārāvalī's teachings on Rājayoga appear to derive from those of the earliest extant Rājayoga text called the Amanaska, which can be dated to the twelfth century. The Amanaska teaches a system of yoga called Rājayoga, which it says is 'the king (rāja) of all yogas.' It omits the first seven auxiliaries of ašţāńgayoga. In fact, the author considers these auxiliaries and the techniques of Haţhayoga to be not just ineffectual practices, but hindrances on the path to attaining samādhi.

In effect, the Amanaska says that there is little point in trying to master difficult postures, breathing exercises and meditation techniques when the goal of these techniques, samādhi, otherwise known as the no-mind state (amanaska), can be achieved easily by Śāmbhavī Mudrā.

The yogin who is practising this Mudrā is described as gazing outwards (bahirdŗšţi) with eyes half open, half closed, while directing the mind inwards to an internal focal point (antarlakšya).

The Yogatārāvalī teaches this technique and calls it Amanaskamudrā. This name and other technical vocabulary common to both texts suggest that the Amanaska influenced the author of the Yogatārāvalī. 4 For this reason, the Yogatārāvalī was probably composed sometime after the twelfth century.

Haţha and Rājayoga were combined in other medieval yoga texts that were probably written only one or two centuries after the Amanaska. Examples of such texts include the Dattātreyayogaśāstra (12-13 th c.) and the Amaraughaprabodha (14 th c.). These texts combined Haţha and Rājayoga in a hierarchical relationship, along with Mantra and Layayoga. The idea behind this fourfold scheme was that the dullest students were taught Mantrayoga, the most gifted, Rājayoga, and those in between, Laya and Haţhayoga. In this system, Mantra, Laya and Haţhayoga were auxiliary practices aimed solely at the achievement of samādhi (i.e., Rājayoga). However, presumably, one could dispense with Haţhayoga altogether and achieve samādhi by practising Mantra or Layayoga. Furthermore, it appears that neither Mantra, Laya nor Haţhayoga were of any use to those gifted students who could readily achieve samādhi. This fourfold system is depicted in figure 1.

Figure 1

The fourfold system of yoga in the Dattātreyayogaśāstra and the Amaraughaprabodha.

The Yogatārāvalī changed this fourfold hierarchical relationship in a subtle yet very significant way. It omitted Mantrayoga and integrated Layayoga with Haţha, thereby creating a system in which Haţhayoga was the sole means to Rājayoga for all practitioners, regardless of their capabilities. In doing so, the physical practices of Haţhayoga became indispensable for achieving samādhi in this system. This new relationship is seen in figure 2.

Figure 2

The fifteenth-century Haţhapradīpikā adopted a scheme similar to the twofold system. 5 The author of this text stated explicitly the mutual dependence of Haţha and Rājayoga as follows:

haţhaµ vinā rājayogo rājayogaµ vinā haţhaģ | na sidhyati tato yugmam ā nišpatteģ samabhyaset ||2.76||

"Without

Haţhayoga, Rāja is not accomplished, and without Rājayoga, nor is Haţha. Therefore, the yogin should practice both until the final stage [of yoga is attained]."

The Yogatārāvalī describes a process whereby the three locks (bandha) are practised during deliberate breath retentions (kumbhaka) to bring about a special type of breath retention called Kevala Kumbhaka. Unlike deliberate breath retentions that are performed by holding the breath according to one's capacity (yathāśakti), Kevala Kumbhaka is a spontaneous cessation of the flow of the breath, along with all sensory and mental activity. In the Yogatārāvalī, Kevala Kumbhaka is the central mechanism that connects Haţha and Rājayoga. This explains why nearly one fifth of its verses are devoted to describing this special type of breath retention.

The similarities between the Yogatārāvalī and the Haţhapradīpikā seem extensive enough to suggest that one influenced the other. In fact, both texts share a verse on immersing the mind in the internal resonance ( Seeing that the author of the Haţhapradīpikā compiled his work from earlier yoga texts, it is likely that he borrowed and slightly modified the above verse from the Yogatārāvalī. If this is the case, the Yogatārāvalī must have been written before the fifteenth century. The Haţhapradīpikā's borrowing from the Yogatārāvalī is further suggested by the similar teachings of these two texts on Kevala Kumbhaka, immersing the mind in the internal resonance (nādānusandhāna) and Śāmbhavī Mudrā.

Reading the Yogatārāvalī in conjunction with yoga texts of the same era sheds much light on its terminology and content. Apart from what I have mentioned above, another striking example is the usage of various words for samādhi in the Yogatārāvalī. The terms manonmanī, unmanī, sahajāmanaska, turīya and yoganidrā appear to have the same meaning as rājayoga and samādhi. This is also the case in the Amanaska and the Haţhapradīpikā, the latter of which explicitly states that these terms are to be understood as synonyms (ekavācaka): rājayogaģ samādhiś ca unmanī ca manonmanī | amaratvaµ layas tattvaµ śūnyāśūnyaµ paraµ padam || amanaskaµ tathādvaitaµ nirālambaµ nirañjanam | jīvanmuktiś ca sahajā turyā cety ekavācakāģ ||4.3-4||

" Rājayoga, samādhi, unmanī, manonmanī, amaratva, laya, tattva, śūnyāśunya, para-pada, amanaska, advaita, nirālamba, nirañjana, jīvanmukti, sahajā and turyā [all] say the same thing." 6 Cf. Haţhapradīpikā 4.66 (śrīādināthena sapādakoţilayaprakārāģ kathitā jayanti | nādānusandhānakam ekam eva manyāmahe mukhyatamaµ layānām). 7 The meaning of avadhānāni as literally 'concentrations' can be found in Abhinavagupta's Tantrāloka (2.12). Jayaratha's commentary glosses avadhāna as, "one-pointedness on a particular object of attention" (pratiniyatāvadheyavišayanišţham ekāgryam).

Figure 2: The twofold system of yoga in the Yogatārāvalī

Medieval discourse on Rājayoga does not mention different levels of samādhi as seen in Buddhism and Patañjali's Yogaśāstra. Rājayoga texts consistently described samādhi as the stone-like state of complete inactivity; mental, respiratory and physical. Like meditating ascetics in the Mahābhārata, the Rājayogin in samādhi is sometimes said to be lifeless like a piece of wood or, according to the Yogatārāvalī, as still as a lamp's flame in a windless place:

"When the movement of the breath has stopped because of the prolonged restraint of the mind and senses, the bodies of the best yogins become unmoving, like the flames [of lamps] in a windless place, and their minds immersed in the no-mind [state] of mind."

As I have mentioned above, the connections between the Yogatārāvalī, the Amanaska and the Haţhapradīpikā indicate that the Yogatārāvalī was probably written between the twelfth and fifteen centuries. However, the colophon 8 of at least one manuscript and several printed editions attribute the Yogatārāvalī's authorship to the great Advaitavedāntin philosopher, Śańkarācārya, who is generally believed to have lived in the eighth century. Colophons are not a certain indication of authorship, for they may be composed and modified after the text was written. In this case, a sectarian claim to the Yogatārāvalī could have prompted someone to compose colophons that attributed the text to Śańkarācārya. Terminology such as ātma, turīya, etc., which is reasonably common in Tantras and medieval yoga texts of this time but also prominent in Advaitavedāntin texts, could have been used to justify the attribution. It is also possible that the author was a yogin by the name of Śańkara and, at some more recent time, a scribe embellished his name in the colophon with the honorific titles reserved for the famous Śańkarācārya. In my opinion, the Yogatārāvalī was not composed before the twelfth century because its technical terminology, such as Rāja and Haţhayoga, is absent in Sanskrit works written before this time, including the famous commentaries and works widely attributed to the great Advaitavedāntin philosopher.

T HE TWELFTH TO THE FIFTEENTH century was a remarkable time in the history of yoga. It saw the emergence and advance of physical yoga techniques and the supremacy of the practice of samādhi as the means to liberation in yoga traditions. Early Haţha and Rājayoga traditions created simple soteriological systems, which were based solely on the practice of yoga (rather than ritual, gnosis or devotion). In doing so, they omitted the doctrinal and ritualistic complexity of earlier tantric and philosophical traditions.

These yoga traditions incorporated some of the techniques and terminology found in earlier Tantras. However, it is misleading to refer to medieval Haţha and Rājayoga as types of tantric yoga because, not only did their early texts omit tantric ritual and doctrine, but they also fashioned new systems of yoga out of simpler methods of meditation and more physical techniques than are found in older traditions of Tantra. Generally speaking, tantric yoga is usually characterised by complex meditative practices that integrate elaborate metaphysics and doctrine with visualisation. 9 The complexity of tantric yoga contrasts sharply with the simple meditation techniques of early Haţha and Rājayoga texts, such as Śāmbhavī Mudrā and nādānusandhāna.

The shift towards physical practice, doctrinal simplicity and stonelike samādhi (as opposed to gnostic meditation techniques 10 ) suggest that Haţha and Rājayoga were influenced by older ascetic traditions. Indeed, the emphasis on retaining semen (bindudhāraņa) and the physical techniques peculiar to Haţhayoga, such as the three locks (bandha) and inversions (viparītakaraņī), is absent in earlier Tantras.

Nonetheless, the label 'ascetic' does not capture the ingenuity of early Haţha and Rājayoga systems, which omitted extreme ascetic practices such as sitting amid five fires, standing on one leg for twelve years, lying on a bed of nails or holding the arms up until they wither away. Some of their texts denounce afflicting the body (kāyakleśa), and those who would practise headstand are advised against fasting, which is a moderate form of asceticism that has been acceptable to most orthodox Hindus.

Haţhayoga, the yoga of force (haţha), was a yoga that could force certain changes to occur. For example, it could force downward moving vitality (i.e., apānavāyu) to rise. 11 Nonetheless, the difficulty of its techniques was probably a matter of one's perspective. 8 A colophon is a brief statement about the text's author, name and, sometimes, its date. Colophons are usually placed at the end of each chapter and the end of the text itself. 9 I am aware of two earlier tantric works which are exceptions to this. They are known as the Vijñānabhairavatantra and the Svabodhodayamañjarī. They teach simple contemplative practices without much doctrine and metaphysics. Their emphasis on dissolving the mind (cittalaya) suggests they may have inspired later Layayoga traditions. 10 On the differences between ascetic and gnostic meditative practices, see Johannes Bronkhorst, On the one hand, Jain ascetics and many wandering renunciants must have seen Haţhayoga as child's play or, perhaps, they thought it was asceticism watered down for the masses. On the other hand, the physical practices of Haţhayoga would have seemed strenuous and difficult for gnostics who believed that liberation could be attained through listening to the Upanišads, contemplative techniques and the like.

In order to appeal to a wide audience, early Haţha and Rājayoga traditions combined tantric and ascetic techniques within a radically simplified doctrinal framework. As an early Haţha text made clear:

brāhmaņaģ śramaņo vāpi bauddho vāpy ārhato 'thavā | kāpāliko vā cārvākaģ śraddhayā sahitaģ sudhīģ || yogābhyāsarato nityaµ sarvasiddhim avāpnuyāt | 12 ""Whether a Brahmin, ascetic, Buddhist, Jain, skull-bearer (kāpālika) or materialist; one who is wise, has confidence [in the teachings of Haţha and Rājayoga] and is devoted to the practice of yoga, will always obtain success in all things."

The strategy of attracting a wide audience was stated explicitly by a more recent Haţha text called the Haţhābhyāsapaddhati, which was probably composed in the eighteenth century. The opening line of this text claims that its teachings are for anyone who is afflicted by the pain of transmigration (saµsāra), much like the Yogatārāvalī's first verse.

However, the Haţhābhyāsapaddhati also mentions specifically that its teachings are for women, as well as those people who are completely attached to sense objects, those fallen from caste and those who do extremely reckless actions. 13 The explicit inclusion of women here has much to do with the fact that they were excluded from certain religious activities in orthodox Hinduism. For example, women were generally prohibited from learning the Vedas, using vedic mantras, renouncing society to become ascetics and so on.

Some other yoga texts such as the Śivasaµhitā and Yogayājñavalkya reveal that householders, women and members of the lowest orthodox caste (śūdra) were among their audience. Though most texts do not identify the people for which their teachings were intended, explicit prohibitions against teaching a particular group of people are absent.

The inclusivity of these yoga traditions is further indicated by the marginal role of theism in their texts. Though the teachings are presented within either a Śaiva or Vaišņava framework, the theistic elements are very much in the background, so much so that a text's orientation may not always be clear to the reader. The Yogatārāvalī is a good example of such minimal theism. In one verse, Śiva is mentioned as the first teacher of the many methods of absorption (laya) and, in two other verses, samādhi is referred to as the state of Višņu (višņupada). Nothing further is said of these deities, nor their pantheons, myths, mantras and the devotional practices associated with them.

On the whole, the texts of early Haţha and Rājayoga traditions reveal varying degrees of universalism. In other words, they reduced the religious and philosophical elements that may have excluded people of different creeds. They offered a minimalist system of yoga for attaining liberation. The only essential requirement was that one practised the methods of these yogas, having learnt them from a guru. The Yogatārāvalī's emphasis on practice and its omission of complex metaphysics, doctrine, ritual and extreme asceticism is in keeping with this.

The emphasis on practice (abhyāsa) is at the heart of Haţha and Rājayoga texts, which frequently reiterate the importance of practice for attaining the goals of yoga. For example, 15 ] in which extraordinary happiness arises from ceaseless practice, blossoms in the yogin whose roots of intentional 16 The main exceptions to this are the publications of the Kaivalyadhama Yoga Institute and the Lonavla Yoga Institute. 17 I wish to thank Eddie Stern for pointing this out to me (p.c. 28.10.2014).

12 Dattātreyayogaśāstra 41-42ab. This text can be dated to the 12-13th c. and is one of the earliest extant texts to teach Haţhayoga (see the article on 'Haţhayoga' by James Mallinson at https://soas.academia.edu/JamesMallinson). 13 The opening lines of the Haţhābhyāsapaddhati state: "For those afflicted by the pain of Saµsāra; those completely attached to sense objects; women; those fallen from their caste and those who do extremely reckless actions; for their sake, this Haţhābhyāsapaddhati, composed by Kapālakuraņţaka, was written [...]." (saµsāratāpataptānāµ || atyantavišayasaktānāµ || straiņānāµ jātibhrašţānām || atisāhasakarmakartřņām || tatkŗte || iyaµ kapālakuraņţakakŗtahaţhābhyāsapaddhatir [...]

likhyate).

14 Dattātreyayogaśāstra 105cd-107ab. Also see Dattātreyayogaśāstra 41-42ab. (footnote 12). 15 For more information on the use of the term yoganidrā as a synonym for samādhi, see: http://theluminescent.blogspot.co.uk/2015/01/yoganidra.html and unintentional thought have been cut away and whose network of karma has been completely uprooted." In particular, a detailed and comprehensive history of Rāja and Haţhayoga has not been written. Indeed, such a history cannot be written until more evidence has been made available. If one looks at the textual sources in most historical accounts of yoga that have been published in the last thirty years, very few new sources have been brought to light. 16 Certainly, new observations and theories have been advanced in regard to yoga's history, but attempts to construct the history of Haţhayoga with only a few widely known texts such as the Haţhapradīpikā, Śivasaµhitā and Gheraņďasaµhitā are failing to advance our knowledge in any significant way. The omission of texts such as the Yogatārāvalī in secondary sources is proof of the premature attempts at writing the history of Haţhayoga. Various other yoga texts remain in Indian libraries, unedited, unstudied and unknown to scholars and practitioners.

A good example of the rudimentary state of scholarship on yoga is the fact that scholars are still wondering why there are so few postures (āsana) in medieval yoga sources compared with the large number known at the beginning of the twentieth century. The Haţhapradīpikā has fifteen āsana and the Gheraņďasaµhitā, thirty-two, but Kŗšņamācārya taught more than two hundred. Part of the answer to this quandary can be found in some Haţhayoga texts written between the sixteenth and eighteenth centuries. In the aggregate, these texts list and describe hundreds of āsana. When this textual evidence is edited, translated and published, only then will more complete histories of yoga be written.

The other part of the answer consists of the influence of European physical culture and Indian traditions such wrestling (mallavidyā) and martial arts, as well as the innovative genius of twentieth-century yoga gurus such as Kŗšņamācārya and Pattabhi Jois. Though some late medieval Haţhayoga traditions practised numerous āsana, their texts do not mention special sequences of āsana, the movements called vinyāsa, nor the sun salutations known as sūryanamaskāra. In fact, the two types of sūryanamaskāra taught by Pattabhi Jois are based on Kŗšņamācārya's vinyāsa format. 17 Nonetheless, it is likely that sūryanamaskāra in one form or another is an ancient practice. Brahmānanda, the nineteenth-century commentator on the Haţhapradīpikā, knew of a rather Their manuscripts tend to be roughly copied. Therefore, rarely can one rely on a single manuscript to read and understand a text. It is more a matter of using several manuscripts of the same text in order to piece together a complete and reasonably coherent version.

I would like to have included a translation of the Yogatārāvalī in this edition of Nāmarūpa, but I am unable to do so because there are many unresolved differences between its printed editions. As far as I am aware, no one has looked at all the available manuscripts of the Yogatārāvalī in order to resolve these differences. Some editions and manuscripts have twenty-eight verses, others twenty-nine or thirty. And there are many textual variations. The Yogatārāvalī's first verse is a good example of this, for a manuscript at a library in Pune and three printed editions have the following reading:

vande gurūņāµ caraņāravinde sandarśitasvātmasukhāvabodhe | janasya ye jāńgalikāyamāne saµsārahālāhalamohaśāntyai || "I pay homage to the gurus' lotus feet [...] which act like toxicologists for people in order to cure their delusion, which is the poison [known as] worldly life."

The first line is the same, but the second begins with janasya ye instead of niģśreyase, which is found in a manuscript of the Yogatārāvalī at a library in Wai. Which reading did the author intend? Though an editor might speculate that the word niģśreyase ("unsurpassed") has greater poetic value than janasya ye ("which for people") and is more comparable with the Yogatārāvalī's register of Sanskrit, such an editor could not make an informed decision on this without examining the available manuscripts and their relationship to one another. Indeed, someone may have changed janasya ye to niģśreyase at a relatively recent time because that person believed the text could be improved this way. Moreover, regardless of whether an editor decides to read janasya ye or niģśreyase, a good critical edition will provide readers with the available textual evidence so that they can judge for themselves.

Owing to problems such as janasya ye in the first verse, all my above translations of the Yogatārāvalī's verses in this article are provisional. Academic funding for five years has been offered to James Mallinson (the principal investigator), Mark Singleton and myself to critically edit and translate ten Haţhayoga texts, including the Yogatārāvalī. We intend to consult all of the Yogatārāvalī's available manuscripts, of which there are at least twenty. If all goes according to plan, the new edition should be completed by 2018.

THE IMPORTANCE OF THE HISTORY OF YOGA

W ITH OVER TWO THOUSAND YEARS of tradition behind the word 'yoga', there is much to learn from its history. When modern yoga is criticized for teaching a physical practice without a sophisticated intellectual philosophy, one might cite the early Haţha and Rājayoga traditions as a precedent for the effectiveness of this approach. Indeed, I suspect the author of the Yogatārāvalī would have agreed with Pattabhi Jois' statement that yoga is ninety-nine percent practice and one percent theory. Medieval yoga traditions were a great inspiration to Kŗšņamācārya and his students, and there's no reason why it will not be so for future yoga practitioners.

The history of yoga might also inform us of the fate of minimalist yoga traditions. After the sixteenth century, Haţha and Rājayoga were absorbed by more orthodox Brahmanical traditions, which integrated the teachings of earlier Haţha and Rājayoga texts with more sophisticated philosophical and metaphysical doctrines. This gave rise to many of the so-called Yoga Upanišads and other large yoga compilations such as the Yogacintāmaņi. As yoga becomes a contemporary mainstream practice, history is repeating itself. Those who believe they are advancing Haţha yoga by combining it with some other therapy, science, philosophy or religion are travelling down a wellworn path.