History of yoga

Almost every history of yoga you will find online recycles the same shape: five periods, five thousand years, a seal from the Indus Valley, then Patanjali, then a jump to the present. I have read a great many of them and they agree with each other far more than they agree with the scholarship, which is a warning sign rather than a comfort.

The problem is not that those pages contain lies. It is that they flatten three completely different kinds of statement into one confident voice. This page keeps them apart throughout, and that is its whole method:

• Documented. There is a text, an inscription, a dated manuscript or a photograph. The Hatha Yoga Pradipika exists, it is fifteenth century, and it describes fifteen postures. That is documented.
• Inferred. Scholars reason from evidence to a probable conclusion, and disagree with each other. The dating of Patanjali is inferred. So is the claim that hatha yoga was first codified in Buddhist circles, though it is now inferred by the best equipped project ever to look at the question.
• Devotional tradition. The account a living tradition gives of itself: Shiva as the first teacher, Hanuman and the sun salutation, a lineage traced to a sage. These are meaningful, they are what most Indian practitioners were actually taught, and they are not history. Calling them history does them no favours.

I am a teacher, not a historian, and I want to be exact about what that means for this page. Every dated claim below comes from a source I have named and you can open. Where scholars disagree, I say so rather than picking the tidier version. And where I have something of my own to add, it is about what forty years in a hall has and has not shown me, which is a small contribution to a large subject.

Start with the number everybody quotes, because if it goes, the whole five period scheme goes with it.

The claim rests almost entirely on one object: a small seal excavated at Mohenjo-daro in 1928 or 1929, dated to roughly 2200 BCE, and now in the National Museum in New Delhi. It shows a seated horned figure surrounded by animals. John Marshall, who led the excavations, identified it as a proto Shiva, and among his four reasons was that the figure is "in a typical yoga attitude". That identification was widely accepted for generations, and it is where the seal acquired the name Pashupati and the reputation as the oldest evidence of yoga.

Here is what has happened to that reading since. The Indus script remains undeciphered, so the seal cannot be made to say anything at all. Geoffrey Samuel has argued that the competing interpretations of the figure cannot all be right and that the evidence for yogic practice in this material is so dependent on reading later practices backwards into it that its historical value is minimal. Wendy Doniger has made the plainer objection: we cannot know what it meant, and we cannot assume the Indus images are the source of the Hindu ones. Current scholarly summaries treat the yoga reading as speculative rather than established.

Two further problems with the number itself, which nobody who quotes it seems to have checked. The urban phase of the Indus civilisation runs from about 2600 to 1900 BCE, so even on the most generous reading the seal is around four thousand two hundred years old rather than five thousand. And a seated figure is not a practice. People have sat cross legged in every culture that sits on floors, which is most of them, and padmasana is a way of sitting before it is a posture, and my grandmother sat in something an enthusiast would photograph and call an asana.

Honest verdict: the seal is genuinely fascinating and genuinely inconclusive. It belongs in a history of yoga as an open question, and the sentence "yoga is five thousand years old, as proved by the Indus Valley seals" is not supportable. What the seal does not do is diminish anything. A tradition with two and a half thousand years of textual history behind it does not need a rounded up number.

The word yoga is old and its early meanings are not ours. In the Rigveda, whose oldest layers are usually placed around 1200 to 900 BCE, the root appears in the sense of yoking: harnessing animals to a cart, and by extension harnessing anything. Hymn 5.81.1 is the usual citation. This is real and it is not a description of practice. Reading a yoga class into the word at that stage is like reading a gymnasium into the Latin for exercise.

The first places where yoga starts to mean something a practitioner would recognise are the middle Upanishads, particularly the Katha and the Maitrayaniya, the latter often placed around the fourth century BCE. The Maitrayaniya describes yoga in terms of breath and meditation. The Katha works with restraint of the senses, and its chariot image, the senses as horses and the mind as reins, is the direct ancestor of the definition Patanjali would later give. That is the earliest documented yoga in any sense we would accept, and it is about two and a half thousand years old, not five.

Two things are worth saying about this period that the standard blog history omits entirely. The first is that the practices appear at the edges of Brahmanism rather than at its centre: they belong to ascetics, wanderers and renouncers. The second is that Buddhist and Jain traditions were developing systematic contemplative methods in the same centuries, and scholars have proposed that some of the earliest practices may have arisen in the Jain tradition. Yoga was never the property of one religion, and the historical record shows several traditions working on the same problems at once and borrowing from each other freely.

What is entirely absent from this whole layer is a practice built out of postures. Restraint, breath, seat, attention: those are the materials, and the nearest thing to a named posture in the whole layer is a cross legged seat such as sukhasana. If you want to know what the ethical and meditative frame of that period actually asked of a practitioner, the eight limbs are the compact version of it, and they were written down at the next stage.

The Yoga Sutras of Patanjali is the text everyone names and, in my experience of talking to students, the text almost nobody has read. Dating is contested across a wide range, with later dates in the early centuries of the common era more commonly accepted now. It contains 195 or 196 sutras depending on how you count, in four chapters.

Now the fact that changes how you read every yoga advertisement in India: roughly three of those sutras concern posture. The definition given is that asana is a seat that is steady and comfortable, and the instruction that follows is about effort relaxing rather than about ranges of movement. There is no sun salutation, no standing sequence, no list of postures, no benefits per pose. The text is a philosophy of mind. Its second sutra defines yoga as the stilling of the movements of consciousness, and the remaining hundred and ninety odd sutras are about how thought behaves and what happens when it settles.

This is not a debunking. It is the opposite: it is the reason the text is still worth reading, and it explains why a reader who arrives from a posture practice finds it so strange. If you take the Yoga Sutras as the origin document of a physical discipline, it reads as a peculiar and disappointing book. Taken as what it is, it reads as one of the sharpest descriptions of attention anybody has written.

The other unexpected part of its history is what happened next: nothing, for a very long time. The text fell into relative obscurity for roughly seven hundred years, from about the twelfth to the nineteenth century. It did not sit at the centre of Indian practice through the medieval period, because the medieval period had other books, which is the next section. Its return to prominence is a modern event, driven by Vivekananda and the Theosophical Society in the late nineteenth century. The scholar Elizabeth de Michelis treats Vivekananda's Raja Yoga of 1896, an interpretation and adaptation of the sutras, as marking the beginning of modern yoga altogether.

If you want the ancestors of what happens in a yoga hall, this is the layer to read, and it is the layer the popular histories skip in a sentence.

Hatha yoga originated among diverse ascetic traditions around 1000 CE, and the single most striking finding of recent scholarship is that physical yoga appears to have been first codified by Buddhists. That is the conclusion of the Hatha Yoga Project, a five year programme at SOAS in London funded by the European Research Council and led by James Mallinson, which combined philology with ethnography and which SOAS says upended the understanding of yoga as an extreme Hindu ascetic practice that only became health oriented under colonialism. The Amritasiddhi, roughly eleventh century with a manuscript dated 1160, comes out of a tantric Buddhist milieu. The Dattatreyayogashastra of the thirteenth century is the earliest systematised hatha text.

The best known of them, the Hatha Yoga Pradipika, was compiled by Svatmarama in the fifteenth century from earlier sources. Read the posture counts and keep them in mind the next time somebody tells you the sequence they teach is ancient:

Two observations from that table. First, fifteen postures in the most famous hatha text is fewer than a general class uses in the first twenty minutes, and the practices those texts spend their pages on are the ones a modern class treats as optional extras: breath retention, the seals, and cleansing practices of the kind kapalbhati belongs to. The inverted practices in those texts, the ancestors of sarvangasana and sirsasana, are described for what they were believed to do to vital essence rather than for the shoulders. Second, the posture counts rise steeply as you approach the nineteenth century, and the biggest jump of all happens in Mysore. That is not a coincidence and it is the subject of the next section.

Mallinson and Singleton's Roots of Yoga, published by Penguin Classics in 2017, gathers passages from over a hundred texts in eleven languages and is the book I would put in the hands of anybody who wants this layer without an intermediary. Its central point is the one this page keeps returning to: yoga is not a static object but many strands, and posture is one element among ethical rules, breath, and meditation rather than the trunk of the tree.

This is the part that annoys people, and it should be stated carefully rather than provocatively.

Modern posture practice, meaning a class of flowing sequences and held postures taught for health, is substantially a creation of the twentieth century. Mark Singleton's Yoga Body: The Origins of Modern Posture Practice, published by Oxford University Press in 2010, is the standard account. His argument, briefly: asana was held in low regard in nineteenth century India, the yoga Vivekananda took to the West in the 1890s contained essentially no asana at all, and from the 1920s onward a posture based practice emerged that drew on Indian nationalism, on European physical culture and gymnastics, and on the reform programmes of teachers such as Shri Yogendra and Swami Kuvalayananda.

The dates that anchor the story are these. Vivekananda addressed the World's Parliament of Religions in Chicago on 11 September 1893 and published Raja Yoga in 1896, and neither event involved postures. Krishnamacharya was given a yogashala at the Jaganmohan Palace in Mysore on 11 August 1933, under the patronage of the Maharaja, and published Yoga Makaranda in 1934 with 42 asanas and 95 photographs. Norman Sjoman's The Yoga Tradition of the Mysore Palace of 1996 established that the hall Krishnamacharya was given had been a gymnasium, ropes still in place, and that some of his postures resemble those in the palace exercise manual and in the Indian pole and rope discipline of mallakhamba. The Wellcome Collection, writing about the same period, notes that palace patronage put Krishnamacharya in competition with fashionable bodybuilding gyms, which pushed him towards spectacular, demonstrable sequences rather than long held seats.

Singleton is careful about his own thesis in a way his critics and his fans both tend to lose, and it is worth quoting the shape of it. He does not say the postures were invented from nothing. He describes adaptation, reinterpretation and bricolage: the old adapted and transformed into something new, fit for a radically different social world. That is a different claim from fraud, and it is also a much more interesting one.

What follows from Mysore is most of the map of a modern timetable. Ashtanga came from Pattabhi Jois, Iyengar from B. K. S. Iyengar, and the whole vinyasa family from the flowing method those two learned in the same hall, which is why hatha and vinyasa are best understood as two ends of one twentieth century argument. The lineage is set out properly in Krishnamacharya and the Mysore lineage, and the sun salutation has its own documented and surprisingly recent history that is worth reading beside it.

Here is the whole history in one table, with the third column doing the work that the five period scheme refuses to do. Read the status column and you will never be taken in by a yoga school's founding myth again.

Notice what the table does to the question "who invented yoga". It dissolves it. There is no inventor because there is no single thing: there is a word that changed meaning, several traditions that borrowed from one another for two thousand years, a philosophy of mind, a medieval body practice, and a modern synthesis. Anybody who names one founder is telling you about their school rather than about history.

I have been careful on this page to keep out of the historians' way, and I want to be equally careful about the one thing I can genuinely report, which is what happens to this material in a room.

Two observations from forty years, offered as experience and not as evidence. The first is that the students who last are almost never the ones who came for the history. Antiquity is a good advertisement and a poor motivation, and the person who stays eleven years came because their back stopped hurting. The second is that the parts of the old material which survive contact with modern life are precisely the parts the popular history ignores: the breathing, the seat, the deliberate slowness, and the ethical observations in the eight limbs about not forcing. The postures change every generation, in every school. The instruction to stop before the edge does not, and I have watched it prove itself in more bodies than I can count.

What I cannot tell you is anything about what I inherited from whom, and I will not manufacture it. There is a strong temptation in this business to attach yourself to a lineage, because a name and a chain of teachers sells a class. What I can say is only this: I have taught in a residents' association hall in north Bangalore for about forty years, to all ages, and I am eighty nine. That is the entire provenance on offer, and a page about historical honesty would be an odd place to start improving on it.

Which leaves the useful question. A tradition this old and this mixed is not something you can join by reading about it, and the seventeen styles on this site, from Sivananda to yin, are all downstream of the history above, all of them respectable, and none of them the original. If you want to feel what any of it actually is, the only way is an hour of it. Come to a free class and bring the awkward questions: ask why we do a posture the modern texts have no name for, or where the breathing came from, and I will tell you what I know and say plainly when I do not know. That is a better use of a Sunday than another article about five thousand years.