For most of the twentieth century, the story of hatha yoga’s origin was reasonably settled. It emerged from the Nath tradition, associated with Gorakhnath, somewhere around the eleventh or twelfth century, within a Shaiva context, and the Hatha Yoga Pradipika of the fifteenth century was its major statement.
Then a manuscript that had been sitting in catalogues for decades was properly edited and translated, and the settled story stopped working.
The Amritasiddhi, the Attainment of Immortality, dates to around the eleventh century, which makes it the earliest known text to teach the techniques that later become hatha yoga. And it is a Buddhist text, composed within a tantric Buddhist environment, whose earliest surviving manuscript is dated 1160 CE and is bilingual, in Sanskrit and Tibetan.
What it contains
The Amritasiddhi teaches a physical, body-based method aimed at preventing the loss of bindu, a vital drop understood as located in the head and constantly dripping downward toward destruction in the fire of the abdomen. Death, in this model, is the exhaustion of bindu. Stop the fall, and you stop dying.
The techniques it gives for this are the ones later hatha texts will repeat: mahamudra, mahabandha and mahavedha, three practices combining posture, breath retention and muscular locks. These appear in the Hatha Yoga Pradipika three or four centuries later in a recognisably continuous form.
What it does not contain is equally significant. There are no asanas in the modern sense. There is no chakra system of the kind popular writing assumes. There is no kundalini as a coiled serpent rising through centres. Those elements arrive later, from other streams, and get retrofitted onto the practices.
Why this changes the picture
The revision came largely through the Hatha Yoga Project at SOAS, University of London, directed by James Mallinson, with the critical edition of the Amritasiddhi produced by Mallinson and Péter-Dániel Szántó. The project spent years examining manuscripts that had never been edited, and the Amritasiddhi was the headline finding.
The implications run in several directions.
Hatha’s core techniques were not originally Shaiva. They were developed in a Buddhist vajrayana milieu and were subsequently adopted, adapted and reframed by Shaiva traditions, which then became their principal carriers.
The techniques also travelled without their theology. The Amritasiddhi’s bindu-preservation rationale is specific and physiological. Later texts keep the practices and attach different explanations to them, which tells you something about how these traditions actually worked: methods were portable, frameworks were not.
And the ascetic yoga of the earlier period, the tapas and posture traditions visible in earlier literature, is a distinct stream that merges with this one rather than producing it.
A chronology worth having
| Text | Approximate date | Tradition | Contribution |
| Amritasiddhi | 11th century | Buddhist vajrayana | Earliest known teaching of mahamudra, mahabandha, mahavedha. Bindu preservation model |
| Dattatreyayogashastra | 13th century | Vaishnava | Earliest text to teach hatha yoga under that name as a systematic method available to all |
| Goraksha Shataka | 13th century | Nath, Shaiva | Kundalini and the subtle body integrated with the physical techniques |
| Vivekamartanda | 13th-14th century | Nath | Further development of the subtle body model |
| Hatha Yoga Pradipika | 15th century | Nath, Shaiva | The great synthesis. Compiles material from at least twenty earlier sources |
| Gheranda Samhita | 17th century | Vaishnava | Sevenfold practice. The largest asana count in the classical corpus, at thirty-two |
| Hatha Ratnavali | 17th century | Shaiva | Lists eighty-four asanas by name, though it describes fewer |
Note the asana numbers at the bottom. Even in the seventeenth century, the classical texts describe a few dozen postures at most, and most are seated. The modern repertoire of standing sequences is a twentieth-century development, and this table is the quickest way to see it.
The honest caveats
This is active scholarship, not settled consensus, and it should be presented that way.
The dating of the Amritasiddhi rests on manuscript evidence and internal features, and dating Sanskrit texts is rarely precise. Some scholars have argued for a more gradual or more mixed origin than a clean Buddhist-to-Shaiva transfer. The relationship between the Amritasiddhi and earlier oral practice is unknowable, since a text can record a technique that predates it by centuries.
What is not seriously disputed is that the Amritasiddhi is the earliest text we have teaching these specific techniques, and that its context is Buddhist. That alone is enough to require the older account to be revised.
Frequently asked questions
What is the Amritasiddhi?
An eleventh-century Sanskrit text, the earliest known to teach the physical techniques that later became hatha yoga, composed in a tantric Buddhist context.
Is hatha yoga originally Buddhist?
Its earliest documented techniques appear in a Buddhist text, and were later adopted and reframed by Shaiva traditions. The picture is one of exchange rather than a single origin.
What techniques does the Amritasiddhi teach?
Mahamudra, mahabandha and mahavedha, combinations of posture, breath retention and muscular locks aimed at preventing the loss of bindu.
Does the Amritasiddhi describe yoga poses?
Not in the modern sense. It contains no asana sequences, no chakra system as popularly understood, and no serpent kundalini imagery.
Who established the dating of this text?
The critical edition was produced through the Hatha Yoga Project at SOAS, University of London, by James Mallinson and Péter-Dániel Szántó, working from manuscripts including a bilingual copy dated 1160 CE.
How many asanas do the classical hatha texts describe?
Very few by modern standards. The Hatha Yoga Pradipika describes about fifteen, the Gheranda Samhita thirty-two, and the Hatha Ratnavali names eighty-four while describing fewer.
Sources
Primary: James Mallinson and Peter-Daniel Szanto, eds. and trans., The Amrtasiddhi and Amrtasiddhimula: The Earliest Texts of the Haöhayoga Tradition (2021), published by the Institut Francais de Pondichery and the Ecole francaise d’Extreme-Orient. This critical edition is the basis for everything in this article and should be cited by anyone repeating the claim.
Modern scholarship: James Mallinson and Mark Singleton, Roots of Yoga (2017), which places the Amritasiddhi within the wider corpus and prints translated extracts from the full range of hatha texts. The Hatha Yoga Project at SOAS, University of London, ran from 2015 to 2020 and its publications and manuscript work are the context for this revision. Jason Birch’s work on the dating of hatha texts, particularly on the meaning of hatha and on the later corpus.
Leave a Reply