Chapter 17 The Amṛtasiddhi: Haṭhayoga’s Tantric Buddhist Source Text

In: Śaivism and the Tantric Traditions
James Mallinson
Search for other papers by James Mallinson in
Current site
Google Scholar
Open Access

Like many of the contributors to this volume, I had the great fortune to have Professor Sanderson as the supervisor of my doctoral thesis, which was a critical edition of an early text on haṭhayoga called the Khecarīvidyā. At the outset of my work on the text, and for several subsequent years, I expected that Sanderson’s encyclopedic knowledge of the Śaiva corpus would enable us to find within it forerunners of khecarīmudrā, the haṭhayogic practice central to the Khecarīvidyā. However, notwithstanding a handful of instances of teachings on similar techniques, the fully-fledged practice does not appear to be taught in earlier Śaiva works. In subsequent years, as I read more broadly in the corpus of early texts on haṭhayoga (which, in comparison to the vast Śaiva corpus, is relatively small and thus may easily be read by one individual), I came to the realisation that almost all of the practices which distinguish haṭhayoga from other methods of yoga were unique to it at the time of their codification and are not to be found in the corpus of earlier Śaiva texts, despite repeated assertions in secondary literature that haṭhayoga was a development from Śaivism (or “tantra” more broadly conceived).

The texts of the haṭhayoga corpus do, however, couch their teachings in tantric language. The name of the haṭhayogic khecarīmudrā, for example, is also that of an earlier but different Śaiva practice. When I was invited to speak at the symposium in Professor Sanderson’s honour held in Toronto in 2015, I decided to try to articulate my rather inchoate thoughts on this subject by presenting a paper entitled “Haṭhayoga’s Śaiva Idiom.” The inadequacy of my theories was brought home to me some months after the symposium when I started to read, together with two other former students of Sanderson, Péter-Dániel Szántó and Jason Birch,1 a twelfth-century manuscript of the Amṛtasiddhi (AS), the earliest text to teach many of the key principles and practices of haṭhayoga.2 I had already read much of the text with Sanderson and others, but only from later manuscript sources. As we read the older manuscript it gradually became clear that the Amṛtasiddhi was composed in a Vajrayāna (tantric Buddhist) milieu.

Thus my notion of haṭhayoga having a Śaiva idiom needed readdressing. One might perhaps talk instead of its “tantric idiom.” But I shall leave reflections on that topic for a later date and in this short paper focus on the Amṛtasiddhi and, in particular, the features of it which make it clear that it was composed in a Vajrayāna milieu. I am currently preparing a critical edition and annotated translation of the text with Dr Szántó; what follows here results from our work in progress. Despite our edition being incomplete, I am confident that the conclusion drawn here about the origins of the text is sound (and that further work on the text will provide additional and complementary evidence) and I think it important enough to warrant preliminary publication. Subsequent publications will address this unique text’s many other remarkable features.

1 The Amṛtasiddhi

The importance of the Amṛtasiddhi was first brought to scholarly attention by Kurtis Schaeffer in an article published in 2002.3 Here I shall reprise as little of his rich and dense article as is necessary to provide the background to what follows. Schaeffer focuses on the twelfth-century4 manuscript of the text, photographs of printouts from a microfilm of which he and Leonard van der Kuijp have kindly shared with me. At the time that the microfilm was made, the manuscript was in Beijing, although Schaeffer believes that it has since been returned to Tibet. The manuscript is unique in that it is bilingual, with three registers: the Sanskrit text in a Nepali or east Indian script, a transliteration of the Sanskrit in Tibetan hand-printing script and a translation5 into Tibetan in the Tibetan cursive script.

This manuscript is referred to in what follows by the siglum C. The other witnesses of the text which have been collated are considerably later than C (the oldest is perhaps the c. 17th-century K1). They present versions of the text in which redaction has removed or obscured some of the Buddhist features evident in C. These witnesses may be divided into two groups. The first is a single Grantha manuscript from the Mysore Government Oriental Library (M2), the second seven north Indian and Nepali manuscripts, two from Jodhpur’s Maharaja Man Singh Pustak Prakash (J1 and J2 = J) and four from the Nepal-German Manuscript Preservation Project (K1–K4 = K).6

The text of the Amṛtasiddhi consists of 303 verses divided into 35 short vivekas.7 The first ten vivekas teach the constituents of the yogic body. Vivekas 11–13 teach three methods of manipulating those constituents (mahāmudrā, mahābandha and mahāvedha) and viveka 14 teaches the practice (abhyāsa), i.e. how the three methods are to be used together. Vivekas 15–18 teach the four grades of aspirant, 19–33 the four states (avasthās) of yoga, and 34–35 the final transformation of the body leading up to nirvāṇa.8

2 The Amṛtasiddhi in the Haṭhayoga Tradition

2.1 Citations and Borrowings

The Amṛtasiddhi is a seminal work in the haṭhayoga textual tradition.9 Schaeffer (2002, 518–519) mentions its citations in the Yogacintāmaṇi (c. 1600 ce)10 and Haṭhapradīpikājyotsnā (1837 ce).11 In addition, several haṭhayoga texts borrow directly from the Amṛtasiddhi without attribution. The c. 13th-century Gorakṣaśataka shares three half-verses with it.12 The Vivekamārtaṇḍa, which is also likely to date to the 13th century redacts four of the Amṛtasiddhi’s verses into three.13 The c. 14th-century Amaraughaprabodha shares six verses with the Amṛtasiddhi and paraphrases it extensively elsewhere.14 The Gorakṣayogaśāstra (15th century or earlier) borrows two and a half verses15 and extensively paraphrases other parts of the text. The c. 15th-century Śivasaṃhitā is much the biggest borrower from the Amṛtasiddhi, sharing 34 verses with it.16 The Haṭhapradīpikā shares five half-verses with the Amṛtasiddhi, but these may be borrowed from the Amaraughaprabodha since all the shared passages are also in that text.17

2.2 Doctrinal Innovations

Several of the Amṛtasiddhi’s teachings have no prior attestation and are central to teachings on haṭhayoga in later texts, where they are either reproduced verbatim, as noted above, or incorporated into new compositions. These may be summarised as follows.

  1. The Yogic Body

    1. The Amṛtasiddhi is the first text to relocate to the body the old tantric triad of sun, moon and fire.18 The idea of a moon in the skull dripping amṛta is found in many earlier tantric works, but that of the sun in the stomach consuming it is new, as is the conflation of the sun and fire.

      1. The Moon

        meruśṛṅge sthitaś candro dviraṣṭakalayā yutaḥ |

        aharniśaṃ tuṣārābhāṃ sudhāṃ varṣaty adhomukhaḥ ||3.1||

        The moon is on the peak of Meru and has sixteen digits.

        Facing downwards, it rains dewy nectar day and night.

      2. The Sun

        madhyamāmūlasaṃsthāne tiṣṭhati sūryamaṇḍalaḥ |

        kalādvādaśasaṃpūrṇo dīpyamānaḥ svaraśmibhiḥ ||4.1||

        ūrdhvaṃ vahati dakṣeṇa tīkṣṇamūrtiḥ prajāpatiḥ |

        vyāpnoti sakalaṃ dehaṃ nāḍyākāśapathāśritaḥ ||4.2||

        grasati candraniryāsaṃ bhramati vāyumaṇḍale |

        dahati sarvadhātūṃś ca sūryaḥ sarvaśarīrake ||4.3||

        2d °pathāśritaḥ ] CK4; yathāśritaṃ K3, yathāśritaḥ cett.

        3b °maṇḍale ] M2; °maṇḍalaiḥ C, °maṇḍalaṃ cett.

        (1) The sphere of the sun is at the base of the Central Channel, complete with twelve digits, shining with its rays. (2) The lord of creatures (Prajāpati), of intense appearance, travels upwards on the right. Staying in the pathways in the spaces (ākāśapatha)19 in the channels it pervades the entire body. (3) The sun consumes the lunar secretion, wanders in the sphere of the wind and burns up all the bodily constituents in all bodies.

      3. Fire

        kalābhir daśabhir yuktaḥ sūryamaṇḍalamadhyataḥ |

        vasati vastideśe ca vahnir annavipācakaḥ ||5.1||

        yo vai vahniḥ sa vai sūryo yaḥ sūryaḥ sa hutāśanaḥ |

        etāv ekatarau dṛṣṭau sūkṣmabhedena bheditau ||5.2||

        1b °madhyataḥ ] CJK; °madhyagaḥ M2Y

        1c vasati vastideśe ] conj.; vasati vatideśe C, vasate vastideśe M2Y, vasatir asthideśe cett.

        2d sūkṣmabhedena bheditau ] C; sūkṣmāt sūkṣmatarau nṛbhiḥ M2, sūkṣmabhedena bhedinau cett.

        (1) Endowed with ten digits, in the middle of the sphere of the sun in the region of the stomach dwells fire, which digests food. (2) Fire is the sun; the sun is fire. The two look almost the same [but] differ subtly.

    2. The use of the word bindu for semen, bindu’s identification with the amṛta dripping from the moon, its preservation being essential for life and its division into male and female are all innovations of the Amṛtasiddhi which are widely adopted in later haṭhayoga texts.

      • adhaś candrāmṛtaṃ yāti tadā mṛtyur nṛṇāṃ bhavet ||4.11||

        11a yāti ] M2K; yati C, °mṛtaṃ yasya J

        The nectar of immortality in the moon goes downwards; as a result men die.

      • bindupātena vṛddhatvaṃ mṛtyur bhavati dehinām ||21.3||

        The fall of bindu makes men grow old [and] die.

      • sa bindur dvividho jñeyaḥ pauruṣo vanitābhavaḥ |

        bījaṃ ca pauruṣaṃ proktaṃ rajaś ca strīsamudbhavam ||7.8||

        anayor bāhyayogena sṛṣṭiḥ saṃjāyate nṛṇām |

        yadābhyantarato yogas tadā yogīti gīyate ||7.9||

        kāmarūpe vased binduḥ kūṭāgārasya koṭare |

        pūrṇagirimudāsparśād vrajati madhyamāpathe ||7.10||

        yonimadhye mahākṣetre javāsindūrasannibham |

        rajo vasati jantūnāṃ devītattvasamādhṛtam ||7.11||

        binduś candramayo jñeyo rajaḥ sūryamayas tathā |

        anayoḥ saṃgamaḥ sādhyaḥ kūṭāgāre ’tidurghaṭe ||7.12||

        9cd yadābhyantarato yogas tadā yogīti gīyate ] CHJ; yadābhyantarato yogas tadā yogī sa gīyate M2, yadā tv abhyantare yogas tadā yogo hi bhaṇyate cett.

        10a kāmarūpe ] CM2; kāmarūpo cett.

        10b kūṭāgārasya° ] CM2; kūṭādhāraṇya J, kūṭādhārasya K

        10d °mudā° ] C; °sadā° J, °guhā° cett.

        10d vrajati ] C; vrajate M2, rājanti cett.

        11d °samādhṛtam ] C; °samāvṛtaṃ M2, °samāvṛta K1, samāvṛtaḥ cett.

        (8) Know bindu to be of two kinds, male and female. Semen (bīja) is said to be the male [bindu] and rajas (female generative fluid) is female. (9) As a result of their external union people are created. When they are united internally, then one is declared a yogi. (10) Bindu resides in Kāmarūpa in the hollow of the multi-storied palace (kūṭāgārasya).20 Through pleasurable contact at Pūrṇagiri it travels along the Central Channel. (11) Rajas resides in the great sacred field in the perineal region (yonimadhye). It is as red as a javā flower21 and is supported by the Goddess element (devītattvasamādhṛtam). (12) Know bindu to be made of the moon and rajas to be made of the sun. Their union is to be brought about in the very inaccessible multi-storeyed palace.

    3. A connection between the mind and breath is taught as early as the Chāndogya Upaniṣad (6.8.2). The Amṛtasiddhi is the first text to teach that mind, breath and bindu are connected, a notion found in many subsequent haṭhayoga texts.

      calaty ayaṃ yadā vāyus tadā binduś calaḥ smṛtaḥ |

      binduś calati yasyāyaṃ cittaṃ tasyaiva cañcalam ||7.17||

      17a calaty ayaṃ yadā ] C; yadāyaṃ calate M2, yadā caṃcalate JK, calaty eṣa yadā YHJ

      17b calaḥ smṛtaḥ ] JKYHJ; calaḥ smṛ˹ta˺ḥ C, ca cañcalaḥ M2

      17c binduś calati yasyāyaṃ ] C; yasyāyaṃ calate binduḥ M2, yasyāyaṃ calate binduś JK, binduś calati yasyāṅge YHJ

      17d tasyaiva ] CKYHJ; tasthyai∗ M2, tathaiva J

      It is taught that when the breath moves bindu moves; the mind of he whose bindu is moving is restless.

    4. The three granthis.

      The Amṛtasiddhi’s system of three granthis, brahma°, viṣṇu° and rudra°, which are situated along the central channel of the body and are to be pierced by the mahāvedha (13.10–11), is very common in subsequent haṭhayoga texts.22

  2. The three practices, mahāmudrā, mahābandha, mahāvedha (vivekas 11–13).

    These practices, which involve bodily postures and breath control, are used to make the breath enter the central channel and rise upwards. They are an innovation of the Amṛtasiddhi and are taught in all subsequent haṭhayoga texts, albeit sometimes with different names.

  3. The four avasthās

    The four avasthās, “states” or “stages” of yoga practice (ārambha, ghaṭa, paricaya, niṣpanna/niṣpatti) introduced in the Amṛtasiddhi (vivekas 19–33), are taught in many Sanskrit haṭhayoga texts; they are also mentioned in the old Hindi Gorakhbāṇī (śabds 136–139).

In addition to these innovations, in viveka 14 (abhyāsa, “practice”) the Amṛtasiddhi describes, at a level of detail unparalleled in other texts, the internal processes brought about by its methods, in particular the movement of the breaths.

3 Buddhist Features of the Amṛtasiddhi

In Schaeffer’s analysis of the Amṛtasiddhi (2002, 521–524), he notes how it is unique amongst Tibetan Buddhist works because its teachings are said to bestow jīvanmukti, “liberation while living,” and make the yogi identical with Śiva. Despite these Śaiva features, however, close reading of manuscript C, the twelfth-century bilingual witness of the text, shows that the text was composed within a Vajrayāna milieu. Furthermore, it pits its teachings against those of other Vajrayāna schools, not Śaiva ones.

As can be seen in the examples given below, manuscript C generally has the best readings of the text and presents its Buddhist teachings intact. In the other manuscripts the specifically Buddhist doctrines found in C are either unwittingly included, misunderstood (and sometimes presented in corrupt forms as a result) or deliberately changed or omitted.

Some of the text’s Buddhist features are ambiguous or obscure enough for them to have been preserved by the redactors of the text as presented in the later witnesses. Thus we find multiple examples of Vajrayāna (or more broadly Buddhist) terminology such as mahāmudrā (viveka 11 and 31), vajrapañjara (7.26d), jñānasaṃbhāra (6.9c, 20.2bc), śūnya (8.2a, 8.8d, 8.10d, 19.15a, 20.7b, 25.1c), niṣpanna (19.2c, 31.1c) and abhiṣeka (13.15a). Similarly, Amṛtasiddhi 7.4 mentions the very specifically Vajrayāna notion of the four blisses:23

ānandā ye prakathyante viramāntāḥ śarīrataḥ |
te ’pi bindūdbhavāḥ sarve jyotsnā candrabhavā yathā ||7.4||

4c °viramāntāḥ ] C; ciram antaś M2, viramāṃtā JK

The [four] bodily blisses whose last is [the bliss of] cessation all arise from bindu, just as moonlight arises from the moon.

Other Buddhist features of the text as found in C are deliberately omitted or altered in the later witnesses. Examples of these are listed below. This list is not exhaustive; further close reading of the text is likely to reveal more examples.

  1. Chinnamastā

    Manuscript C opens with a sragdharā maṅgala verse in praise of the goddess Chinnamastā:

    nābhau śubhrāravindaṃ tadupari vimalaṃ maṇḍalaṃ caṇḍaraśmeḥ

    saṃsārasyaikasārā tribhuvanajananī dharmavartmodayā yā |

    tasmin madhye trimārge tritayatanudharā chinnamastā praśastā

    tāṃ vande jñānarūpāṃ maraṇabhayaharāṃ yoginīṃ yogamudrām ||

    a śubhrā° ] C; candrā° M2 ●vimalaṃ ] C; vivaraṃ M2

    c tasmin ] C; tasyāṃ M2 ●tri° ] M2; tre° C ●chinnamastā praśastā ] C; cittahasthāṃ praśastāṃ M2

    d tāṃ vande jñānarūpāṃ ] C; vande jñānasvarūpāṃ M2

    At the navel is a white lotus. On top of that is the spotless orb of the sun. In the middle of that, at the triple pathway, is she who is the sole essence of samsara [and] the creator of the three worlds, who arises on the path of dharma, who has three bodies [and] who is lauded as Chinnamastā, “she whose head is cut.” I worship her, she who has the form of knowledge, who removes the danger of death, the yoginī, the seal of yoga.

    Until the 16th century, Chinnamastā is not mentioned in non-Buddhist texts (Bühnemann 2000, 37). Her Vajrayāna origins have been demonstrated by Sanderson (2009, 240–241), who notes how the epithet dharmodayā, found in the Amṛtasiddhi as dharmavartmodayā, is “strictly Buddhist.” One might argue that this maṅgala verse could be an addition to the text when it was redacted by a Vajrayāna tradition, but the verse is also found in the Grantha manuscript M2 in a corrupt form. Chinnamastā’s name is given therein as Cittahasthā, but the epithets dharmavartmodayā and tritayatanudharā are preserved. The Rajasthani and Nepali manuscripts omit the verse.

  2. chandoha

    At Amṛtasiddhi 1.16, manuscript C uses the specifically Buddhist term chandoha:24

    sāgarāḥ saritas tatra kṣetrāṇi kṣetrapālakāḥ |

    chandohāḥ puṇyatīrthāni pīṭhāni pīṭhadevatāḥ ||1.16||

    16c chandohāḥ ] em.; chandohā C, saṃbhedāḥ M2JK

    There are oceans, rivers, regions [and] guardians of the regions; gathering places (chandohāḥ), sacred sites, seats [of deities and] the deities of the seats

    In Śaiva texts chandoha is found as saṃdoha.25 That the manuscripts other than C read saṃbhedāḥ, which makes no sense, suggests that they may derive from an archetype that had saṃdohāḥ, which subsequent copyists did not understand.

  3. The four elements

    Amṛtasiddhi 6.2 refers to four physical elements:

    pṛthivyādīni catvāri vidhṛtāni pṛthak pṛthak ||6.2||

    2a catvāri ] C; tattvāni cett.

    The four [elements] earth etc. are kept separate [by the breath].

    In Śaiva and other Hindu traditions there are five primary physical elements. The later manuscripts therefore change catvāri, “four,” to tattvāni, “elements.”

  4. kūṭāgāra

    This is a common term in the Pali Canon, meaning “a building with a peaked roof or pinnacles, possibly gabled; or with an upper storey” (Rhys Davis and Stede 1921–1925, s.v. kūṭāgāra). It is also found in several Vajrayāna texts, where it refers to a “multi-storeyed palace” in the middle of a maṇḍala (Reigle 2012, 442). It is not found in Śaiva texts and is not recognised by the later north Indian and Nepali witnesses of the Amṛtasiddhi.

    7.10ab kāmarūpe vased binduḥ kūṭāgārasya koṭare |

    10a °rūpe ] CM2; °rūpo JK

    10b kūṭāgārasya ] C, ∗ūṭāgārasya M2, kūṭādhāraṇya° J, kūṭādhārasya K

    Bindu resides at Kāmarūpa,26 in the hollow of the multi-storeyed palace.

  5. trivajra

    8.21 in C mentions the three vajras, i.e. the common Vajrayāna triad of kāya, vāk and citta. In the other witnesses trivajrāṇāṃ is found as trivargāṇāṃ.

    trivajrāṇāṃ samāveśas tadā vai jāyate dhruvam ||8.21||

    21c trivajrāṇāṃ ] C; trivargāṇāṃ M2JK

    Then absorption into the three vajras is sure to arise.

  6. trikāya

    A reference to the Buddhist notion of the triple body is expunged in the later witnesses:

    sarvajñatvaṃ trikāyasya sarvajñānāvabodhakam |

    lakṣaṇaṃ siddhacittasya jñātavyaṃ jñānaśālibhiḥ ||29.2||

    2a °kāyasya ] C; °kālasya M2, °kāryasya JK

    2b °bodhakam ] CM2; °bodhanam JK 2c siddhacittasya ] C; siddhivit tasya JK

    Omniscience, which brings about complete understanding of the triple body, should be known by the knowledgable to be the mark of he whose mind has been mastered.

  7. buddha

    Verses in which C has (or its archetype is likely to have had) buddha are reworked in the later witnesses.

    bindur buddhaḥ śivo bindur bindur viṣṇuḥ prajāpatiḥ |

    binduḥ sarvagato devo bindus trailokyadarpaṇaḥ ||7.15||

    15a buddhaḥ ] em.; vṛddhaḥ C, ūrdhvaḥ cett.

    Bindu is Buddha, bindu is Śiva, bindu is Viṣṇu, the lord of creatures, bindu is the omnipresent god, bindu is the mirror of the three worlds.

    tāvad buddho ’py asiddho ’sau naraḥ sāṃsāriko mataḥ | 32.3ab

    3a buddho ] C; ∗ddho M2, ⊔dvo J1, siddho J2, vaddho K

    Even a Buddha, as long as [he remains] unperfected [by means of the practice taught in the Amṛtasiddhi ], is considered a worldly man.

  8. svādhiṣṭhāna yoga

    In two places the Amṛtasiddhi mentions svādhiṣṭhāna yoga. This is a method of visualising oneself as a deity which is central to the teachings of a wide variety of Vajrayāna texts (e.g. Guhyasamāja 7.2, where it is called svādhidaivatayoga, and the Pañcakrama, whose third krama is called the svādhiṣṭhānakrama). In the two verses from the Amṛtasiddhi given below, the methods of svādhiṣṭhāna yoga are said to be ineffective; to achieve the goals of yoga one must use the practice taught in the Amṛtasiddhi. The later witnesses of the text do not understand the phrase svādhiṣṭhānena yogena and, presumably surmising svādhiṣṭhāna to refer to the second of the six cakras in a system taught in many haṭhayoga texts (but not in the Amṛtasiddhi, which makes no mention of cakras), they change yogena to mārgeṇa in an attempt to make the phrase refer to a pathway in the yogic body.

    svādhiṣṭhānena yogena yasya cittaṃ prasādhyate |

    śilāṃ carvati mohena tṛṣitaḥ khaṃ pibaty api ||8.9||

    9a yogena ] C; mārgeṇa M2JK 9b yasya ] JK; yastuś C, yatna M2 ●prasādhyate ] M2JK; prasādhyati C

    He who tries to master his mind by means of self-established yoga deludedly chews a rock and, thirsty, drinks the sky.

    svādhiṣṭhānena yogena na kṣīyete guṇau nṛṇām |

    asti mudrā viśeṣeṇa gurumukhābjasaṃbhavā ||10.11||

    11a yogena ] C; mārgeṇa M2JK 11b na kṣīyete ] em.; na kṣīyate C, prakṣīyante M2, nākṣipeti JK ●guṇau ] C; guṇā M2, guṇo JK 11c viśeṣeṇa ] CJK; viśeṣād vā M2 11d guru° ] CJK; guror M2 ●°mukhābja° ] C; °vaktrābja° M2, °mukhāt tu JK ●°saṃbhavā ] JK; °saṃbhavāṃ C, °saṃbhavāt M2

    The two [unwanted] guṇas [rajas and tamas] in men are not destroyed by self-established yoga. There is a mudrā especially [for that], born from the lotus-mouth of the guru.

4 Conclusion

The Amṛtasiddhi was composed in a Vajrayāna Buddhist milieu and its intended audience was other Vajrayāna Buddhists. Its teachings are subsequently found in haṭhayoga texts from a wide range of non-Buddhist traditions. This does not mean, however, that haṭhayoga itself was a product of Vajrayāna Buddhists. I have argued elsewhere (e.g. Mallinson 2015) that some haṭhayoga techniques were current among ascetics long before their codification. The Amṛtasiddhi was the first text to codify many of haṭhayoga’s distinctive principles and practices and was thus the first to assign names to them. As a result the Amaraughaprabodha, the first text to teach physical yoga methods under the name haṭha, includes among its techniques the Amṛtasiddhi’s mahāmudrā, mahābandha and mahāvedha (with slight variations in their methods). In addition to these physical techniques, the Amaraughaprabodha also adopts from the Amṛtasiddhi the more theoretical doctrine of the four avasthās or stages of yoga, showing that the Amṛtasiddhi’s influence was more than simply terminological.

Because they share traditions of 84 siddhas, several scholars have posited connections between Vajrayāna Buddhists and Nāth yogis,27 with whom the practice of haṭhayoga has long been associated. The Amṛtasiddhi’s Vajrayāna origins and its borrowings in subsequent haṭhayoga texts, some of which are products of Nāth traditions, provide the first known doctrinal basis for this connection and a stimulus for its further investigation.28

5 Witnesses of the Amṛtasiddhi

5.1 Manuscripts Collated

  • (C) China Nationalities Library of the Cultural Palace of Nationalities Ms. No. 005125 (21). Paper. Sanskrit text in both Nepali (or perhaps East Indian) and Tibetan hand-print scripts, Tibetan translation in Tibetan cursive script.

  • Maharaja Man Singh Pustak Prakash, Jodhpur

    1. (J1) 1242. Paper. Devanagari.

    2. (J2) 1243. Paper. Devanagari.

  • Nepal-German Manuscript Preservation Project. All entitled Amṛtasiddhi.

    1. (K1) E655/39. Paper. Devanagari.

    2. (K2) E1501/11. Paper. Devanagari.

    3. (K3) H232/37. Paper. Newari.

    4. (K4) E68/7. Paper. Devanagari.

  • (M2) Mysore Government Oriental Manuscripts Library D-4342 (ff. 21v–40v). Palm leaf. Grantha.

5.2 Other Collated Witnesses

These two texts are mentioned in the apparatus only in the small number of instances that they provide readings.

  • (Y) Yogacintāmaṇi ed. Haridās Śarmā, Calcutta Oriental Press, n.d.

  • (HJ) Haṭhapradīpikājyotsnā of Brahmānanda. Svāmī Maheśānand, Dr. Bāburām Śarmā, Jñānaśaṃkar Sahāy, and Ravindranāth Bodhe, eds. Brahmānandakr̥tā Haṭhapradīpikā Jyotsnā: Locanātmaka Saṃskaraṇa (Hindī). Lonavla: Kaivalyadhām S.M.Y.M. Samiti. 2002.

5.3 Manuscripts Not Yet Collated

  1. Mysore Government Oriental Library D-4341. Paper. Grantha.

  2. Mysore Government Oriental Library R-2881(n). Palm leaf. Grantha. Incomplete.

  3. Adyar Library 75278. Palm leaf. Grantha.

  4. Baroda Oriental Institute 7970(b). Palm leaf. Grantha.


I thank Dominik Wujastyk for his comments on a draft of this article, the research for which was carried out as part of the Hatha Yoga Project ( This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement no. 647963).


We were joined at our reading sessions by Sam Grimes, Diwakar Acharya, Camillo Formigatti, Anand Venkatkrishnan and Paul Gerstmayr, whom I thank for their valuable comments.


I thank Professors Kurtis Schaeffer and Leonard van der Kuijp for sharing with me photographs of printouts from a microfilm copy of this manuscript. Schaeffer also kindly shared his draft edition of the Tibetan translation of the Amṛtasiddhi given in this witness. We read the manuscript together with a collation of other witnesses, including a transcription of the Grantha manuscript M2 prepared by Viswanath Gupta, whom I thank for his assistance. Concerning manuscript sigla for the Amṛtasiddhi, please refer below to the section “Witnesses of the Amṛtasiddhi.”


Prior to Schaeffer’s article, the only mention of the text of which I am aware (other than in manuscript catalogues) is Gode 1954, 22, in which its citations in the Yogacintāmaṇi are noted.


Schaeffer (2002, 517) says that the manuscript’s colophon gives a date which “may read 1159 c.e.” The reading is clear: ekāśītijute [°jute is Newar scribal dialect for Sanskrit °yute] śāke sahāsraike tu phālgune | kṛṣṇāṣṭamyāṃ samāpto ’yaṃ kṛtvāmṛtasiddhir mayā || (f.37v). The eighth day of the dark fortnight of the lunar month of Phālguna in Śāka 1081 corresponds to March 2nd 1160 ce (according to the calculator at It is possible that the colophon has been copied from an examplar and that the manuscript itself does not date to 1160. The mansucript’s Tibetan colophon says that the Tibetan translation is that of the “monk of the Bya [clan]” (Bya ban de) Pad ma ’od zer, who worked towards the end of the eleventh century, which provides us with an earlier terminus ante quem for the text than the date of the manuscript itself.


As noted in the manuscript’s Tibetan colophon, the translation is of a different recension of the Sanskrit text from that given in the manuscript. At some places, e.g. 7.10 and 7.26, the translation corresponds to the text as found in the other witnesses, but not that in C.


Full details of these witnesses are given at the end of this article.


There are 35 vivekas in the Beijing ms and 38 in the others. All verse numbering given here corresponds to the order of verses in C (which does not itself give verse numbers).


Vivekas 19–35 are interspersed with very short chapters on a variety of topics. In the first viveka (vv. 10–13) there is a list of the topics to be taught in the text. The list corresponds exactly to the vivekas up to viveka 19, but then goes awry. More analysis is needed to be sure, but it seems likely that at least some of the viveka divisions after 19 are later additions to the text.


Despite the compound haṭhayoga being found in earlier Vajrayāna works (Birch 2011, 535–536) and its teachings being central to later haṭhayogic texts, the Amṛtasiddhi does not call its yoga method haṭha. This paradox will be addressed in subsequent publications.


Yogacintāmaṇi p. 13 [AS 6.11, 6.13], p. 26 [14.4, 14.8–12, 14.16c–19d], p. 34 [7.25, 7.15, 7.18, 7.8–9, 7.12, 7.5, 7.2cd, 7.3cd, 7.6–7, 7.16c–17d, 7.19ab, 7.17cd, 7.20–24], p. 39 [32.3–4], p. 40 [33.1, 34.1] p. 101 [6.10], 107 [3.1–4, 4.1a–4.2b, 4.4c–4.12d], p. 112 [5.1, 5.3–4], p. 213 [25.3c–4d, 26.1–2, 27.1, 28.1, 29.1, 31.3c–4b, 31.5ab, 31.5c–7b (with significant differences), 31.8c–9b, 3.10, 3.12], p. 218 [15.3a–4b, 16.1ab, 16.2a–3b, 17.1, 17.2ac (with differences), 17.3, 18.1–5, 19.1ab].


Haṭhapradīpikājyotsnā ad 3.100 [AS 7.8c–9d, 7.12, 7.5a–, 7.2cd] and 4.1 [AS 33.1, 32.3–4, 7.23, 34.1, 7.17, 7.20, 7.6c–7d, 7.16cd].


AS 6.7a–6.8b = Gorakṣaśataka 34a–35b. This verse is also found at Gheraṇḍasaṃhitā 5.62.


AS 7.16–20 ≈ Vivekamārtaṇḍa 90–92.


AS 16.1c–2b, 11.3, 11.9cd, 11.3cd, 14.6, 13.5cd, 13.7cd, 19.2 = Amaraughaprabodha 20, 29, 32cd, 37ab, 38, 39cd, 40ab, 45.


AS 3.1, 6.11ab, 11.4 = Gorakṣayogaśāstra 5, 13ab, 17.


AS 1.15b–1.16d, 1.17c–1.18b, 1.19ab, 3.1–4, 4.3–4, 11.1ab, 11.3cd, 11.4bc, 11.5ab, 11.6, 11.7cd, 12.6, 15.1, 16.1–3, 19.2 = ß 2.1b–2.2d, 2.3, 2.4ab, 2.6c–9, 2.11–12, 4.28cd, 4.27ab, 4.27dc, 4.28ab, 4.31, 4.34cb, 4.38, 5.13, 5.17c–5.20b, 3.31.


AS 11.3, 11.9cd, 19.2 = Haṭhapradīpikā 3.9, 3.13cd, 4.69.


This triad is mentioned at Niśvāsatattvasaṃhitā Nayasūtra 4.147 and in many subsequent tantric works.


This is a śleṣa: ākāśapatha can also mean the sun’s orbit in the sky.


On the kūṭāgāra, see below, p. 418.


The bright red javā flower (Hibiscus rosa-sinensis L.), popularly known as the China Rose, is common throughout south, southeast and east Asia.


Granthis are mentioned in many earlier Śaiva texts, some of whose lists include brahma, viṣṇu and rudra granthis but not in the Amṛtasiddhi’s configuration. See e.g. Kubjikāmatatantra 17.61–84, in which there are sixteen granthis and Netratantra 7.22–25, in which there are twelve.


On the four blisses see Isaacson and Sferra 2014, passim.


Sanderson 2009, 180 note 436.


Sanderson loc. cit.: “This substitution of initial ch- for s-/ś- is probably an east-Indianism.”


The Mahāmudrātilaka (draft edition of Staatsbibliothek zu Berlin Preuss. Kulturbesitz Orientabteilung Hs. or. 8711, folio 17 verso) locates the bodily Kāmarūpa between the eyebrows.


Although such usage is not found in pre-modern texts, to avoid confusion I use the word “Nāth” to refer to ascetics usually called yogīs or jogīs in texts and travellers’ reports and whose traditions, with some exceptions such as those which trace their lineages to Kānhapa or Kṛṣṇācārya, came, by the sixteenth century at the latest, to be grouped together in twelve panths or lineages. On the Nāth Saṃpradāya, see Mallinson 2011.


The historical context of this connection is explored in Mallinson 2019, in which the Konkan site of Kadri (in present-day Mangalore) is proposed as the location of the transition from Vajrayāna Buddhism to Nāth Śaivism evinced by the Amaraughaprabodha’s reworking of the teachings of the Amṛtasiddhi.


This edition was read with Professor Alexis Sanderson, Jason Birch, Péter-Dániel Szántó and Andrea Acri in Oxford in early 2012, all of whom I thank for their valuable emendations and suggestions.


Primary Sources

  • Amaraughaprabodha of Gorakṣanātha. K. Mallik, ed. The Siddha Siddhānta Paddhati and Other Works of Nath Yogis. Poona: Poona Oriental Book House, 1954.

  • Kadalīmañjunāthamāhātmyam. Śambhu Śarmā Kaḍava, ed. Kāśī: Gorakṣa Ṭilla Yoga Pracāriṇī, 1957.

  • Kubjikāmatatantra. T. Goudriaan and J.A. Schoterman, eds. The Kubjikāmatatantra: Kulālikāmnāya Version. Leiden: E.J. Brill, 1988.

  • Guhyasamājatantra. Yukei Matsunaga, ed. Osaka: Toho Shuppan, 1978.

  • Gorakṣayogaśāstra. National Archives of Kathmandu 5–332; Nepal-German Manuscript Preservation Project 24/44.

  • Gorakṣaśataka. Government Oriental Manuscripts Library, Madras, Ms. No. R 7874.

  • Gorakhbānī. P.D. Baḍathvāl, ed. Prayāg: Hindī Sāhity Sammelan, 1960.

  • Gheraṇḍasaṃhitā. J. Mallinson, ed., trans. New York:, 2004.

  • Chāndogya Upaniṣad. Patrick Olivelle, ed., trans. The Early Upaniṣads: Annotated Text and Translation. New York; Oxford: Oxford University Press, 1998.

  • Dattātreyayogaśāstra. Unpublished edition by James Mallinson.29

  • Navanāthacaritramu of Gauraṇa. K. Ramakrishnaiya, ed. Madras University Telegu Series, no. 7. Madras, 1937.

  • Niśvāsatattvasaṃhitā. Dominic Goodall, Harunaga Isaacson and Alexis Sanderson, eds. The Niśvāsatattvasaṃhitā: The Earliest Surviving Śaiva Tantra, volume 1. A critical edition and annotated translation of the Mūlasūtra, Uttarasūtra, and Nayasūtra. Collection Indologie, no. 128 (Early Tantra Series, no. 1.) Pondicherry: Institut Français d’ Indologie/École française d’ Extrême-Orient, 2015.

  • Netratantra. Madhusūdan Kaul Śāstrī, ed. The Netra Tantram, with Commentary by Kshemarāja. 2 vols. Kashmir Series of Texts and Studies, nos. 46, 61. Bombay: the Research Department of Jammu and Kashmir State, 1926, 1939.

  • Pañcakrama. Katsumi Mimaki and Toru Tomabechi, eds. Pañcakrama: Sanskrit and Tibetan Texts Critically Edited with Verse Index and Facsimile Edition of the Sanskrit Manuscripts. Tokyo: The Centre for East Asian Cultural Studies for Unesco, 1994.

  • Yogacintāmaṇi of Śivānandasarasvatī. Haridās Śarmā, ed. Calcutta: Calcutta Oriental Press, n.d.

  • Vivekamārtaṇḍa of Gorakṣadeva. Oriental Institute of Baroda Library, Acc. No. 4110.

  • Haṭhapradīpikā of Svātmārāma. Svāmī Digambarjī and Dr Pītambar Jhā, eds. Lonavla: Kaivalyadhām S.M.Y.M. Samiti, 1970.

  • Haṭhapradīpikājyotsnā of Brahmānanda. Svāmī Maheśānand, Dr. Bāburām Śarmā, Jñānaśaṃkar Sahāy, and Ravindranāth Bodhe, eds. Brahmānandakr̥tā Haṭhapradīpikā Jyotsnā: Ālocanātmaka Saṃskaraṇa (Hindī). Lonavla: Kaivalyadhām S.M.Y.M. Samiti, 2002.

Secondary Literature

  • Birch, Jason. 2011. “The Meaning of haṭha in Early Haṭhayoga.” Journal of the American Oriental Society 131 (4): 527–554.

  • Bühnemann, Gudrun. 2000. The Iconography of Hindu Tantric Deities, volume I. The Pantheon of the Mantramahodadhi. Groningen: Egbert Forsten.

    • Crossref
    • Export Citation
  • Gode, P.K. 1954. Studies in Indian Literary History, Vol. II. Bombay: Bharatiya Vidya Bhavan.

  • Isaacson, Harunaga and Francesco Sferra. 2014. The Sekanirdeśa of Maitreyanātha (Advayavajra) with the Sekanirdeśapañjikā of Rāmapāla. Critical Edition of the Sanskrit and Tibetan Texts with English Translation and Reproductions of the MSS. (With contributions by Klaus-Dieter Mathes and Marco Passavanti). Serie Orientale Roma fondata da Giuseppe Tucci Vol. CVII. Napoli: Universita degli Studi di Napoli “L’ Orientale.”

  • Mallinson, James. 2011. “Nāth Saṃpradāya.” In Brill’s Encyclopedia of Hinduism, Vol. 3, edited by Knut A. Jacobsen, 407–428. Leiden: Brill.

  • Mallinson, James. 2015. “Śāktism and Haṭhayoga.” In Goddess Traditions in Tantric Hinduism, edited by Bjarne Wernicke Olesen, 109–140. London: Routledge.

  • Mallinson, James. 2019. “Kālavañcana in the Konkan: How a Vajrayāna Haṭhayoga Tradition Cheated Buddhism’s Death in India.” Religions 10 (4), 273.

  • Reigle, David. 2012. “The Kālacakra Tantra on the Sādhana and Maṇḍala.” Journal of the Royal Asiatic Society Series 3, 22 (2): 439–463.

  • Rhys Davis, T.W. and William Stede. 1921–1925. Pali-English Dictionary. London: Pali Text Society.

    • Crossref
    • Export Citation
  • Sanderson, Alexis. 2009. “The Śaiva Age.” In Genesis and Development of Tantrism, edited by Shingo Einoo, 41–349. Tokyo: Institute of Oriental Culture, University of Tokyo.

  • Sastri, V.V. Ramana. 1956. “The Doctrinal Culture and Tradition of the Siddhas.” In The Cultural Heritage of India. Vol. 4: The Religions, edited by Haridas Bhattacharyya, 300–308. Calcutta: The Ramakrishna Mission Institute of Culture.

    • Crossref
    • Export Citation
  • Schaeffer, Kurtis R. 2002. “The Attainment of Immortality: From Nāthas in India to Buddhists in Tibet.” Journal of Indian Philosophy 30 (6): 515–533.

  • Collapse
  • Expand

Śaivism and the Tantric Traditions

Essays in Honour of Alexis G.J.S. Sanderson

Series:  Gonda Indological Studies, Volume: 22


All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 142 67 15
PDF Views & Downloads 172 67 11