Joining the Clubs: Continuities between Pāśupata and Nāth Asceticism1
James Mallinson*
Balliol College, University of Oxford
Abstract
This article adduces textual, epigraphic, and iconographic evidence to identify vestiges of the Pāśupata ascetic tradition in Nāth Yogī lineages. The Pampāmāhātmya reveals links between Pāśupata and Nāth teachings on yoga. A combination of all three varieties of sources identifies forerunners of the Nāth Yogī Dhaj Panth subdivision within the Pāśupatas. Further, Nāth iconography from its beginnings to the early modern period depicts yogis carrying clubs similar to those seen in material representations of the Pāśupata teacher Lakulīśa. Together this evidence indicates that the Pāśupatas were among the various tantric traditions that survived in modified forms in lineages of the Nāth Yogīs.
Keywords: yoga, asceticism, Hinduism
* Contact: jim.mallinson@ames.ox.ac.uk
© 2025 Mallinson. This open access article is distributed under a Creative Commons Attribution 4.0 License (https://creativecommons.org/licenses/by/4.0/)
Introduction
The Pāśupatas were the first Śaiva ascetic tradition in India, and are considered to have ceased to exist by the thirteenth century.2 Already in the middle of the nineteenth century, Horace Hayman Wilson observed that rather than dying out, the Pāśupatas “merged into other sects, and particularly into that of the Kānphaṭā Yogīs,” i.e. the loose collection of lineages now identified as constituting the Nāth Saṃpradāya.3 Wilson based his observation on the traditions’ sharing of common features of male Śaiva asceticism, and various subsequent authors have made the same observation.4
The shared features that constitute the evidence for these inferences include celibacy, frequenting cremation grounds, worshipping the śivaliṅga, wearing ash, jaṭā or earrings, and carrying accoutrements such as the triśūla. To these might be added certain locations that, having been Pāśupata strongholds, became home to Nāth Yogī5 lineages, such as Vijayanagara,6 Eklingji near Udaipur,7 Ujjain,8 Kalinjar,9 the Salt Lake region west of Jaipur,10 Karvan and Dabhoi,11 and Jalandhar.12
In the remainder of this article, I shall identify some more specific evidence for links between Pāśupatas and Nāths that I have stumbled across in the course of my research on the history of physical yoga and its practitioners. No doubt there is a great deal more to be found.
1 Gorakṣa in the Cintra Praśasti
A clear connection between the two traditions may be found in the Cintra Praśasti (so called because it was brought from India to Cintra in Portugal in the 18th century), an inscription from Somnath in Kathiawad dated 1287 CE that, among many other things, records the establishment of images of five deities, including one of Gorakṣa,13 the second human guru of the Nāths, by a Pāśupata ascetic called Tripurāntaka.14
2 Techniques of Haṭhayoga in the Pampāmāhātmya
Yoga is central to Pāśupata practice from its earliest teaching in the circa second-century CE Pāśupatasūtra.15 The distinctive feature of yoga as propounded by the Nāths is the use of complex methods of breath control and physical techniques for manipulating the vital energies that, from the fourteenth century onward, are classified as mudrās.16 These methods are first taught in the eleventh-century Amṛtasiddhi and so are not found in any early Pāśupata texts, such as the Pāśupatasūtra, Kauṇḍinya’s Pañcārthabhāṣya thereon, the Gaṇakārikā or the Pāśupatayogavidhi of the Skandapurāṇa. They are taught in perhaps the latest text to teach a specifically Pāśupata yoga, the Pampāmāhātmya, a glorification of the goddess Pampā, from whose name is derived that of the village of Hampi that lies close to the famous city of Vijayanagara. The date of the Pampāmāhātmya is uncertain; the Nāth yoga methods it includes under the rubric of mudrā indicate that the passage in which they are found postdates the twelfth-century Vivekamārtaṇḍa, from which they are clearly derived (see below). It is likely to be at least two centuries younger than the Vivekamārtaṇḍa, because it is not until the c.1400 CE Śivasaṃhitā and Haṭhapradīpikā that we find these methods collectively termed mudrās.17
The titles of some of the last six chapters of the Pampāmāhātmya (chapters 11–16 of the Uttarabhāga) indicate their Pāśupata nature: pañcārthacaraṇapraśaṃsā, pāśupatavratanirūpaṇam, śaivacaryā, mahāvratanirūpaṇam, kālāmukhapraśaṃsā, kālāmukhamatanirūpaṇam, aṣṭāṅgayoganirūpaṇam.18 The final chapter, on aṣṭāṅgayoga, opens with brief descriptions of five types of yoga (mantra, sparśa, bhāva, abhāva, mahā), followed by the aṣṭāṅgayoga itself (vv. 10–103) and then teachings on five physical mudrās (vv. 104–128).
The teaching of five types of yoga that precedes aṣṭāṅgayoga further indicates a link with Pāśupata traditions because a system of five yogas is mentioned in the Pāśupatayogavidhi of the Skandapurāṇa (179.36ab).19 There the five kinds are not specified, but groupings of five yogas with names the same as those of the Pampāmāhātmya are taught in two other Śaiva Purāṇas, the Śivapurāṇa (Vāyavīyasaṃhitā 2.29.5–13) and the Liṅgapurāṇa (2.55.7–8).
The five mudrās taught in the Pampāmāhātmya are identical to five yoga methods taught together in the Vivekamārtaṇḍa. The Pampāmāhātmya introduces them thus:
anantyaṃ khalu mudrāṇām ātmavidbhir viniścitam |
mukhyās tatra vadāmy atra pañca mudrāḥ prasaṅgataḥ ||105||
uḍḍīyāṇaṃ nabhomudrāṃ mahāmudrāṃ jalandharām |
mūlamudrāṃ ca yo vetti sa yogī sarvaveditā ||106||
(105) Experts in self-knowledge have determined an infinite number of mudrās.
In that regard I shall teach here the five most important among them.
(106) The yogi who knows Uḍḍīyāṇa, the sky mudrā, the great mudrā, Jalandharā, and the root mudrā knows everything.
The second verse is very close to verse 40 of the Vivekamārtaṇḍa:
mahāmudrāṃ nabhomudrām uḍḍīyāṇaṃ jalandharam|
mūlabandhaṃ ca yo vetti sa yogī muktibhājanam ||40||
The only noteworthy difference is that the Pampāmāhātmya has the highly unusual mūlamudrā for the Vivekamārtaṇḍa’s mūlabandha (the author of the Pampāmāhātmya turns all three of the Vivekamārtaṇḍa’s bandhas into mudrās).
The Pampāmāhātmya’s teachings on the individual practices are clearly calqued on those of the Vivekamārtaṇḍa. Here, for example, is the Vivekamārtaṇḍa’s teaching on what it calls jālandharabandha, “the Jālandhara Lock” (vv. 45–46):
badhnāti hi śirājālam adhogāminabhojalam |
tato jālandharo bandhaḥ kaṇṭhaduḥkhaughanāśanaḥ ||45||
jālandhare kṛte bandhe kaṇṭhasaṃkocalakṣaṇe|
na pīyūṣaṃ pataty agnau na ca vāyuḥ pradhāvati ||46||
(45) Because it blocks off the network of channels from which the liquid from the void flows down, it is [called] the Jālandhara Lock. It gets rid of all problems in the throat.
(46) When the Jālandhara Lock is performed, its defining feature being the constriction of the throat, nectar does not fall in the fire and the breath does not rush forth.
At 16.123a–125b the Pampāmāhātmya teaches the same practice as follows:
baddhvā sarvasirājālam adhogāminabhojalam |
kaṇṭhadeśe nibadhnīyān mudrā jālandharī smṛtā ||123||
jālandharīkṛtaṃ samyakkaṇṭhaduḥkhavināśanam |
nāgnau patati yat kiṃ cid api vyomnaḥ parāvṛtam ||124||
na cittaṃ dhāvati svairaṃ na ca vāyuḥ pradhāvati |
(123) When [the yogi] blocks off the entire network of channels from which the liquid from the void flows down and locks it in the region of the throat, it is called the Jālandharī mudrā.
(124) Jālandharī completely gets rid of problems of the throat. Nothing whatsoever falls in the fire, having returned from the void.20 The mind does not run about willfully, nor does the breath rush forth.
Again, there are no significant changes other than the author of the Pampāmāhātmya making Jālandhara a mudrā (and accordingly giving it the feminine name Jālandharī) rather than a bandha as it is said to be in the Vivekamārtaṇḍa.
The Vivekamārtaṇḍa is a seminal text of the physical yoga tradition. Its teachings are attributed to Gorakṣa and it is found in hundreds of manuscripts from all over the subcontinent and beyond, and in several different recensions, the best known of which is entitled Gorakṣaśataka. The author of the Pampāmāhātmya incorporated its teachings into Pāśupata yoga, indicating a confluence of Pāśupata and Nātha teachings in the medieval Deccan.
3 The Dhvajin
In the Pampāmāhātmya’s description of the Mahāvrata observance of the Pāśupata ascetic, we find the following practice described at 13.75–76:21
jāte paramavijñāne svairāvasthāṃ cared vaśī |
svairāvasthāṃ praviṣṭo ’sau dhvajīti parigaṇyate ||75||
dhvajam ekaṃ kare bibhran maruddhūtatadaṃśukam |
caled yatas tato gacched dhvajīti śivaśāsanam ||76||
When the supreme knowledge has arisen, the ascetic enters the “willful” stage. When he has entered the “willful” stage he is known as a “flag-bearer” (dhvajin).
According to the teaching of Śiva, the “flag-bearer,” carrying a flag in his hand, should go in whatever direction the cloth of that [flag] is blown by the wind.
An inscription from Omkareshwar dated 1063 CE mentions a Pāśupata called Gāndhadhvaja, perhaps referring to an ascetic of a lineage whose practices included this observance.22 Some subsequent Nāth lineages also added the suffix -dhvaja to their names and in the earliest known list of the panths or lineages of the Nāths, found in the 1570 CE Nujūm al-‘ulūm from Bijapur, one of the twelve panths is called dhaj,23 a vernacular (here Persian) form of the Sanskrit dhvaja. Briggs reports the inclusion of the Dhajjanāth panth among the twelve panths identified in 1924 as constituting the Nāth Saṃpradāya by the Mahant of Ṭillā, then the headquarters of the Nāths, but finds little evidence for its continued existence: he relies upon a terse notice in Panjab Notes and Queries from 1884 to add only that “[i]t is asserted that they are to be found in Ceylon.”24 This assertion may have arisen through confusion over the identity of the best-known Dhvajapanthin, Magaradhaja, an itinerant yogi who left graffiti in at least 41 pilgrimage sites during his travels across north India and the Deccan in the fifteenth century CE.25 The Dhvajapanth is no longer included in authoritative lists of Nāth panths, but there is at least one Dhvajapanthin institution still extant, the Makaradhvaja Bālājī temple at Beawar in Rajasthan26 (in the Salt Lakes region identified with Jambumārga in Cecil’s survey of early Pāśupata sites).27 Old Hindi poetry indicates that the siddha Haṇuvanta was the guru of Magaradhaja28 and at Beawar Haṇuvanta is revered as the founder of the Dhvaja Panth. The temple’s primary images, however, are of the deity Hanumān29 and Makaradhvaja, who, according to a post-Rāmāyaṇa legend followed at the temple, is Hanumān’s son.30 It is perhaps because of this identification that Briggs was told that the Dhajjapanth was extant only in Sri Lanka, which was famously visited by Hanumān. In light of the teachings of the Pampāmāhātmya, it seems reasonable to infer that Magaradhaja and other dhvajins of the 15th century31 were observing the Pāśupata vow described therein, journeying to pilgrimage sites in whatever direction the wind took them.
The material record other than their graffiti suggests that the number of such dhvajins has never been great. Despite the very large number of representations of yogis in sculpture and painting, I know of only two depictions of yogis carrying flags of the sort that a dhvajin might use.32 One is from the early seventeenth-century Gulshan Album compiled for Prince Salīm, the future Emperor Jahāngīr (fig. 1).33 The other is in a monumental illustrated manuscript of the Rāmcaritmānas produced in Jodhpur in about 1785 CE (fig. 2).
Figure 1. Gulshan Album, fol. 13b (detail). Staatsbibliothek, Berlin.
Note: This image is the property of the Staatsbibliothek zu Berlin- Preussischer Kulturbesitz, Orientabteilung and is published under a Public Domain Mark 1.0 license. The Gulshan Album has the shelf number A 117.
4 The Club
The final parallel between Pāśupatas and Nāths to which I shall draw attention is the most immediately striking, but also the least conclusive: the shared attribute of a club.
The use of a club is implicit in the name of the foremost Pāśupata teacher, Lakulīśa: the Sanskrit word lakula (more commonly found as laguḍa) means “club.” The name Lakulīśa and its cognates such as Laguḍi are attested only from the time of the Skandapurāṇa (c. 550–650 CE) onwards, but the earliest Pāśupata image, which is on a pilaster from Mathurā dating to 380 CE, depicts a three-eyed deity, perhaps Śiva as Caṇḍa, holding a staff in his right hand.34 The staff is straight, not tapering like the clubs found in subsequent depictions of Lakulīśa, but is held in a similar way to them. A large number of statues of Lakulīśa holding a club were produced over the course of the rest of the first millennium. They are found from Mathurā and Kanauj and Odisha to as far south as Badami in Karnataka.35 The following example (fig. 3) is from Kalinjar.
Figure 2. Rāmcaritmānas (detail). Meharangarh Museum Trust, c. 1775 CE.
In modern depictions of Lakulīśa at his temple at his place of incarnation, Karohan, his club is a short straight stick. I do not know of other examples of him carrying a straight stick from the period between the 380 CE Mathurā pilaster and this modern image (fig. 4).36
Figure 3. Lakulīśa with club and disciples, Kalinjar, c. 9th century CE. Photo by author, December 2023.
Figure 4. Contemporary sign at Lakulīśa temple, Karohan. Photo by author, March 2016.
Nāth Yogīs are rarely associated with the carrying of clubs in recent centuries,37 and I know of only one report of their carrying them today38—yet in their early iconography, dating from the thirteenth century onwards, they are often depicted carrying a range of clubs and shorter straight sticks. Bakker notes that the Amarakośa gives laguḍa as a synonym of daṇḍa, the usual term for an ascetic’s stick or staff, and textual typographies of Nāth insignia also often use the term daṇḍa for the stick carried by a yogi, but yogis are almost never depicted carrying the long staffs typical of Brahmanical ascetics.39
4.1 Panhale Kaji
Among the earliest such images is one of Gorakṣa at Panhale Kaji in southern Maharashtra (figs. 5 and 6).40 This stone stele is damaged and found among other statue fragments in an otherwise unused and unadorned cave. Its large size (63 cm in height) relative to other statues at the site suggests that it was originally an object of worship, and Deshpande (1986, 64) convincingly proposes that it was originally installed in cave 14. The figure may be identified as Gorakṣa from the cows adorning its pedestal. The object in his left hand is likely to be a horn, which in various forms has been a key identifier of Nāth Yogīs to this day. The stick in his right hand is damaged at its top, so could have been longer and perhaps something other than a simple stick, but comparison with the other two images of yogis carrying clubs from Panhale Kaji (shown below) and with Matsyendra at Kadri and a yogi at Dabhoi (also shown below) suggests that it is a short staff.
Outside cave 14 are images of 12 siddhas, including one carrying a club (fig. 7). He is in a similar standing position to the larger image of Gorakṣa, so it is tempting to identify him as Gorakṣa as well, but this is unlikely if, as suggested by Deshpande, the primary image inside the cave was of Gorakṣa.41
In the hamlet of Gaur Lena, about 1.5 kilometers from the main group of caves at Panhale Kaji, is a cave that contains eroded images of 84 siddhas, including two carrying a club or short stick (figs. 8 and 9). One is likely to be Gorakṣa, as he appears to be accompanied by a cow and has a similar stance to the Gorakṣa found in cave 14 at the main site.42
Figures 5 and 6. Gorakṣa torso and pedestal, Panhale Kaji. Photos by author, December 2023.
4.2 Kalinjar
At Kalinjar, in the Vindhya range, there is a rock-cut image of Gorakṣa (fig. 10) in which he is leaning on a staff similar in length to a Brahmanical ascetic’s daṇḍa, but which has the ridged rounded ends typical of those found in the other early depictions of Gorakṣa noted in this article. The inscription accompanying the image is dated 1279 CE.43
4.3 Kadri
At Kadri, in Mangalore—where in perhaps the twelfth century a Vajrayāna Buddhist tradition gave way to, or was absorbed into, a Nāth tradition44—there are several images of uncertain age of Nāth Yogīs carrying short sticks or clubs. Perhaps the oldest is a broken Matsyendra now found in the Mangalore Government Museum (fig. 11). While some features of this statue, such as the icon on its jaṭāmukuṭa, appear to be in imitation of an older image of Lokeśvara at the Kadri Mañjunātha temple, the straight club with widening ends held in Matsyendra’s left hand is distinctive.
The replacement image of Matsyendra in the Kadri monastery (fig. 12) is clearly in imitation of the earlier image, but the club now has an overall taper.
Figure 7. Yogi outside Panhale Kaji cave 14. Photo by author, March 2016.
Figure 8. Gorakṣa in Panhale Kaji cave 29 (Gaur Lena). Photo by author, March 2016.
Figure 9. Yogi in Panhale Kaji cave 29 (Gaur Lena). Photo by author, March 2016.
Images at Kadri of the two other yogis most important in the early Nāth tradition, Gorakṣa (fig. 13) and Cauraṅgi (fig. 14), which appear to have been created at the same time as the newer Matsyendra, also depict them carrying clubs, both with a slight taper. Gorakṣa’s is unusual. Foregrounded in the center of the image, it is very long, three-quarters as tall as he is, and has a certain phallic symbolism.
4.4 Dabhoi
The Mahudi Gate at Dabhoi, which was decorated between 1230 and 1240 CE, has depictions of 12 primary siddhas, including Matsyendra, Gorakṣa, and Cauraṅgi, but they are damaged to the extent that it is impossible to determine whether they originally carried clubs. One of the 84 yogis depicted in the gate’s upper brackets, however, is carrying a club with similar features to that carried by the Panhale Kaji Gorakṣa (fig.15).
4.5 Vijayanagara
There are hundreds of depictions of yogis on the columns of the various fifteenth- and early sixteenth-century temples at Vijayanagara, several of which show them holding, or resting or balancing on, short sticks with widening ends (figs. 16–19).45
Figure 10. Gorakṣa at Kalinjar. 1279 CE. Photo by author, December 2023.
Figure 11. Matsyendra, Mangalore Government Museum. Photo by author, March 2012.
Figure 12. Matsyendra, Kadri Monastery. Photo by author, March 2012.
This preponderance of clubs or short staffs in early statuary depicting Nāth Yogīs is succeeded by a marked absence in the hundreds of depictions of Nāths in Mughal miniature painting. It appears that their use almost died out, but there are a small number of indications that it did not disappear completely.
An image of Gorakṣa at his temple in Odadar, Gujarat, of uncertain age but not of great antiquity, depicts him in a seated posture resting his hands on a short straight stick.46
Figure 13. Gorakṣa, Kadri Monastery. Photo by author, March 2016.
Figure 14. Cauraṅgi, Kadri Monastery. Photo by author, March 2016.
An early nineteenth-century painting of Guru Nānak with a group of Nāth Yogīs shows one of them carrying a short club-like stick (fig. 20).
Drawing on his observations in India in the first half of the nineteenth century, Wilson reports that a group of Śaiva ascetics called Sūkharas, who display some similarities to Nāth Yogīs, distinguish themselves by carrying “a stick three spans in length.”47
Finally, Briggs, reporting on a visit he made in 1924 to the Satnāthi establishment at Puri, says that the Mahant there carried a “‘club,’ made of straw and covered with cloth, called sudarṣan [sic].”48
The reasons for the carrying of these clubs are very rarely explicit. Scholars have made a range of complementary proposals about that of Lakulīśa. Bisschop suggests that it may have developed from early iconography of Śiva, itself perhaps connected to that of the Kuśāṇa deity Oésho.49 Bakker argues that the phallic nature of the clubs carried by Pāśupatas and Heracles are intended to convey virility and, in the Pāśupata case, an association with Śiva, his emblem being the phallic liṅga.50 Nandi asserts that Lakulīśa’s club indicates a hostile attitude toward other sects and that differences in its prominence in his iconography reflect differences in the threat posed to Pāśupatas by other religions: held to the fore, it indicates increased rivalry.51 Cecil suggests that the clubs sported by Śiva and Lakulīśa were symbolic of their roles as protectors of their devotees.52 In the Skandapurāṇa,53 Lakulīśa is said to have picked up a firebrand before initiating his pupils. Bisschop has tentatively suggested that this indicates that the firebrand was used in the initiation and that this role may have been taken on by the club.54
Figure 15. Yogi, Mahudi Gate, Dabhoi. Photo by Rafique Shaikh, January 2016.
Figure 16. Matsyendra, Virūpākṣa Temple, Hampi. Photo by author, March 2016.
Figure 17. Yogi with short staff, Hazārarāma Temple, Hampi. Photo by author, March 2016.
Figure 18. Yogi with short staff, Paṭṭābhirāma Temple, Hampi. Photo by author, March 2016.
Figure 19. Yogi balancing on short staff, Viṭṭhala Temple, Hampi. Photo by author, March 2016.
Figure 20. Guru Nānak debates with a group of yogis (detail). Guler, early 19th century. Toor Collection.
A significant difference between the use of clubs in the two traditions is that Pāśupata ascetics are not depicted as carrying them, and their use is not prescribed in any Pāśupata texts.55 Meanwhile both the semi-divine siddhas of the Nāth tradition, such as Matsyendra and Gorakṣa, and run-of-the-mill Nāth ascetics are depicted with a variety of clubs and short staffs.
Within the Nāth tradition, clubs and sticks have had a variety of uses, including fighting. In the Śrīnāthatīrthāvalī, a Nāth pilgrimage guide commissioned by Maharaja Man Singh of Jodhpur in the early nineteenth century, at two of the 97 sites described there are said to be images of Gorakṣa carrying a mudgara.56 This weapon, which is often listed as an attribute of Śaiva and Śākta deities,57 is usually understood to mean some sort of hammer or mallet.58 It is sometimes differentiated from a musala, a club,59 but in current vernaculars, mudgar is used to describe the clubs popularly used by wrestlers for training and occasionally by ascetics of various orders. A large wooden mudgar on display at the temple of Hiṅglāj Mātā in Siddhpur in Gujarat (fig. 21) is said to have been used in the mid-twentieth century by the Dasnāmī Saṃnyāsī yogi and wrestler Guru Mastrām to see off ghosts at night.
At the 2025 Kumbh Melā I saw two Vairāgī sādhus carrying large clubs of the kind used by wrestlers for training (fig. 22).
It seems likely, then, that the mudgaras of Gorakṣa referred to in the Śrīnāthatīrthāvalī are clubs or sticks of the kind found in the images of him shown above. A text called Gorakṣamudgara is found in a manuscript in the National Archives, Kathmandu, which may be dated to approximately the fifteenth century on palaeographic grounds.60 The text itself makes no mention of Gorakṣa, and the word mudgara occurs in it only once, when people in hell are said to be mudgaracūrṇitāḥ,61 “pounded by mudgaras,” suggesting that here too a mudgara is a club and that the text’s name means “The Club of Gorakṣa.” This is appropriate: the text is a polemic against worldly existence and a paean to the renunciate life, so one may understand its title to refer to Gorakṣa figuratively smashing saṃsāra.
Figure 21. The mudgar of Guru Mastrām at Hiṅglāj Mātā Mandir, Siddhpur. Photo by author, 2016.
Figure 22. Vairāgī sādhus carrying clubs at the Prayagraj Kumbh Melā 2025. Photo by author.
Nāth Yogīs have also used their sticks or clubs for purposes that we can rule out for Lakulīśa for historical reasons. Lakulīśa may have rested his chin or arm on a club in order to meditate, as do some of the yogis portrayed at Vijayanagar, but it is unlikely that he would have used one for balance in more difficult postures, because there is no evidence for yogis using balancing postures prior to about the eleventh century.62 Similarly, in images of yogis from 1600 onwards, they are often to be seen using clubs to grind bhāṅg, cannabis, into a paste before imbibing it, as seen in the Mughal-era example below (fig. 23). The use of cannabis in this way in India is not attested until the second millennium CE.63
One further use of the yogi’s stick or club is explained in commentaries on Jāyasī’s Padmāvat. In the late 19th century, Sudhākar Dvivedī glosses a mention of a yogi’s ḍaṇḍa with soṁṭā, adding that “yogis keep [it] for performing magic; it is one or one-and-a-half hand-spans in length like a black ‘ruler’ (a long, straight, round, wooden stick for marking straight lines). Some yogis call this the rod (soṁṭā) of Bhairav Nāth, others that of Gorakh Nāth.”64 In a more recent commentary, Agravāl, remarking on the same reference, says that the daṇḍā is “a small stick made of ebony by waving which yogis display miracles.”65
5 Concluding Remarks
Of the four distinct indicators of continuity between the Pāśupatas and Nāths presented in this article, the first three demonstrate conclusively that the Nāth Yogīs, or at least their predecessors before the order became formalized in the second half of the second millennium CE, included lineages that were direct heirs to the Pāśupatas. Not all Pāśupata lineages fed exclusively into the Nāths, nor are all Nāths heirs to the Pāśupatas. The Kālamukhas, for example, who developed from the Pāśupatas, had a formative influence on the Vīraśaivas of the Deccan, a religious tradition distinct from that of the Nāths.66 Meanwhile the Nāths became home to a wide range of tantric practitioners, including many who were altogether distinct from Pāśupatas, such as Vajrayāna Buddhists or Gāruḍika snake magicians. Notably many Nāth and Siddha traditions were open to men and women from a wide range of castes, unlike the Pāśupatas, who accepted only male Brahmins.
Why the Pāśupata name disappeared is unclear, but the transition from Pāśupata to Nāth coincides with a new dispensation in Indian asceticism, marked in particular by the adoption of new methods of physical yoga, of which the Nāths were among the foremost practitioners, and the concomitant appearance and rise to prominence of the Siddha traditions that flourished in South Asia and Tibet from the eleventh century onwards and became part of the Nāth lineages. At the same time the dominance of tantra as a state religion waned rapidly, most probably as a result of the turbulence caused by Islamic incursion. The tantric mantle was taken up by monastic traditions, which flourished during this period. Formal Pāśupata lineages faded away, but certain distinctive aspects of their doctrine and practice survived among the fledgling Nāths.
Figure 23. A gathering of ascetics preparing bhāṅg, 1600–1610, with later additions. Unknown artist. Source: Topsfield 2004, plate 198.
6 Works Consulted
6.1 Primary Sources
The Amṛtasiddhi and Amṛtasiddhimūla: The Earliest Texts of the Haṭhayoga Tradition, eds. James Mallinson and Péter-Dániel Szántó. Collection Indologie 150. Haṭha Yoga Series 2. Pondicherry: Institut Français de Pondichéry; Paris: École française d’Extrême-Orient. 2021.
Āpastambadharmasūtra, ed. G. Bühler. Bombay Sanskrit Series Nos. LIV and L, 3rd ed. Poona: Bhandarkar Oriental Research Institute. 1932.
Gorakṣamudgara, National Archives Kathmandu ms no. 5–332 (= Nepal-German Manuscript Preservation Project B 24/6).
Padmāvat = The Padumawāti of Malik Muḥammad Jaisi, Edited, with a Commentary, Translation, and Critical Notes, by G. A. Grierson and Mahāmāhopādhyāya Sudhākara Dvivedī. Calcutta: Baptist Mission Press. 1896.
Pampāmāhātmya, ed. K. V. Śāstri as Skāndapurāṇantargata Hēmakūtạ khaṇḍātmaka Saptarṣi Yātra Prakāśaka Pampā Māhātmyamu. Ballari. 1933.
Pāśupatasūtra = Pasupata Sutras with Pancharthabhashya of Kaundinya, ed. R. Ananthakrishna Sastri. Trivandrum Sanskrit Series 143. Trivandrum: University of Travancore. 1940.
Liṅgapurāṇa. Gurumandal Series 15. Calcutta. 1960.
Vāyavīyasaṃhitā, ed. Christèle Barois. In La Vāyavīyasaṃhitā: Doctrine et rituels śivaïtes en contexte purāṇique, vol. 3. PhD diss., École Pratique des Hautes Études. 2012.
Vivekamārtaṇḍa, forthcoming edition by James Mallinson based primarily on Oriental Institute of Baroda Library acc. no. 4110.
Śrīnāthatīrthāvalī, ed. Yogī Śaṅkarnāth Jī. Calcutta: United Commercial Press. 1950.
Siddhasiddhāntapaddhati of Gorakṣanātha, ed. M. L. Gharote and G. K. Pai. Lonavla: Lonavla Yoga Institute. 2005.
Skandapurāṇa Pāśupatayogavidhi. Draft edition by Yuko Yokochi.
6.2 Secondary Sources
Acharya, Diwakar. 2005. “The Role of Caṇḍa in the Early History of the Pāśupata Cult and the Image on the Mathurā Pillar Dated Year 61.” Indo-Iranian Journal 48: 207–222. https://doi.org/10.1007/s10783-005-2197-8
Acharya, Diwakar. 2011. “Pāśupatas.” In Brill Encyclopedia of Hinduism, vol. 3, edited by Knut A. Jacobsen, 458–466. Leiden: Brill.
Acri, Andrea. 2022. “Pāśupatas.” Hinduism and Tribal Religions (Encyclopedia of Indian Religions), edited by J. D. Long, R. D. Sherma, P. Jain, and M. Khanna, 1124–1129. Dordrecht: Springer Netherlands.
Agravāla, Vāsudeva Śaraṇa. 1998. Padmāvata, Malika Muhammad Jāyasī kṛta mahākāvya. Jhānsī: Sāhitya Sadan.
Bakker, Hans. 2019. “Origin and Spread of the Pāśupata Movement: About Heracles, Lakulīśa and Symbols of Masculinity.” In Holy Ground: Where Art and Text Meet: Studies in the Cultural History of India, 553–565. Leiden: Brill. https://doi.org/10.1163/9789004412071_030
Bankar, Amol. 2022. “The Mystery of the Wandering Jogīs from 15th Century India.” Pushpam 5: 34–41.
Beal, Samuel. 1884. Si-Yu-Ki: Buddhist Records of the Western World Translated from the Chinese of Hiuen Tsiang (A.D. 629). 2 vols. London: Trübner & Co.
Bhandarkar, D. R. 1931. “Mathura Pillar Inscription of Chandragupta II: G.E. 61.” In Epigraphia Indica, vol. XXI, 1–9. Delhi: Government of India.
Bisschop, Peter. 2009. “Śiva.” In Brill Encyclopedia of Hinduism, vol. 1, 741–754. Leiden: Brill.
Bisschop, Peter. 2020. “From Mantramārga Back to Atimārga: Atimārga as a Self-referential Term.” In Śaivism and the Tantric Traditions, edited by S. Hatley, D. Goodall, H. Isaacson and S. Raman, 15–32. Leiden: Brill.
Brick, David. 2012. “The Origin of the Khaṭvāṅga Staff.” Journal of the American Oriental Society 132, no. 1: 31–39. https://doi.org/10.7817/jameroriesoci.132.1.0031
Briggs, George Weston. 1989 [1938]. Gorakhnāth and the Kānphaṭa Yogīs. Delhi: Motilal Banarsidass.
Bühler, G. 1892. “The Cintra Praśasti of the reign of Sarangadeva.” In Epigraphia Indica, vol. 1, 271–287. Calcutta: Archaeological Survey of India.
Cecil, Elizabeth. 2014. “Seeking the Lord with a Club: Locating Lakulīśa in the Early History of Pāśupata Śaivism (Sixth to Ninth century CE).” South Asian Studies 30, no. 2: 142–158. https://doi.org/10.1080/02666030.2014.962308
Cecil, Elizabeth. 2020. Mapping the Pāśupata Landscape: Narrative, Place, and the Śaiva Imaginary in Early Medieval North India. Gonda Indological Studies. Leiden: Brill.
Chakravarti, N. P. 1948. “A Note on the Halayudha Stotra in the Amaresvara Temple.” In Epigraphia Indica, vol. XXV, 185. Calcutta: Government of India.
Derrett, J. Duncan M. 1974. “Modes of Sannyāsīs and the Reform of a South Indian Maṭha Carried out in 1584.” Journal of the American Oriental Society 94, no. 1: 65–72. https://doi.org/10.2307/599731
Deshpande, M. N. 1986. The Caves of Panhāle-Kāji (Ancient Pranālaka). New Delhi: Archaeological Survey of India.
Diamond, Debra, Catherine Glynn, and Karni Singh Jasol. 2008. Garden and Cosmos. The Royal Paintings of Jodhpur. Washington: Smithsonian/Arthur M. Sackler Gallery.
Dvivedī, Sudhākar. 1896. Sudhākaracandrikā, commentary on the Padmāvat (q.v.).
Flatt, Emma J. 2011. “The Authorship and Significance of the Nujūm al-‘ulūm: A Sixteenth-century Astrological Encyclopaedia from Bijapur.” Journal of the American Oriental Society 131, no. 2: 23–24. https://www.jstor.org/stable/23044643
Hatley, Shaman. 2018. The Brahmayāmalatantra or Picumata, Volume I, Chapters 1–2, 39–40 and 83. Revelation, Ritual, and Material Culture in an Early Saiva Tantra. Collection Indologie 133. Early Tantra Series 5. Pondicherry: École française d’Extreme-Orient.
Ingalls, Daniel H. H. 1962. “Cynics and Pāśupatas: The Seeking of Dishonor.” Harvard Theological Review 55, no. 4: 281–298. https://doi.org/10.1017/S001781600000794X
Kumar, Vijay, and Ranjan. 2021. “Inscriptions of the Kalinjar Fort, Kalinjar District Banda, Uttar Pradesh.” Indian Journal of Archaeology 6, no. 2: 613–1280.
Ladrech, Karin. 2002. “Bhairava à la massue.” Bulletin d’Études Indiennes 20, no. 1: 163–192. HAL Id: hal-03906240, version 1.
Lyons, Tryna. 1999. “The Changing Faces of Ekliṅgjī: A Dynastic Shrine and its Artists.” Artibus Asiae 58, nos. 3–4: 253–271. https://doi.org/10.2307/3250019
Mallinson, James. 2011. “Nāth Saṃpradāya.” In the Brill Encyclopedia of Hinduism, vol. 3, edited by Knut A. Jacobsen, 407–428. Leiden: Brill.
Mallinson, James. 2019. “Kālavañcana in the Konkan: How a Vajrayāna Haṭhayoga Tradition Cheated Buddhism’s Death in India.” Religions 10, no. 273: 1–33. https://doi.org/10.3390/rel10040273
Mallinson, James. 2020. “Haṭhayoga’s Early History: from Vajrayāna Sexual Restraint to Universal Somatic Soteriology.” In Hindu Practice, edited by Gavin Flood, 177–199. Oxford: Oxford University Press. https://doi.org/10.1093/oso/9780198733508.003.0008
Mallinson, James. 2022. “Nath Yogis and their ‘Amazing Apparel’ in Early Material and Textual Sources.” In Objects, Images, Stories: Simon Digby’s Historical Method, edited by Francesca Orsini, 65–95. London: Oxford University Press.
Murphy, James Cavanah. 1795. Travels in Portugal: Through the Provinces of Entre Douro E Minho, Beira, Estremadura, and Alem-tejo, in the Years 1789 and 1790, Consisting of Observations on the Manners, Customs, Trade, Public Buildings, Arts, Antiquities, &c. of that Kingdom. London: Strahan & Cadell.
Nandi, R. N. 1977. “Origin of Śaivite Monasticism: the Case of the Kālamukhas.” In Indian Society: Historical Probings in Memory of D.D. Kosambi, edited by R. S. Sharma, 190–201. New Delhi: People’s Publishing House.
Nandi, R. N. 1986. Social Roots of Religion in Ancient India. Delhi/Calcutta: K. P. Bagchi & Company.
Newid, Mehr-Ali. 1986. Waffen und Rüstungen im Alten Indien Dargestellt aufgrund der Quellen in Literatur, Kunst und Archälogie. lnaugural-Dissertation zur Erlangung des Doktorgrades der Philosophischen Fakultät der Ludwig-Maximilians-Universität zu München.
Powell, Seth. 2023. “Yogi Sculptures: Complex Āsanas Across the Deccan.” In “Yoga and the Traditional Physical Practices of South Asia,” edited by Daniela Bevilacqua and Mark Singleton. Special issue, Journal of Yoga Studies 4: 85–111. https://journalofyogastudies.org/index.php/JoYS/article/view/JoYS.2023.V4.02
Sarde, Vijay. 2019. “Archaeological Investigations of the Nātha Sampradāya in Maharashtra (c. 12th to 15th Century CE).” PhD diss., Deccan College Post Graduate and Research Institute (Deemed University), Pune.
Shah, U. P. 1957. “Nāth Siddhoṃ kī Prācīn Śilpamūrtiyāṃ.” Nāgarīpracāriṇī Patrikā varṣa 62: 174–207.
Shanthamurthy, Shubha. 2015. “An Epigraphical Account of the Śaiva Institutions of Baḷḷigāve (11cAD–12cAD).” DPhil. diss., University of Oxford.
Topsfield, Andrew. 2004. In the Realm of Gods and Kings: Arts of India. London: Philip Wilson Publishers.
Wilson, Horace Hayman. 1861. A Sketch of the Religious Sects of the Hindus. London: Trübner & Co.
Wujastyk, Dominik. 2002. “Cannabis in Traditional Indian Herbal Medicine.” In Ayurveda at the Crossroads of Care and Cure. Proceedings of the Indo-European Seminar on Ayurveda held at Arrábida, Portugal, in November 2001, edited by Ana Salema, 45–73. Lisbon: Centro de História de Além-Mar.
1 I first presented on this topic in a paper at a conference at All Souls, Oxford, held in May 2024 in honor of HRH Princess Maha Chakri Sirindhorn of Thailand. I thank the organizer of that conference, Diwakar Acharya, for inviting me to present at it, and for the perceptive comments I received from Her Highness and Pongsit Pangsrivongse. I presented another version of the paper at the Yoga Darśana Yoga Sādhana conference held at the University of Hamburg later in May 2024, and I thank Lubomir Ondračka for his comments there and on a subsequent draft. Peter Bisschop’s comments and corrections to the various drafts I have presented for publication since have been particularly helpful and I thank him for them. In addition to other very useful remarks, Vijay Sarde pointed out important features of the iconography of the yogis depicted at Panhale Kaji that I had missed, for which I am very grateful. Amol Bankar gave me some helpful comments on a draft of the article, for which I thank him. Thanks too to Davinder Toor of the Toor Collection, John Guy at the Metropolitan Museum, New York, Mahendra Singh Tanwar of the Mehrangarh Museum, Jodhpur, and Kristina Münchow at the Staatsbibliothek, Berlin, for their help with obtaining images and permissions for reproducing them.
2 Acharya 2011, 465. Acharya’s encyclopedia entry gives a brief overview of the Pāśupatas. A more up-to-date summary may be found in Acri 2020. References to Pāśupatas are found after the thirteenth century, such as in a list of ascetic lineages in the seventeenth- or eighteenth-century Siddhasiddhāntapaddhati (6.41), and the pāśupata vrata continued to be observed (leading some scholars to identify its adherents as Pāśupatas, e.g. Derrett 1974), but there is no firm evidence for their existence (or that of their subdivisions, e.g. Lākulas and Mausulas, Vaimalas and Kārukas) as a formal order after the thirteenth century.
3 For a brief overview of the Nāth Saṃpradāya, see Mallinson 2011.
4 E.g. Briggs (1938, 218–227), Eliade (1958, 303), and White (1996:97). Ghurye (1953, 145) identifies the Kāpālika tradition, which is derivative of that of the Pāśupatas (Sanderson 2006, Törzsök 2011), as the forerunner of the Nāths, and Lorenzen (1991, 35) similarly notes practices shared by Kāpālikas and Nāths.
5 I use “Nāth Yogī” to refer to a member of one of the various Nāth panths or their forerunners and “yogi” as a generic term for a practitioner of yoga.
6 See below on the Pāśupata Pampāmāhātmya from Hampi and depictions of Nāth Yogīs on temples from Vijayanagara.
7 Lyons 1999, 254n8. See also White 1996, 120–121 on legends connecting the two traditions at Eklingji.
8 After incarnating at Kāyāvataraṇa, Śiva is said to have walked to Ujjain and there initiated Kuśika as his disciple (e.g. Kauṇḍinya ad Pāśupatasūtra 1.1), and Xuanzang identifies Pāśupatas as being the dominant religious sect in Mālavā (Beal 1884, 260–261). The cave of Bhartṛhari in Ujjain remains an important Nāth site to this day.
9 Kalinjar has many Pāśupata inscriptions and images of Lakulīśa with his disciples (see e.g. fig. 3) as well as thirteenth-century images of Gorakṣa (see e.g. fig. 10) and Matsyendra, together with inscriptions.
10 This is the likely location of Jambumārga, a region said to be important for Pāśupatas in the Skandapurāṇa (on which see Cecil 2020, ch. 3) and where today there are numerous Nāth ashrams.
11 Kāyāvataraṇa, where Śiva is said by Kauṇḍinya to have incarnated, is modern-day Karohan or Karvan, which is close to the fortified city of Darbhavatī, modern-day Dabhoi, whose northern gate was decorated with Nāth images c.1230 CE (Shah 1957).
12 Xuanzang mentions Jalandhar as a region inhabited by Pāśupatas (Lorenzen 1991, 182). Today it is home to the important Nāth monastery of Jwalamukhi.
13 Here called Gorakṣaka, perhaps metri causa, although this form of the name is often found in prose works (I thank Lubomír Ondračka for this observation).
14 The inscription is reproduced by Murphy (1798, plate xiii) and edited, together with a summary of its contents, by Bühler (1892, 284). The verse mentioning Gorakṣa is as follows:
gorakṣakaṃ bhairavam āñjaneyaṃ sarasvatīṃ siddhivināyakaṃ ca |
cakāra pañcāyatanāntarāle bālendumaulisthitamānaso yaḥ ||45||
15 On this date see Bhandarkar 1931, 7.
16 On the early history of physical methods of yoga, see Mallinson 2020.
17 This passage comes at the very end of the text, before three closing verses, and is not alluded to in the chapter’s introductory section, so could be a later addition; however, elements of the aṣṭāṅgayoga teaching that precedes it, in particular details of the locations for meditation, also have parallels with the Vivekamārtaṇḍa.
18 Sanderson (2014, 10–11) notes that chapters 11 and 12 of the Uttarabhāga of the Pampāmāhātmya include teachings derivative of those of three Pāśupata texts, the Pañcārthabhāṣya, Gaṇakārikā, and Prāyaścittavidhi of Gārgya.
19 I thank Yuko Yokochi for sharing with me her draft edition of the Pāśupatayogavidhi.
20 I am unsure of the meaning of vyomnaḥ parāvṛtam, and in my tentative translation have understood vyoman, “the void,” to refer to the space in the central channel.
21 I thank Anthony Evensen for providing me with this reference.
22 Charkravarti 1948. See also Bisschop 2020, 28–29.
23 Flatt 2011, 242. In this article I use dhvaja, dhaja/dhaj, and dhajja as found in the sources I refer to.
24 Briggs 1938, 68–69.
25 Several scholars have recorded Magaradhaja’s graffiti. For a summary and lucid overview, see Bankar 2022, who also notes references to the Dhajjapanth in Old Hindi literature.
26 As of December 2024, the mahant of the temple, Dr. Prakash Nath, was active on Facebook, at https://www.facebook.com/mahantprakashnath.
27 See footnote 8.
28 In the Haṇavant Jī kī Sabadī, Magaradhaja asks Haṇavanta a question (2.5 / 742, p.72).
29 In second-millennium iconography of Hanumān, he is very often depicted carrying a flag (cf. the fifth verse of the Old Hindi Hanumān Cālīsā: hātha vajra aur dhvajā birājai, kāṃdhe mūñja janeū sājai).
30 Ghoḍācolī identifies Haṇuvanta as Hanumān (p. 12, v. 136 = v.11 in Ghoḍā Colī Jī kī Sabadī). He also gives an esoteric understanding of dhaja as referring to stretching the breath up to the sky, i.e. the void in the head (p. 12, v. 133 = v. 8 in Ghoḍā Colī Jī kī Sabadī).
31 Bankar (2022) also notes graffiti from that period made by the yogis Ratanadhaja and Acyantadhaja.
32 There are numerous depictions of yogis carrying short staffs topped with round peacock-feather trimmed banners, but these are likely to be fans (which are also used for giving blessings) and would not be suitable for indicating wind direction. I thank HRH Princess Maha Chakri Sirindhorn for this observation.
33 Staatsbibliothek Gulshan Album, fol. 13b.
34 On the identity of this figure, see Acharya 2005.
35 See e.g. Nandi 1986, 102–105; Bakker 2019; and Cecil 2014.
36 The khaṭvāṅga, a straight staff topped with a skull that is a characteristic accoutrement of Bhairava and Kāpālika ascetics, is of a similar size and shape to Pāśupata clubs and may, as Hatley suggests (2018, 290–292), be an “inflection” of the Pāśupata club, but it has older origins: it is listed as one of the insignia to be taken up by the Brahmin-killer in the Āpastambadharmasūtra (1.28.21–29.1), where it is most probably a simple bedpost. By the time of the Baudhāyanadharmasūtra (second to first century BCE), the khaṭvāṅga is topped with a skull (Brick 2012). Kāpālikas, or rather traditions derivative of them such as that of the Aghorīs, have also been included under the aegis of the Nāths, usually associated with lineages derived from Kānha and Jālandhara.
37 A few exceptions are noted below.
38 Vijay Sarde saw a Nāth Yogī carrying a club at the 2016 Siṃhasth Kumbh Melā in Nasik (email communication, January 21, 2025).
39 See Mallinson 2022 for several examples of daṇḍa being used in vernacular sources to denote a typical accoutrement of the yogi, especially p. 77n24.
40 See Sarde 2019, figs. 4.79, 4.82 for two other early images from Maharashtra of Nāth Yogīs carrying clubs or short sticks.
41 I thank Vijay Sarde for this observation.
42 I thank Vijay Sarde for this observation.
43 Kumar & Ranjan 2021, 741. I thank Peter Bisschop for drawing my attention to this image and inscription.
44 Mallinson 2019.
45 See Powell 2023, 101–103 for further examples of sticks being used for balancing in Vijayanagar yogi statuary.
46 https://fr.wikipedia.org/wiki/Gorakhnath accessed 10th January 2025.
47 Wilson 1861, 236.
48 Briggs 1938, 124. Plate v contains a photograph of the mahant with his sudarṣan.
49 Bisschop 2009, 752.
50 Bakker 2019, 555–559. This phallic connotation of the Sanskrit word for “club,” lakuṭa or lakuḍa, from which Lakulīśa gets his name, is supported by the cognate modern Hindi word lauṛā (लौड़ा) being a common slang term for the penis.
51 Nandi 1986, 103–105.
52 Cecil 2014, 150–151. From the eighth century to the present day, Śiva as Bhairava is occasionally depicted carrying a club or a staff, which would also appear to indicate protection or aggression. On early instances of Bhairava carrying a club, see Ladrech 2002. The eight different images of Bhairava at Dabhoi show him variously carrying staffs and khaṭvāṅgas, and the monumental eight-armed Bhairava at Kalinjar holds a staff. Bhairava is more often depicted carrying a club in early modern and modern images. See e.g. https://collections.vam.ac.uk/item/O70303/shiva-as-bhairava-painting-unknown/
In an illustration to a manuscript of the Śivarahasya created in Jodhpur in 1827 (fig. 54 on Diamond, Glynn, and Jasol 2008, 236–237), Śiva’s Gaṇas are depicted carrying various weapons and one has a large club, also indicating the club’s protective rôle in a Śaiva context. I thank Peter Bisschop for drawing my attention to this image.
53 Skandapurāṇa 167.27. For the text and translation, see Cecil 2020, 19–21.
54 Personal communication, January 13, 2025.
55 This is in contrast to the Cynics, who carried clubs in imitation of Heracles (Ingalls 1962, 292–293). Pace Bakker, who says that Ingalls wrongly “assumes that the Greek Heracles launched the Indian Lakulīśa” (2019, 558), Ingalls (ibid., 296) sees Cynicism and Pāśupatism as “parallel developments” that may have known each other but did not have a genetic relationship.
56 Vv. 121 and 166. The first is at an unnamed village in Karnataka, the second 18 kos from Pañcavatī (modern-day Nasik), on a mountain called Īṣārā.
57 E.g. Netratantra 9.21; Mantramahodadhi 7.33, 10.5.
58 For references and analysis, see Newid 1986, 15–16.
59 E.g. Mahābhārata 3.268.5.
60 I thank Péter-Dániel Szántó for this dating and for providing me with a transcription of the text as found in this manuscript.
61 Verse 45, f. 10r.
62 Mallinson 2020.
63 See e.g. Wujastyk 2002.
64 Dvivedī 1911, 240 (translated from Hindi).
65 Agravāla 1998, 122 (translated from Hindi).
66 See e.g. Nandi 1977, 199; Shanthamurthy 2015, 59, 65.