Śaivism in the Gupta-Vākāt.

aka Age∗

PETER BISSCHOP

Abstract

One of the features of the Gupta-Vākāt.aka age is the growth of Śaivism. In this article some of
the epigraphical evidence for this process is assembled and discussed. While the direct evidence for the
adoption of Śiva worship among the Guptas is limited to ministers of the Gupta court, it is clear
that the Vākāt.aka kings were predominantly Māheśvaras. New fragmentary wall inscriptions uncovered
from Mansar, the site of Pravarasena II’s palace, hint at a possible connection with the teachings of the
Śvetāśvatara-Upanis.ad. Two post-Gupta inscriptions from the area around Mandasor are discussed
in the light of a tendency towards religious hierarchisation, an attitude which came to be increasingly
characteristic of early medieval Śaivism. In the second part attention is drawn to the variety of Pāśupata
and Māheśvara worship in the Gupta-Vākāt.aka age, as well as to the trifold organisation of the
Pāśupata movement. The article ends with a note on the interaction with non-Śaiva traditions, in
particular Buddhism, and its possible impact upon the formation of the Pāśupata movement.

Introduction: Śiva worship under the Guptas and Vākāt.akas

In a recent important study Alexis Sanderson shows, through detailed analysis, how Śaivism
became the dominant Brahmanical religion in the early medieval period (Sanderson 2009).
While Sanderson’s study is concerned with the post-Gupta period, it is clear that this
development did not come out of nowhere. In this paper I will present, by way of a few
examples, some of the evidence attesting to the growing influence of Śaivism in the courts
of the royal houses of the Gupta-Vākāt.aka age. One of the main questions which this brings
up is why and how was the ideology of Vis.n.u as the model of kingship abandoned in favour
of the complex character of Śiva?1 In the second part I will consider the identity of the kind
of Śaivism which flourished in this period.

∗ This article is an extended version of a paper I gave at the Symposium ‘The Gupta-Vākātaka Age’, British
.
Museum, London, June 29–30, 2009. I would like to thank the organisers, Hans Bakker and Michael Willis, for
inviting me to give a presentation on the present subject. This is the first publication to appear in the context of
the research project ‘Early Śaiva Mythology: A study of the formative period of an integrated religious vision’, a
collaboration between Peter Bisschop and Harunaga Isaacson, kindly funded by a three year grant of the Arts and
Humanities Research Council (AHRC) and the Deutsche Forschungsgemeinschaft (DFG). I am grateful to Hans
Bakker, Harunaga Isaacson and Michael Willis for their critical comments on an earlier draft.
1 Cf. also Sanderson 2009: 58ff.

JRAS, Series 3, 20, 4 (2010), pp. 477–488 

C The Royal Asiatic Society 2010

doi:10.1017/S1356186310000295

https://doi.org/10.1017/S1356186310000295 Published online by Cambridge University Press

 478 Peter Bisschop

As is well-known, the Gupta epigraphical records refer to the kings of the Gupta dynasty
as paramabhāgavatas, which we can safely assume refers to their personal devotion to Vis.n.u.2
However, it is also clear that theirs was not an exclusivist religion, for within the Gupta
empire there is abundant evidence for the support of other religions as well, most notably
Buddhism, Jainism, and Śaivism. Two Gupta inscriptions in particular are important, because
they show the support of Śaivism by prominent members of the Gupta court:

• Karamdanda Inscription of the Reign of Kumāragupta.3

On octagonal base of liṅga at Karamdanda (12 miles from Faizabad), dated [Gupta] Sam . vat
117 (= CE 436). Records a gift, on 10th day of Kārttika, by Pr.thivı̄s.en.a, Kumāragupta’s
minister (mantrin), for the worship of Mahādeva Pr.thivı̄śvara (i.e. his chosen deity); also
mentions Mahādeva Śaileśvara.
• Udayagiri Cave Inscription of Candragupta II.4
Records excavation of a cave, out of devotion (bhaktyā), for Bhagavat Śambhu by Vı̄rasena,
who came from Pāt.aliputra and was a minister of Candragupta II. The latter is reported
to have come to the site with him (rājñaiveha cāgatah.).

These two records attest to the adoption of Śiva worship by primary members of the Gupta
court, while the Udayagiri inscription makes it clear that they were supported in their
religious activities by the Gupta kings. On the other hand, no evidence exists to show that
any of the Gupta kings themselves favoured Śiva as their is..tadevatā. On the contrary, the
Bhāgavata faith remained a central characteristic of this royal house.
The situation is significantly different for the Vākāt.akas, their neighbours ruling to the
south of the Vindhyas. Members of the dominant, eastern branch of this dynasty were
predominantly Maheśvara worshippers, with the noteworthy exception of Rudrasena II and
his remarkable Gupta wife Prabhāvatı̄ Guptā, both of whom followed the Bhāgavata faith
of her family. Rudrasena I, for example, is standardly referred to in the Vākāt.aka pedigrees
as “foremost among the devotees of Lord Mahābhairava” (atyantasvāmimahābhairavabhakta).
This is in itself an important piece of evidence for the history of early Śaivism, because it is
among the first attestations for the terrifying form of Śiva as Mahābhairava, quite probably
referring to a deity housed in a specific temple, as Hans Bakker has argued.5 Subsequently the
Vākāt.aka rulers used more general royal titles, such as atyantamāheśvara or paramamāheśvara,
to express their faith.6
The name Śiva itself occurs in connection with the ancestral claim of Rudrasena that he
descended from the Bhāraśivas, a “House that was installed by Śiva, who was pleased that its
members wore His emblem, the liṅga, placed as a load on their shoulders”.7 In inscriptions

2 On the epithet paramabhāgavata “supreme devotee of the Bhagavat”, see Willis 2009: 65ff.
3 Konow 1909–10: pp. 70–72; Sircar 1965: pp. 289–290. For a new reconstruction of the problematic portion
towards the end of this inscription, see Willis 2009: p. 303, n. 263.
4 CII III: 33–36; Sircar 1965: pp. 279–280. See Willis 2009: p. 40 for a picture of ‘Cave 8’ at Udayagiri.
5 Bakker 1997: p. 13, n. 23. This on account of the words svāmi and bhakta.
6 The epithet paramamāheśvara seems to appear for the first time, in Prakrit form, in an inscription of the
Śālaṅkāyana king Devavarman of Veṅgı̄pura. For references to the publication of this text and discussion, see
Sanderson 2009: p. 44, n. 7.
7 amsabhārasamniveśitaliṅgodvahanaśivasuparitustasamutpāditarājavamśa (CII V: 12, ll. 4 f.; translation Bakker 1997:
. . .. .
p. 20, n. 50).

https://doi.org/10.1017/S1356186310000295 Published online by Cambridge University Press

 Śaivism in the Gupta-Vākāt.aka Age 479

of Pravarasena II, son of Rudrasena and Prabhāvatı̄ Guptā, we encounter the repeated
claim that he “established the Kr.ta yuga [on earth] by the grace (prasāda) of Śambhu”.8 In
the Patna Museum Plate he also claims that he “carried as his weapon the lance by the
grace (prasāda) of Śambhu”.9 It is interesting to note that the name Śambhu also turns up
in one of the new inscriptions uncovered from Mansar, at the site of Pravarasena’s royal
sanctuary.10 Another such fragmentary wall inscription contains the word prasāda (‘grace’),11
which once again brings to mind Pravarasena’s royal inscriptions. Although these new wall
inscriptions are very fragmentary they do give us some insight in the kind of Śaivism
followed by Pravarasena, in that they seem to highlight once again the concept of ‘grace of
Śambhu’ (śambhuprasāda), a benevolent god who granted Pravarasena the authority to rule.12
The fragmentary inscriptions also contain an intriguing reference to the Vedānta, i.e. the
doctrine of the Upanis.ads.13 In this connection it is noteworthy that it is in the teachings of
the Śiva-oriented Śvetāśvatara-Upanis.ad that the concept of grace (prasāda) plays a key role.14
In this Vedic text the former outsider Rudra is presented for the first time in a Vedic context
as the One God (eko devah.) upon whose grace final liberation depends. If we also take into
account the fact that the Śvetāśvatara-Upanis.ad is a text associated with the Taittirı̄ya school
of the Black Yajurveda and that the copper plate charters of Pravarasena show the king’s
support of Taittirı̄ya brahmins we can surmise that it was this Vedic school that provided the
religious basis for Pravarasena’s Śaivism.15
Another potential source for tracing the Śiva devotion of Pravarasena and the Vākāt.akas,
in addition to the epigraphical and art-historical material, is the Setubandha or Rāvan.avāha,
a Prakrit kāvya attributed to Pravarasena himself. The evidence for this attribution may be
inconclusive but it is quite likely that it is a product of the Vākāt.aka period.16 Although the
subject of the poem – the building of the bridge to Laṅkā – is by definition Vais.n.ava, it is
striking that after first invoking Vis.n.u in four verses, the author continues with four verses
dedicated to Śiva. As Handiqui observes, this may well reflect the author’s Śaiva leanings:

8 śambhoh prasādadhrti(ta)kārttayugasya (CII V: 12, ll. 15–16; 19, ll. 11–12; 24, l. 16; 30, l. 16; etc.). Mirashi
. .
corrects ◦ dhr.ti◦ to ◦ dhr.ta◦ .
9 śambhoh prasādadhrti(ta)śūla(lā)yudhasya (CII V: 71, ll. 1–2).
. .
10 . . . [m] na vivarttinam// śambhor . . . // . . . rūpam / dvāra . . . (Kropman 2008: p. 6, pl. 15). The new inscri-
ptions were found on the south side of the temple at ‘Mansar III’. See Kropman 2008, for details and photographs
of these inscriptions.
11 siddham/ manāh p . . . // vedāntād dhyā[n] . . . // . . . prasādaś ca . . . // tair upahriya . . . // (Kropman 2008:
.
7, pl. 17). Two more wall fragments are preserved: tasya cārādhanāt prāpta . . . , prāptavyam // . . . va . . . / ni . . .
(ibid.: 6, pl. 16); . . . dyat[e]/ rūpādigrāha . . . (ibid.: 7, p. 18).
12 This may also be relevant to the question why Pravarasena II changed the expression of Rudrasena II,
bhagavataś cakralaks.mapratis..thitaśāsanasya, referring to Rudrasena’s Bhāgavata religion, into bhagavataś cakrapān.eh.
prasādopārjitaśrı̄samudayasya. This change has been noted by Bakker in his contribution to the British Museum
symposium.
13 See n. 11 above.
14 Śvetāśvatara-Upanisad 3.20 dhātuh prasādāt (cf. Katha-Upanisad 2.20 dhātuprasādāt), Śvetāśvatara-Upanisad 6.21
. . . . .
devaprasādāt. Cf. also Pāśupatasūtra 5.40 apramādı̄ gacched duh.khānām antam ı̄śaprasādāt.
15 Willis (2009: pp. 221–222) reaches a similar conclusion on the basis of Pravarasena’s support of Taittirı̄yas
in general and the ācārya Devaśarman in particular. Cf. also Mirashi’s observation regarding the śākhā affiliation of
Vākāt.aka donees: “It is noteworthy that among the donees of copper-plate grants the Rigvēdins and the Sāmavedins
are conspicuous by their absence, not a single grant being made to them. Among the Yajurvēdins, the followers of
the Taittirı̄ya śākhā predominate, as many as six grants having been made to them” (CII: xlv). Of the six grants
Mirashi refers to, five (nos. 3–5, 7, 15) were issued by Pravarasena II, while one (no. 8) was issued by his mother,
Prabhāvatı̄ Guptā.
16 Cf. Handiqui 1976: pp. 15–30 and Bakker 2008b.

https://doi.org/10.1017/S1356186310000295 Published online by Cambridge University Press

 480 Peter Bisschop

“the emotional fervour of the verses (1.5,7,8) in which he invokes Nat.arāja Śiva might be
an indication that, even though he venerated Vis.n.u as Rāma, his is..tadevatā was probably
Maheśvara in his dancing form”.17 More research on this question is needed, but it is not
altogether impossible to see in these eight opening verses an allusion to the concept of
Harihara,18 which would have been particularly apposite in a Vākāt.aka context, because
of the support of both Bhāgavatism and Śaivism by this royal house during this transitional
period.
To express his devotion to his Lord, Pravarasena built the Pravareśvara temple at the
new capital Pravarapura, which has yielded some of the most intriguing and beautiful Śaiva
sculptures in existence, including the justly famous ‘Mansar Śiva’.19 Given the references to
Śambhu in Pravarasena’s inscriptions it is tempting to see in this unique image an expression
of the concept of Śambhu ‘the Benevolent’.20 The name of the temple (devakulasthāna),
Pravareśvara, is among the first examples of royal sanctuaries dedicated to a chosen deity
(is..tadevatā) incorporating the ruler’s name. These temples became a characteristic feature of
most early medieval kingdoms, in which Śaivism was the norm.

Two Inscriptions from Mandasor

While these records all attest to the royal support of Śiva worship during the Gupta-Vākāt.aka
age, they do not show one particular feature of later Śaivism, which, arguably, made it so
successful, namely its hierarchical, all-encompassing stance, integrating Brahmanism and
Śaivism. Sanderson has supplied much evidence for this attitude in Tantric Śaivism in the
early medieval period, but it is also central to early Śaiva Purān.ic literature. We can observe a
trend moving in this direction in an inscription from the early sixth century: the ‘Mandasor
Stone Inscription of Yaśodharman and Vis.n.uvardhana’, dated Mālava (= Vikrama) Year 589
(= CE 532).21 The second benedictory verse reads as follows:

svayambhūr bhūtānām . sthitilaya[samu]tpattividhis.u prayukto

yenājñām. vahati bhuvanānām . vidhr.taye |
pitr.tvam
. cānı̄to jagati garimān am
. . gamayatā sa śambhur
bhūyānsi pratidiśatu bhadrān.i bhava[tām] 

This is translated by Fleet, the editor of the inscription, as follows:

May he, (the god) Śaṁbhu, confer many auspicious gifts upon you, – employed by whom in
the rites of (effecting the) continuance and the destruction and the production of (all) things that
exist, (the god) Svayaṁbhû, is obedient to (his) commands, for the sake of the maintenance of

17 Handiqui 1976: pp. 24–25. In addition to Natarāja another aspect of the deity invoked in these verses is
.
Ardhanārı̄śvara: Setubandha 1.6.
18 The number eight itself has a particular resonance in a Śaiva context, recalling as it does Śiva’s astamūrti, a
..
concept which plays a prominent role in the works of Kālidāsa. Similarly one may note that the Kumārasambhava,
Kālidāsa’s only work dedicated to a Śaiva subject, is constituted of eight sargas.
19 This sculpture has been published a number of times. Cf. e.g. Bakker 1997: Plate XXXVII. For the excavations
at Pravarapura and Pravareśvara, see Bakker 2008a.
20 For an interpretation of this icon connecting it to the figure of Nı̄lalohita, see Bisschop 2008.
21 CII III: pp. 150–158; Sircar 1965: pp. 411–417.

https://doi.org/10.1017/S1356186310000295 Published online by Cambridge University Press

 Śaivism in the Gupta-Vākāt.aka Age 481

(all) the worlds; and by whom, leading (him) to dignity in the world, he has been brought to the
condition of being the father (of the universe)!22

This verse makes, it appears, two central statements:

1. Śaṁbhu employed Svayambhū (= Brahmā) for the tasks of creation, maintenance and
reabsorption of the universe;
2. He bestowed ‘fatherhood’ (pitr.tva) upon him, causing him to be respected in the world.

Both statements indicate a hierarchical relationship between the two deities: Śiva is the
ultimate master and Brahmā owes his position to him. The second statement requires some
comment: while Fleet supplies “of the universe” it is in my opinion more likely that the poet
refers here to the Brāhman.a myth according to which Prajāpati (here Svayambhū) is the
father of Rudra.23 While the inscription apparently accepts this father-son relation, it makes
Śiva the active agent and thus reinterprets their relation.
The author(s) of the strongly Śaiva Skandapurān.a, a text datable to the end of the sixth or
early seventh century24 and as such after the period under discussion, went a lot further. In
its opening chapters the Skandapurān.a turns the Brāhman.a mythology on its head, for here
Brahmā, initially regarding himself as the first being out of ignorance, is made to realise that
he is in fact the son of Śiva (SP 3.1–9). It also makes up for the Brāhman.a presentation of
their relationship, for it tells that Śiva granted Brahmā a boon to be born as his son after
he had been propitiated by him (SP 4.1–7). This ‘son’ is not really Śiva himself but a Gan.a
named Nı̄lalohita Rudra, who chops off Brahmā’s fifth arrogant head (SP 4.11–20, 5.22–66).
This story shows a significant development in the religious imagination, in which the notion
of Śiva’s birth from the creator god was not acceptable.25
A second example of this process of religious hierarchisation is provided by another
inscription from the Mandasor area: the Rı̄sthal inscription, dated CE 512. This rich
inscription contains a wealth of important information on the history of the Aulikaras
of Mandasor, but is also relevant for the history of the later Guptas, as Richard Salomon
has demonstrated (Salomon 1989). It records, among other things, that Bhagavaddos.a, the
viceroy of Prakāśadharman, “constructed in Daśapura the Prakāśeśvara Temple, the symbol
of Bhāratavars.a” (22). It continues to note that he built, “within that same city, a beautiful
shrine of Brahman” (23) but also that he “built a shrine to Kr.s.n.a and one to Bujjuka
as a refuge for ascetics who devoted themselves to (the practice of) Sām . khya and Yoga”
26
(24). Although this might be viewed as attesting to Prakāśadharman’s tolerant attitudes
towards religion in general, it is also clear that a hierarchical order is expressed, for, while

22 The verse following this invocation in the Mandasor Stone Inscription contains an intriguing early reference
to the ‘chaplet of bones’ (asthimālā) on Śiva’s head, which attests to the Kāpālika type of development of Śiva’s
iconography in the Mandasor area.
23 Cf. Deppert 1977.
24 For this date of the text, see SP IIA: 52.
25 Note, however, that a similar criticism is already attested earlier in Kālidāsa’s Kumārasambhava 5.81cd, in
Pārvatı̄’s famous defense of Śiva: yam āmananty ātmabhuvo ’pi kāran.am . katham. sa laks.yaprabhavo bhavis.yati “He whom
they honour as cause of Self-born Brahma, how can his origin be determined?” (tr. Smith: 2005: p. 203).
26 laksma bhāratavarsasya nideśāt tasya bhūksitah| akārayad daśapure prakāśeśvarasadma yah 22 tasyaiva ca
. . . . .
purasyāntar brahman.aś cāru mandiram| unmāpayad iva vyoma śikharair gghanarodhibhih. 23 āśrayāya yatı̄nām . ca
. bujjukāvasathañ ca yah. 24 . Translations and text of this inscription
sāṅkhyayogābhiyog(in)ām| vyadhatta kr..sn.āvasatham
all by Salomon 1989.

https://doi.org/10.1017/S1356186310000295 Published online by Cambridge University Press

 482 Peter Bisschop

the Śaiva temple established for Prakāśeśvara, Prakāśadharman’s is..tadevatā, is the main object
of description and receives extensive praise, to the extent that it is even called ‘the symbol
of Bhāratavars.a’, the shrines to Brahman, Kr.s.n.a and the (probably local deity) Bujjuka, are
simply listed as other examples of the numerous religious works performed by him. The
incorporation of the royal founder’s name as the first element of the name of the temple (X-
ı̄śvara) became increasingly characteristic for the building activities of kings in the medieval
period and played an important role in the expression of royal ideology.27 Prakāśeśvara may
stand as an early example of the later royal temples which came to dominate the medieval
landscape of India and beyond.
In the aforementioned study ‘The Śaiva Age’ Sanderson identifies five “key elements
of the social, political and economic process that characterises the early medieval period”,
arguing that Śaivism was so successful in that period because it “legitimated, empowered,
or promoted” these elements.28 One of these elements is the “multiplication of land-
owning temples”, which involves among other things the phenomenon of royal temple
building which I have just discussed. Another element is, in his words, “the expansion of
the agrarian base through the creation of villages, land reclamation, and the construction of
water-reservoirs, wells, and other means of irrigation, with the steady growth in population
that these developments imply”.29 Now this element also plays a key role in the Rı̄sthal
inscription, for one of the other main activities recorded in this inscription is the construction
by Prakāśadharman of “this broad Vibhı̄s.an.a Lake, which is a mirror-image of the Bindu
Lake, dedicating to his grandfather, King Vibhı̄s.an.avardhana, its great meritorious fruit of
excellent dignity” (19).30 Indeed the two main objects of the inscription are the temple and
the lake, as becomes clear in the final verse:

As long as the wind blows, twirling the leaves of the vines and wafting the sweet perfume of
the flowers, so long may this beautiful lake and this temple of Śambhu remain, spreading their
glories and blocking the path of misfortune. (28)31

Thus in this early sixth-century inscription from Mandasor we find some of the quintessential
activities of early medieval Śaiva kings referred to. As such it is important to note that the
Śaiva kings of medieval India were not introducing a new practice, but were simply following
a model that had already been established earlier, as is indicated by this inscription, recording
the religious deeds of former feudatories of the Guptas. In fact, as Willis shows at length
in his recent book The Archaeology of Ritual, the whole system of land grants and agrarian
expansion under land-owning temples effectively starts with the early Guptas.32 One crucial
element is still missing though: the king’s initiation (dı̄ks.ā) into the Śaiva fold by a tantric

27 Cf. Sanderson 2009: 274. There are earlier examples for the practice of eponymous naming of deities: cf.
Willis 2009: p. 141. See also the example of Pravareśvara above.
28 Sanderson 2009: p. 253.
29 Sanderson 2009: p. 253.
30 rājñe pitāmahavibhı̄sanavardhanāya ślāghyānubhāvagurupunyaphalam nivedya| vistāri bindusarasah pratibimbabhūtam
. . . . .
etad vibhı̄s.an.asaras samakhāni tena 19.
31 kisalayaparivarttı̄ vı̄rudhām vāti yāvat surabhikusumagandhāmodavāhı̄ nabha(svān|) sara i(da)m abhirāmaṁ sadma
śambhoś ca tāvad vihataduritamārgge kı̄rttivistārin.ı̄ stām. ◦ vistārin.ı̄ is printed as ◦ vistarin.ı̄ (unmetrical) in Salomon’s
edition, but the accompanying note on p. 6 indicates that this must be a typo.
32 See in particular Chapter 2 in Willis 2009.

https://doi.org/10.1017/S1356186310000295 Published online by Cambridge University Press

 Śaivism in the Gupta-Vākāt.aka Age 483

rājaguru (cf. Sanderson 2009: 254 ff.).33 But for that most important element we find in the
Rı̄sthal inscription early medieval Śaiva kingship in a nutshell.

The Pāśupata Movement

As for the question of the kind of Śaivism in existence in the Gupta-Vākāt.aka age, as
mentioned there is no evidence as yet of the existence of the tantric form of Śaivism,
which involved, among other things, initiation of the king into the tantric fold. The Śaiva
Siddhānta tradition was still in its initial stage of development, although the most ancient core
of the earliest surviving Śaiva Siddhānta Tantra, the Niśvāsatattvasam . hitā (still unpublished),
has been recently dated to ca. CE 450–550 (Goodall & Isaacson 2007). Instead, for the main
agents involved in this period we have to look at the Pāśupatas, an ascetic movement of Śiva
worshippers, whose basic text was the Pāśupatasūtra. Its commentary, the Pañcārthabhās.ya by
Kaun.d.inya, may well stem from the Gupta period.
On the other hand, the precise identity of the Pāśupatas remains a complex topic. First
of all there is the issue of Lakulı̄śa. In scholarly literature it is generally held that Lakulı̄śa,
an incarnation of Śiva, was the historical founder of the Pāśupata movement and that he
lived sometime in the second century CE. This view mainly rests on the famous Mathurā
. vat 61 = CE 380. However, in
34
Pillar Inscription of Candragupta II, dated [Gupta] Sam
fact there is no mention of the name Lakulı̄śa in that inscription nor of the term Pāśupata.
What we do get is a lineage of teachers (ācāryas), the tenth of which is a certain Uditācārya,
the donee mentioned in the inscription, who trace their origins back to Bhagavat Kuśika.
This Kuśika is usually identified with the first of the four pupils of Lakulı̄śa. However, we
have no evidence that the notion of Lakulı̄śa as an incarnation of Śiva existed at the time.
The name Lakulı̄śa, or a variant of it, is attested for the first time only around the sixth
century,35 while the earliest images seem to stem from about the same period. Moreover,
the inscription makes no direct reference to Pāśupatas but only to Māheśvaras, who are asked
to do worship (pūjā) in the ‘teacher’s shrine’ (gurvāyatane). One of the intriguing aspects of
the inscription is that it refers to some of the ācāryas mentioned as the ‘spotless students’
(vimalaśis.ya) of their own preceptor. This indicates first of all a fascination with an unbroken
lineage of teachers, which remains characteristic for all forms of later Śaivism, but also, in
stressing the word ‘spotless’ (vimala), a possible connection with the mysterious Vaimalas
(‘followers of the Spotless one’), who are sometimes mentioned as a group of Pāśupatas in
Śaiva tantric sources.36

33 However, see Willis 2009: pp. 221–222, regarding the Vākātaka king Pravarasena II and his ācārya Devaśarman.
.
The earliest unmistakable epigraphical reference to Śaiva dı̄ks.ā occurs in the ‘Amudalapadu Plates of Vikramāditya
I, Year 5’ (Epigraphia Indica 32: pp. 175–184), dated CE 660. See Sanderson 2001: pp. 8–10, n. 6, for a discussion of
this and two other inscriptions from the second half of the seventh-century recording the Śaiva Siddhānta initiation
of three major kings. For the Amudalapadu Plates, see also Willis 2009: p. 270, n. 164, who notes that “[d]espite
the Vais.n.ava invocation, the plates record the gift of a village to Sudarśanācārya on the occasion of the king taking
Śaiva dı̄ks.ā”.
34 Bhandarkar 1931; Sircar 1965: pp. 277–279.
35 Cf. Bisschop 2006: p. 46, where I refer to Skandapurāna 166.25b, 166.29a, 167.129d and 167.169d. To this
.
should be added the reference to Nakulı̄śvara in the Śivadharmaśāstra (cf. Bisschop 2006: 30). The date of the
Śivadharmaśāstra remains to be settled: Hazra (1985) dates it between 200 and 500 CE, which seems to me too early.
A sixth or seventh-century dating may be more probable.
36 For this tradition, see Acri 2008.

https://doi.org/10.1017/S1356186310000295 Published online by Cambridge University Press

 484 Peter Bisschop

It is evident that much remains to be done on the Mathurā Pillar Inscription.37 In a recent
article, Diwakar Acharya has suggested a new reading of the crucial last line, proposing to
read can.d.a instead of Bhandarkar’s dan.d.a:
38
jayati ca bhagavā[ñ can.d.ah.] rudradan.d.o [’]gra[nā]yako nitya[m
.]
And always victorious is Lord Can.d.a, [he who is] the rod of Rudra,39 the foremost leader [of
the Gan.as].

As Acharya argues, this is relevant for the interpretation of the figure depicted on the pillar
in front of the triśūla. Rather than it being a representation of an unknown deity Dan.d.a,
as earlier scholars have taken it, he argues that it may in fact represent an early pre-Lakulı̄śa
Pāśupata deity Can.d.a, also known as Can.d.eśa.40 The figure of Can.d.eśa itself has been put
into a new light more recently by Dominic Goodall, in an article called ‘Who is Can.d.eśa?’
(Goodall 2009), where he shows that, contrary to what has long been supposed, Can.d.eśa
is not exclusively a Śaiva Siddhānta deity from the Tamil-speaking South of India, but has
a more complex historical origin, independent from Śaiva Siddhānta and not confined to
Tamil Nadu. This again has a bearing on the interpretation of images of Lakulı̄śa, for given
the potential for confusion between the two, as is convincingly shown by the studies of
Acharya and Goodall, some of the images so far identified as Lakulı̄śa may in fact represent
Can.d.eśa instead. A possible Gupta example of this is a loose image from Nāchnā in Madhya
Pradesh, which Joanna Williams in her The Art of Gupta India identifies as Lakulı̄śa, but for
which, given that the main attribute is clearly an axe and not a club, an identification of
Can.d.eśa may be more appropriate.41
In a way the issue concerning the identity of the deity Can.d.eśa is illustrative of a larger and
complex subject, namely the still little-understood variety of Pāśupata and Māheśvara worship
in this period. While there has been a tendency in scholarly literature to narrowly identify the
Pāśupata cult with the religious system of the sādhaka outlined in Kaun.d.inya’s commentary

37 For the latest treatment, see Willis 2009: pp. 134–139. However, the suggestion put forward there, that
the eponymous names of the two liṅgas Upamiteśvara and Kapileśvara should be taken “as both tatpurus.a and
karmadhāraya compounds” does not seem convincing. I see no reason why they should not be taken as regular
tatpurus.as alone. As a consequence I am not convinced that it is the teachers Upamita and Kapila, who have become
ı̄śvaras themselves, that receive the pūjā specified in the inscription. The two objects of worship are rather Śiva-liṅgas
named after earlier ācāryas, following the well-established model of eponymous naming practice.
38 Acharya 2005: p. 208. As Hans Bakker has pointed out to me, even with this correction, the reading remains
doubtful, because it is an unmetrical half-line of an āryā: a short syllable is needed before rudradan.d.o (sa?) to restore
the metre and sandhi. On the other hand the assumption that it is an āryā is not unproblematic: see Goodall 2009:
p. 380, n. 92.
39 I understand rudradanda as a tatpurusa compound, in contrast to Acharya and Bhandarkar, who both take it
.. .
as a bahuvrı̄hi (respectively “holder of the terrifying staff” and “whose staff is terrific”). I would argue that Can.d.a
(Can.d.eśa) is invoked here as the personification of Rudra’s rod, “a symbol of judicial authority and punishment”
(Monier-Williams, s.v.).This interpretation fits well with Can.d.eśa’s role in early Śaivism, as chastiser of transgressions
(see Goodall 2009: pp. 396–398). It is also in line with the immediately preceding passage in the inscription, which
warns: yaś ca kı̄rtyabhidroham . kuryy[ā]d ya[ś cā]bhilikhita[m upa]rry adho vā [sa] pam
. cabhir mah[ā]pātakair upapātakaiś ca
sam . yuktas syāt (ll. 15–16) “And the one who damages the memorial, and the writing above or below, he will be
invested with the five major and minor sins”. The invocation of Can.d.eśa seems appropriate in such a context.
40 A problem with this interpretation is that the figure does not seem to carry an axe (Candeśa’s characteristic
..
attribute) but a stick. There are also other weaknesses in this theory: cf. Goodall 2009: p. 380.
41 Williams 1982: p. 113, pl. 163. Bakker (1997: p. 100), who discusses the image in connection with an image
from Mandhal which he provisionally identifies as ‘Gan.ādhyaks.a’ (plates V and VI), also dismisses the Lakulı̄śa
identification, without however considering the possibility that it might represent Can.d.eśa.

https://doi.org/10.1017/S1356186310000295 Published online by Cambridge University Press

 Śaivism in the Gupta-Vākāt.aka Age 485

on the Pāśupatasūtra, this in fact represents only one element of the tradition. The Pāśupata
system as outlined by Kaun.d.inya involves a lifelong career of extreme asceticism, which is
hard to reconcile with other early references to Pāśupatas, in particular epigraphical records.
Thus, for example, the earliest explicit epigraphical references to Pāśupatas that we possess are
at the same time among the earliest examples of copper-plate grants recording endowments
for temple worship. These are the copper plates from Bāgh, which record the land grants
given by the mahārājas of the Valkhās, who were very probably subordinates of the Guptas.
The inscriptions themselves stem from the second half of the fourth century. Seven copper
plates in total in this collection refer to Pāśupatas as recipients of grants for the performance
of worship in temples (Nos. III, V, VI, IX, X, XII, XIV).42 Other religious communities are
mentioned as recipients of these grants as well.43 The deities under worship are not limited
to the figure of Śiva alone, however, for among the names of gods to whom the grants
were dedicated we encounter Nārāyan.adeva (no. III), Mahāsenadeva (i.e. Skanda; no. IX)
and Bappapiśācadeva (nos. V, VI, XII and XIV). The last one, perhaps a local form of Śiva,
seems to be connected with the Pāśupatas in particular, as in two of the grants (nos. V and
VI) only the Pāśupatas are mentioned as recipients. One of the grants (no. X) also records
that a shrine to the Mothers (mātr.sthānadevakula) had been established by the Pāśupatācārya
Lokodadhi in the village of Piñcchikānaka.
What these grants show is that not all Pāśupatas followed the rigorous ascetic system of
Kaun.d.inya, but that there were others who served the needs of a larger, lay Śaiva community.
The mention of Nārāyan.adeva among the gods worshipped moreover suggests that they could
also fulfill priestly services in temples dedicated to non-Śaiva deities. The Pāśupata career
outlined in the scholastic work of Kaun.d.inya, with its emphasis on lifelong asceticism, as
such only represents one strand in a larger, complex religious field. In fact Kaun.d.inya’s system
itself clearly requires the existence of such ācāryas, who by definition can not themselves be
engaged in the kinds of activities that the sādhaka has to perform on the stages of the Pāśupata
path to liberation. Likewise there would have been a community also of those who have faith
in such Pāśupata ācāryas and in the Pāśupata teaching, but who are tied to their own lives
and can not take the large step of renunciation and consequent adherence to the Pāśupata
sādhaka’s rules.44 As such one may conceive of a Pāśupata community consisting of three
segments: 1) ācāryas (such as Kaun.d.inya himself), 2) sādhakas (practicing the system outlined in
the Pāśupatasūtra and Kaun.d.inya’s commentary thereon), and 3) a community of the faithful
(consisting of uninitiated, non-ascetic supporters/devotees).45 For information on the third
segment one has to look at more popular texts, such as the Skandapurān.a, but in particular at
the various texts which together make up the Śivadharma corpus. These texts, such as the
Śivadharmaśāstra, the Śivadharmottara and the Śiva-Upanis.ad, are still much neglected, with the

42 Ramesh & Tiwari 1990. On the Valkhās and their records, see also Virkus 2004: pp. 108–115.
43 Besides Pāśupatas are mentioned the Ārya-Coksas (no. III; an early Vaisnava sect), the Mantraganācaryas (no.
. .. .
IX) and the Bhagavacchis.t.as (nos. XII, XIV).
44 Note that the Mathurā Pillar Inscription also appears to distinguish between ācāryas on the one hand and lay
Māheśvaras on the other. The latter are asked to guard the property (parigraha) of the ācāryas: naitat khyātyartham
abhili[khyate] (|) [atha] māheśvarān.ām . vijñaptih. kriyate sambodhanam . ca (|) yathākā[le]nācāryān.ām
. parigraham iti matvā
viśaṅka[m. ] [pū]jāpuraskāra[m
. ] parigrahapāripālyam. [kuryyā]d iti vijñaptir iti (ll. 10–14).
45 I reproduce and expand here some of the points made by Harunaga Isaacson in an e-mail, dated 30 August
2009.

https://doi.org/10.1017/S1356186310000295 Published online by Cambridge University Press

 486 Peter Bisschop

larger part of this corpus not having been properly edited.46 In addition, Diwakar Acharya
has recently discovered a number of Pāśupata manuals (vidhis), which likewise hint at a more
diverse Pāśupata religious milieu, involving both an ascetic and a lay community.47

Interaction with non-Śaiva traditions

The division into ascetic and lay community calls to mind the religious traditions of
Buddhism and Jainism, where such a division was already in place for a longer time. It
may well be that the Śaivas modelled themselves on these traditions, although we have no
direct evidence for this. It is striking, however, that the iconography and life-story of Lakulı̄śa
and the Buddha have shared characteristics. Both icons depict a human being in a seated
posture, possessed of various divine or auspicious attributes. But for the distinctive features
of the erect penis, the club and the matted hair, there is a strong resemblance between early
depictions of Lakulı̄śa and the Buddha, which suggests that Buddhist (but possibly also Jain
iconography) had a big impact in this formative period on the iconography of Lakulı̄śa.
When we look at the life-story of Lakulı̄śa, the most striking feature is its emphasis on the
humanity of God’s descent. This is not a cosmic type of avatāra, as tends to be the case with
the avatāras of Vis.n.u, but it is the story of a God taking on human form, who wanders from
place to place, initiates his four pupils at different places in the north of India and instructs
them in His (Paśupati’s) teachings. This conjures up the image of the Buddha wandering
through Magadha. The similarity may not have escaped the Pāśupatas themselves, for in
a late passage of the Skandapurān.a mention is made of Lakulı̄śa’s (lagud.ı̄śvara) wanderings,
surrounded by his pupils, in Magadha. Eight sites (as..tau sthānāni) in Magadha are said to be
connected with these wanderings, which once again brings to mind a Buddhist tradition,
namely the eight great sites (as..tamahāsthāna) of the Buddha’s life story.48
That there should be such an interaction between these traditions in this period should
not come as a surpise. It is noteworthy, for example, that some of the major Pāśupata sites are
also places with a strong Buddhist connection. As examples may be mentioned the two cases
just discussed, Mathurā and Bāgh,49 but also Vārān.ası̄, arguably one of the most important
Śaiva places of all time, which is not too far from the Buddhist site of Sārnāth. Anyone
reading Hsiuen-Tsang’s travelogue cannot fail to note the constant references to Pāśupatas
whom the Buddhist pilgrim encounters on his travels through seventh-century India.50 The
evidence is not limited to iconography and topography, but there is textual evidence as
well, for one of the earliest quotations of the Pāśupatasūtra itself occurs in a Buddhist work,
Bhā(va)viveka’s Mādhyamakahr.dayakārikā.51 The name of the author itself, it may be noted,

46 For a recent update on this material, see Goodall 2009: pp. 374–375, n. 88. The Śivadharmaśāstra, the earliest
of these texts, is particularly important, because, as Hazra has shown, “it is totally free from Tantric influence”
(Hazra 1985: p. 296). As for the Śivadharmottara, Hazra observes that “[i]t belongs to those Pāśupatas who looked
upon Lakulı̄śa as a great teacher (guru) identical with Śiva himself” (Hazra 1983: p. 204).
47 For an introduction to these vidhis and an edition and translation of the first of these, the Samskāravidhi, see
.
Acharya 2007.
48 Skandapurāna 167.169: magadhāsu smrtāny astau sthānāni śaśimaulinah| śisyaih parivrto yāni babhrāma lagudı̄śvarah|
. . .. . . . . . .
tāni dr..s.tvā bhavet sadyah. pumān pāpavivarjitah.. Cf. Bisschop 2006: p. 218, where a different explanation of these eight
sanctuaries is put forward.
49 For Mathurā, see Srinivasan 1989; for Bāgh, see Verma 2007.
50 Cf. Beal 1884, Index, s.v. Pāśupatas (Po-shu-po-to).
51 Cf. Bisschop 2007: pp. 14–18. Curiously, the references to the Pāśupatasūtra are found in Chapter Nine of
this work, which is directed to the Mı̄mām . sā school.

https://doi.org/10.1017/S1356186310000295 Published online by Cambridge University Press

 Śaivism in the Gupta-Vākāt.aka Age 487

could hint at a possible earlier Pāśupata background, for names with Bhā-x or Bhāva-x
are very common among Pāśupatas.52 In any case, it is intriguing that this sixth-century
Buddhist author had access to the Pāśupatasūtra, an esoteric text in principle meant only for
an initiated community of sādhakas. It certainly indicates the extensive exchange that took
place between these different religious traditions during the period under discussion.

Bibliography
Diwakar Acharya, 2005. “The Role of Can.d.a in the Early History of the Pāśupata Cult and the Image
on the Mathurā Pillar dated Gupta Year 61,” in Indo-Iranian Journal, Vol. 48: pp. 207–222.
Diwakar Acharya, 2007. “The Sam . skāravidhi: A Manual on the Transformatory Rite of the Lakulı̄śa
Pāśupatas,” in Dominic Goodall & André Padoux (eds.), Mélanges tantriques à la mémoire d’Hélène
Brunner. Collection Indologie 106 (Pondichéry), pp. 27–48.
Andrea Acri, 2008. “The Vaimala Sect of the Pāśupatas. New Data from Old Javanese Sources,” in
Tantric Studies, Vol. 1: pp. 193–208.
Hans T. Bakker, 1997. The Vākāt.akas. An Essay in Hindu Iconology. Gonda Indological Studies V.
(Groningen).
Hans T. Bakker, 2008a. “Mansar. The Discovery of Pravareśvara and Pravarapura. Temple and
Residence of the Vākāt.aka King Pravarasena II,” Proceedings of a Symposium at the British Museum.
Groningen. E-book: http://mansar.eldoc.ub.rug.nl/
Hans T. Bakker, 2008b. Mansar, Pravarasena and his Capital. An Introduction, in Bakker 2008a.
Samuel Beal, 1884. Si-Yu-Ki. Buddhist Records of the Western World. Translated from the Chinese of
Hiuen Tsiang (A.D. 629). (London). [Reprint, Delhi 1969]
D. R. Bhandarkar, 1931. “Mathura Pillar Inscription of Chandragupta II: G.E. 61,” in Epigraphia Indica,
Vol. XXI: pp. 1–9.
Peter Bisschop, 2006. Early Śaivism and the Skandapurān.a. Sects and Centres. Groningen Oriental Studies
XXI. (Groningen).
Peter Bisschop, 2007. “The Sūtrapāt.ha of the Pāśupatasūtra,” in Indo-Iranian Journal, Vol. 49: pp. 1–21.
Peter Bisschop, 2008. The Skull on Śiva’s Head. Preliminary Observations on a Theme in the Śaiva Art of
Mansar, in Bakker 2008a.
Corpus Inscriptionum Indicarum 1888. Vol. III. Inscriptions of the Early Gupta Kings and their Successors.
Edited by J. F. Fleet. (Calcutta). [= CII III]
Corpus Inscriptionum Indicarum 1963. Vol. V. Inscriptions of the Vākāt.akas. Edited by Vasudev Vishnu
Mirashi. (Ootacamund). [= CII V]
Joachim Deppert, 1977. Rudras Geburt: systematische Untersuchungen zum Inzest in der Mythologie der
Brāhman.as. (Wiesbaden).
Dominic Goodall, 2009. “Who is Can.d.eśa?” In Shingo Einoo (ed.), Genesis and Development of Tantrism
(Tokyo), pp. 351–423.
Dominic Goodall & Harunaga Isaacson 2007. “Workshop on the Niśvāsatattvasam . hitā,” in Newsletter of
the NGMCP, Vol. 3: pp. 4–6.

52 As examples may be mentioned the Pāśupata teachers Bhāvavālmı̄ka, Bhāvasamudra and Bhāvavirimchi,
.
mentioned in the Amareśvara Temple Inscription (Epigraphia Indica XXV: pp. 173–185), and Bhāvatejas and his
disciple Bhāvabrahman, mentioned in the Tewar Inscription (Indian Antiquary XVIII: pp. 209–211). As for the name
Bhā-x the best example is Bhāsarvajña, the author of the Ratnat.ı̄kā commentary on the Pāśupata Gan.akārikā. A
passage of the Nāt.yadarpan.avr.tti (kindly pointed out to me by Somadeva Vasudeva) makes this convention explicit:
aparah. pāśupatādivratı̄ svasamayaprasiddhanāmabhir vācyah.| yathā pāśupatasya bhāpūrvam
. bhāsarvajña ityādi sambhās.an.am|
(Vr.tti ad Nāt.yadarpan.a 4. pp. 50–52).

https://doi.org/10.1017/S1356186310000295 Published online by Cambridge University Press

 488 Peter Bisschop

Krishna Kanta, Handiqui, 1976. Pravarasena’s Setubandha. Translated into English with extracts from
unpublished commentaries, critical notes and an Introduction. Ahmedabad.
R. C. Hazra, 1983. “The Śiva-dharmottara,” in Purān.a, Vol. 27: pp. 181–210. [= Journal of the Ganganatha
Jha Research Institute 13 (1952–53): pp. 19–50.]
R. C. Hazra, 1985. “The Śiva-dharma,” in Purān.a, Vol. 27: 282–299. [= Journal of the Ganganatha Jha
Research Institute 10 (1952–53): pp. 1–20.
Sten Konow, 1909–10. “Karamdanda Inscription of the Reign of Kumaragupta: [Gupta-] Samvat 117,”
in Epigraphia Indica, Vol. X: pp. 70–72.
Martine Kropman, 2008. The Seals and Inscriptions from Mansar, in Bakker 2008a.
V. P. Limaye & R. D. Vadekar (eds.) 1958. Eighteen Principal Upanis.ads. Vol. I. Upanis.adic Text with
Parallels from extant Vedic Literature, Exegetical and Grammatical Notes. Gandhi Memorial Edition.
(Poona).
Sir Monier Monier-Williams, 1899. A Sanskrit-English Dictionary. Etymologically and Philologically
Arranged with Special Reference to Cognate Indo-European Languages. (Oxford). [Reprint, Hong Kong
1995.]
Pāśupatasūtra. Pāśupatasūtra with the Pañcārthabhās.ya of Kaun.d.inya. Edited by R. Ananthakrishna Sastri.
Trivandrum Sanskrit Series CXLIII. (Trivandrum, 1940).
K. V. Ramesh & S. P. Tiwari 1990. A Copper-Plate Hoard of the Gupta Period from Bagh, Madhya Pradesh.
(New Delhi).
Richard Salomon, 1989. “New Inscriptional Evidence for the History of the Aulikaras of Mandasor”,
in Indo-Iranian Journal, Vol. 32: pp. 1–36.
Alexis Sanderson, 2001. “History through Textual Criticism in the study of Śaivism, the Pañcarātra
and the Buddhist Yoginı̄tantras,” in François Grimal (ed.), Les Sources et le Temps. Sources and Time.
A Colloquium. Pondicherry 11–13 January 1997 (Pondicherry), pp. 1–47.
Alexis Sanderson, 2009. “The Śaiva Age: an explanation of the rise and dominance of Śaivism during
the early medieval period,” in Shingo Einoo (ed.), Genesis and Development of Tantrism (Tokyo),
pp. 41–350.
Setubandha. The Setubandha of Pravarasena. Edited by Pan.d.it S’ivadatta & Kâshinâth Pân.durang Parab.
Kâvyamâlâ 47. (Bombay, 1895).
Dines Chandra Sircar, 1965. Select Inscriptions bearing on Indian History and Civilization. Volume 1. From
the Sixth Century B.C. to the Sixth Century A.D. (Calcutta).
Skandapurān.a (SP I). The Skandapurān.a. Volume I. Adhyāyas 1–25. Critically Edited with Prolegomena
and English Synopsis by R. Adriaensen, H. T. Bakker & H. Isaacson. Supplement to Groningen
Oriental Studies. (Groningen, 1998).
Skandapurān.a (SP IIA). The Skandapurān.a. Volume IIA. Adhyāyas 26–31.14. The Vārān.ası̄ Cycle. Critical
Edition with an Introduction, English Synopsis & Philological and Historical Commentary by Hans
T. Bakker & Harunaga Isaacson. Supplement to Groningen Oriental Studies. (Groningen, 2004).
David Smith, 2005. The Birth of Kumāra by Kālidāsa. Clay Sanskrit Library. (New York).
Doris Meth Srinivasan, 1989. Mathura: the cultural heritage. (New Delhi).
Archana Verma, 2007. Cultural and visual flux at early historical Bagh in central India. (Oxford).
Joanna Williams, 1982. The Art of Gupta India. Empire and Province. (Princeton).
Michael Willis, 2009. The Archaeology of Hindu Ritual. Temples and the Establishment of the Gods.
(Cambridge).

Peter Bisschop
Leiden University

https://doi.org/10.1017/S1356186310000295 Published online by Cambridge University Press