DIWAKAR ACHARYA

THE ROLE OF CANDA IN THE EARLY HISTORY OF THE

S
PA UPATA CULT _AND
_ THE IMAGE ON THE MATHURA 
PILLAR DATED GUPTA YEAR 61*

The image on the Mathura pillar bearing an inscription with the Gupta
year 61 (equivalent to 380/381 AD) was identified early in 1931 by D.R.
Bhandarkar as a standing figure of Lakullśa.1 Most scholars still accept this
identification, though in recent decades some have raised suspicions about
it and are themselves inclined to uncertainty.2 The image on the pillar does
not bear the standard attributes of Lakullśa, namely a sitting pose and
ithyphallic appearance. Likewise, the inscription engraved on the pillar
makes no mention of Lakullśa, nor does it even refer to the Paśupatas
directly. There is no evidence that the notion of Lakullśa as an incarnation
of Śiva existed at that time, even though the name Lakullśa or a variant of
that name is attested in early Śaiva scriptures. And, there is another striking
point to be noted: namely, that a fairly regular iconography of Lakullśa
begins to emerge only by about the sixth century.3
Bhandarkar’s ground for identifying the image as Lakullśa was that
the inscription on the pillar records a donation by one Uditacarya, tenth in
line from Bhagavat Kuśika, who in turn is purportedly a direct disciple
of Lakullśa.4 R.C. Agrawal (1969: 355) has rejected Bhandarkar’s

* I would like to express my gratitude to Prof. Hans Bakker (Groningen University),

Prof. Harunaga Isaacson (University of Pennsylvania), Dr. Habil. Dominic Goodall
(École Française d’Extrême-Orient in Pondicherry) Prof. Arlo Griffiths (Leiden
University) and Dr. Peter Bisschop (Wolfson College, Oxford) for their critical
comments on an earlier draft of this paper. I am also grateful to Mr. Philip Pierce
(Nepal Research Centre) for going over my English.
1
Bhandarkar 1931: 1Y9. Bhandarkar reproduces a rubbing of the inscription but not a
picture of the image. I reproduce here a picture of the image with the kind permission of the
Mathura State Museum. I am grateful to Prof. Madhav Sharan Upadhyaya (Mahendra Sanskrit
University) for approaching the museum authority on my behalf and getting this permission.
2
Shah 1984: 97, Dyczkowski 1989: 20,144. Lorenzen (1991: 180) also appears unsure but
in the end terms the image a BLakullśa-like standing figure^ and finds it reasonable to assume
the Mathura pillar record to be Lakullśa-Paśupata in origin. See also below, footnote 5.
3
Shah 1984: 97.
4
Bhandarkar (1931: 5Y6) has identified this Bhagavat Kuśika with Kuśika, one of the
direct disciples of Lakullśa, but Lorenzen (1991: 180Y181), finding the identification
problematic, raises the possibility of two Kuśika-s. Pathak (1960: 9) identifies him with
Kuśika II mentioned in the list of 18 avat aras transmitted in Jain sources. See Bisschop
2006: 45f.,49f.

Indo-Iranian Journal 48: 207Y222, 2005. # Springer 2006

DOI: 10.1007/s10783-005-2197-8
208 DIWAKAR ACHARYA

identification of the figure and attempted to prove that it is not Lakullśa

but BŚiva as Bhairava.^ T.S. Maxwell (1988: 7) in turn has openly
criticized Agrawal’s view, stating that Bone of the grounds for this change
of mind Y which is not acknowledged Y is that no icons of Lakullśa are
found in Kusana sculpture with a flabby belly, which is a prominent
feature of the_ Mathura pillar figure. This is not only an unacknowledged
reversal of a previously advanced identification, but also an inaccurate
description.^
Using the information existing in the inscription itself can solve the
problem of the identification of the image on the pillar. The clue is found
in the concluding line of the inscription, which is nothing but a verse
composed in the ary a metre in praise of the deity sculpted below on the
pillar. Surprisingly, Bhandarkar and Agrawal have not written a single
word on the relation of this line to the rest of the inscription or to the
figure.5 However, some letters in this line, as in other lines, are partly
effaced. Bhandarkar has restored these effaced letters tentatively within
brackets, and other scholars have accepted his reading. First of all, then,
it is necessary to examine Bhandarkar’s reading of this line, which runs
thus:
6   
a n ¼d andah  rudra-dando¼gra- n
jayati cha bhagav a yako
 6 _ _ __ _ __
nitya m:
_
Here, the Lord (bhagavat),
the foremost leader and the holder of the
terrifying staff rudradanda , is praised, but unfortunately the name of
__
the Lord is effaced. Bhandarkar erroneously deciphered this word of two
letters to be dand a. He reads this name along with the initial adjective as
bhagav an d an d_ ah
_ , and immediately observing that this reading is
_ _ _
wrong, he suggests_ _ what he takes to be a grammatically and metrically
correct reading in a footnote: bhagav an dandah sa.7 But this reading
cannot be accepted even on paleographic grounds. _ _ _ The ligature following
bhagav a is different than the ligature ndah immediately following or
with another ligature ndo in rudrad and_o_ in_ the same line. Bhandarkar
_ _ between the
fell prey to the similarity _ ligatures
__ ñca and nda in the Gupta
__
5
Lorenzen (1991: 180) has unquestionably understood the importance of the
concluding line, but his conclusion is wrong. He writes: B[t]he only reliable means of
identifying Uditacarya’s sectarian allegiance are the inscription’s concluding line of
praise to Lord Danda, who bears the staff of Rudra, and the Lakullśa-like standing figure
engraved on the _pillar.
_ These render it reasonable to assume that this is a Lakullśa-
Paśupata record, but there is still a problem about the identification of Bhagavat Kuśika.^
6
Bhandarkar 1931: 9. Note that ca is transliterated as cha.
7
Bhandarkar 1931: 9, footnote 5.
 THE ROLE OF CANDA ETC. 209
__
 
a n~ candah , and thus, the concluding
script. So I propose to read bhagav
line of the inscription can be read as: _ _ _
 8  
a yako nitya m: 8
a ~n candah rudradan d o ½’ gra½n
jayati ca bhagav
__ _ __ _
And always victorious is Lord Canda, the foremost leader and the
holder of the terrifying staff. __

The image depicted below the inscription on the base of the pillar
matches the description of Canda in the ary a: Fthe foremost leader and
_ _ Now, the relation of this line to the rest
the holder of the terrifying staff._
of the inscription becomes clear, and we can safely conclude that the
image on the Mathura pillar is Canda, not Lakullśa.
An association of Canda with _the _ Paśupatas is evident also on an
_ _
inscription from Nepal from the early seventh century. There are two
shrines of Cand eśvara in the courtyard of the famous Paśupati temple in
Kathmandu, one _ _ in the north-east and the other in the south-east corner
of the main temple. The first shrine is simply called ‘Can d eśvara’s
shrine’ and a small odd-shaped linga _ is worshipped there. The _remains
_ of
offerings (nirm alya) made to Paśupati in the inner sanctum of the temple
are offered to Candeśvara in this shrine, which is in the proper place as
prescribed in the_ _ sivopanisad, Pratist h arasamuccaya9 and
alaks an as
_
other śaiva texts. No inscription _ _ historical
or other _ _ document is present
10
in this shrine.
The other shrine is called ‘Chatracandesvara’s shrine’, and in line with
__
the name of the deity, the image of Chatracan devvara stands beneath a
parasol (chatra) made of stone at the top of a _pillar. The pillar next to
_
the image of Chatracandevvara is engraved with an inscription.11 This
__
inscription records a donation, by a Paśupata teacher ( ac
arya) called
Bhagavat Pranardanapran a Kauśika in the reign of King Jis n ugupta
_ __
8
 The reading  is still grammatically incorrect; the correct reading should be bhaga
a ms candah .
v
9_ __ _
See below, footnotes 33 and 34.
10
It is highly probable that this shrine was built later when the Paśupati temple in
Kathmandu got influence from the South and accepted Saiddhantika ritual scheme. It is
sure that Candevvara is made the consumer of nirm alya only later in Siddhanta Śaivism. I
__
even see a reflection of such a transformation of Candevvara from the supreme Lord to
the consumer of nirmalya in the account of Buddhists _turning
_ Paśupati into Ucchistevvara
found in the Bh 
asavamsavali. __
11
Vajracarya _1973:_ inscription no. 112, Regmi 1983: inscription no. 104. Lorenzen
(1991: 30) is aware of this inscription, but he assumes a Kapalika connection here,
ignoring the direct reference to the Paśupatas and their assembly.
210 DIWAKAR ACHARYA

(624Y632 AD),12 of a few fields for the sake of maintaining the shrine of
Chatracandevvara and a water conduit in some village. It is worthwhile to
note that _he
_ makes his donation to some responsible ascetics in the mu-
nd asr nkhalikap
_ asupat
acaryaparsad Y the assembly of Paśupata teachers
__ _
belonging to the Mun d a and Sr_ nkhalika
_ sub-sects.13 This indicates that
Cand esvara was a deity_ _ commonly
_ worshipped by both divisions of Paśu-
_ _
patas associated with the parsad, and probably others connected with
them. _
Here I cannot help but point out the striking similarities in the contents
of this inscription and the Mathura pillar inscription. Both inscriptions
praise Cand a in his particular forms; they are silent about Lakullśa but
indicate a_ _ close relation between the sectarian teachers and Kuśika.
Further, the sectarian teachers in both are given the honorific designation
of Bhagavat.
More important in the Chatracan d esvara inscription is the opening
verse, an invocation to Chatracan d e _s_vara. Though this invocation does
_ _
not provide any iconographic detail of the deity, it does help us to understand
the theological understanding and the supremacy accorded to the deity.
It runs thus:
samyagj~nan
 adiyuktah saka(la)gunaganam ksobhayitva pra(dh a)nam
brahmadisth
  _ ~ jagad idam
avarantan _ akhilam
_ _ _yo’sr jad viśvarupam j
_ _
12
As the inscription is damaged, the exact date of its installation is not known.
13
Bühler and Indraji (1880: 174) took this term to mean an assembly of teach-
ers belonging to a sect of Paśupatas whose members wore a chain of heads. Though
the reading of the Chatracandesvara inscription has been improved on in recent years,
scholars still follow the interpretation__ of this term Bühler and Indraji suggested 125 years
ago. This interpretation goes against grammatical intuition and appears to me hasty.
The than
 affix never can be applied to the term śsrn_ khal a and following the sam asa rules
for bahuvr l hi compounds, the type as suggested by the above interpretation, the word
formed will be mun daśsrn_ khala or mun daśsrn_ khalaka or mun daś  srn_ khalaka, but never

mun daśsrn_ khalika. The term mundaśsrnkhalikap aśsupata can be formed as a tatpurusa
_ 
having a dvandva as one of its components, or else as a pure tatpurusa. Following the
first, it would stand for an assembly of teachers of two distinct Paśupata sects,
Mundapasupatas and Ś Sr n_ kha-likapaśsupatas, respectively the sect of Paśupatas whose
__
members shaved their head and the sect of Paśupatas who wore a waistband. Fol-
lowing the second, it would mean an assembly of a single sect of Paśupatas, named
 sr n_ khalika-pasupatas, who shaved their head and also wore a waistband. The sub-
Mun daś
sect of Ś Sr n_ khalikapasupatas, is separately referred to in the same Chatracan deś  svara pillar
inscription a few lines below and again in one undated inscription of Narendradeva’s
time (Vajracarya 1973: inscription no. 125, Regmi 1983: inscription no. 118). This ref-
erence, together with some other logical points, leads me to follow the first option and
accept two independent sub-sects of Mundapasupatas and Ś Sr n_ khalikapasupatas. For
further details, see Acharya 1998. __
 THE ROLE OF CANDA ETC. 211
__
ajl vyam sarvapums
am giritarugahanam yah karoty ekarupam
_ ’dya prasanna
payat so _ _ h smaratanudahanaś
_ _ chatracandeśsvaro vah k
  
Chatracan deś
 svara, who burnt the body of Kama, who possesses true
knowledge and the other [virtues], and who, having differentiated the
pradhana, the entire sum of [three] qualities, created this whole world in
its different forms, from Brahma to inert matter, and created everything Y
the thicket (gahana) of mountains and trees, in one unique form af-
fording a livelihood for all men Y today may he be pleased and protect
you.

Can d a is not described as inhabiting the ultimate realm or as being

the supreme deity in the available Paśupata texts or in accounts of
the Paśupata cult in Puranic and Tantric Śaiva sources. But in the above
verse, he is clearly identified with the ultimate Śiva. Can d a is one of
Śiva’s names and it is possible that the Paśupatas worshipped him
as Can d a. It is quite likely that the supreme deity was also treated
as the primordial teacher. In the Svacchandatantra14 and Jayadratha-

14
Svacchandatantra 10.1047Y1052:
gurupan_ ktitrayam  divyam gunair
 antaritam  sthitam j
prathama tamasah pan_ ktir uparist ad vyavasthita jj
tesam
 namani kathyante yathavad anup urvaśsah j
 s caiva pratapavan jj . . .
śsivah prabhur vamadevaś can daś
gan do ldanah j
 naro yamo mal l gahaneśaś ca p
 
prathama pan_ ktir uddist
a rudrair dv atri mś
 sat a jj
a smrt
BThe three groups of divine teachers are concealed by the [three] gunas.  The first group is
situated above tamas. The names of the [teachers included in this group] are, mentioned
properly in successive order: Śiva, Prabhu, Vamadeva, Can d a and Pratapavat, . . . Gan da,

Nara, Yama, Malin, Gahaneśa and Pldana.  This first group is said to be consist of 32
Rudras.[
The Svacchandatantra in this section appears to draw heavily upon the Guhyasutra
section (7.144Y158) of the Niśsv asatattvasamhit
 a. Unfortunately, the text of the
Guhyasutra is disrupted exactly at the point where Can d a would have featured. The
reading of the line is also slightly different, with Ananta, a new name not found in the
Svacchandatantra or Jayadrathayamala, being inserted after Śiva. The text reads śivo
’nantah prabhur vamo ++++ prat apanah, so that we are not in a position to say whether
Can d a featured in the list of the Guhyasatra or not. However, it is interesting to note that
Lakullśa does not appear in any of the three groups of teachers. Kun dal  lva features in his
place, though Lakuli/Lakullśa is mentioned elsewhere in the text. This suggests that it
was important for the author of the Svacchandatantra to include Lakullśa in the group of
teachers, whereas for the author of the Guhyasutra it was not.
212 DIWAKAR ACHARYA

yamala,15 Can d a appears in the first of three groups of teachers

(gurupan_ ktitraya) in the domain of the three gun as, while Lakullśa
appears in the second group of thirty teachers. As _ Goodall, Törzsök
and Rastelli write, FAghoraśiva prescibes that at the end of the
Nirvan adlks a the initiate should be led to Can d eśa, where he is to
listen to the post-initiary rules in front of the god._16 One can easily
argue that in this Siddhanta Śaiva practice Can d eśa is still regarded as
the primordial teacher in front of whom the initiate accepts the post-
initiary rules of the cult (samayas).
Furthermore, it is interesting that Can d a in the above verse is called
the creator of a thicket (gahana). This appears to allude to Gahaneśa,
whom some of the early Tantras place in the mayatattva alongside
Tejeśa and Dhruveśa,17 who are regarded as the supreme deity or the
ultimate goal of the Praman a division of the Paśupatas.18 Some other
 gamas and early Tantras place him in the first of the three groups of
A

15
Jayadrathayamala, fol. 112r1Y4:
gunanur agam adi gurupamktitrayam mahat j
atr
_ antoparist at prathamam _namany e_ s
dhv am
_ brahmado bh  j
 yatha śsrnu
śivaś can dah_ _ prabhur vamo lma uttamah j. . .
yamo ma_ l_l gahaneś
_ sah pl d anaś
s ceti te sm rtah j _
ete dv _
atrim śsad akhyata. _ _ _
_
BThe great triple group of teachers comes ahead the mere attachment to the gunas. The
first [of them] is above tamas, darkness. Listen to their names as they are: Śiva,_ Can d a,
Prabhu, Vama, Brahmada, Bhlma, Uttama, . . . Yama, Malin, Gahaneśa and Pldana. Thus
they may be recalled, and are 32 [in number].[ _
16
Tantrikabhidhanakośa II, s.v. Can d eśa and Ca n d ~ 
17
 ajnabhanga.
The Mrgendratantra, Parakhyatantra, Matan_ gap arameśsvar agama and Ramakan tha

in the Kiranav   rtti place Gahaneśa in the ma ya
 tattva, but the Niśva
saguhya, Svacchanda-
tantra, and Jayadrathayamala place him in the gurupan_ kti. For example, Mrgendra-
tantra, vidyapada 13.153Y154:
 l śso brahmana
anantas trikalo gopta k sem  h patih j
dhruvatejodhisau
 rudrau gahaneśaś ca viśsvar at jj
m adhikarino
ay  rudra man dal
 adhipat l ś
svar ah j
 aracakrakar
sams udhabh
 utagramavivartakah jj
BAnanta, Trikala, Goptr;
  lśa; Brahmanaspati,
K sem  Dhruveśa, Tejeśa, and Gahaneśa,
who reigns over the world, are the Rudras who govern the domain of Maya. [Each of
them] is the omnipotent lord of his dominion, and transforms multitude of beings mounted
on the wheel of transmigration.
18
The Svacchandatantra at one point makes both Tejeśa and Dhruveśa the ultimate
goal of the Pramaana division of the Paśupatas (10.1174ab), but later transfers Tejeśa to
the Vaimala division (11.71Y72).
 THE ROLE OF CANDA ETC. 213
__
teachers alongside Can d a.19 He is regarded as the cause of primordial
matter ðm ulaprakr tik
aranaÞ
 in these Tantras and is designated as the
_
overlord of all Rudras (sarvarudramaheśvara).20
Now we come to iconographic details of the Mathura image and the
Chatracan deśvara
 image from Kathmandu. The image of Can d a on the
Mathura pillar (see plate II) is shown standing

. . . with two hands, the right of which is let down catching a staff or club
and the other held akimbo but also bearing some unidentifiable object. The
hair on the head is matted with some curls falling on both the shoulders.
The statue bears, apparently, two garments, the upper or uttarlya being
made fast to the body by a band passing round between the chest and the
belly and with one end flowing loose spirally at the proper left. Although
the lower part of the body is apparently clothed with a Dhotl, the privates
are clearly shown, like the breast nipples appearing through the upper
garment. The last but not least important point that we have to notice is the
third eye in the forehead.21

The Chatracan deśvara

 image in the Paśupati temple has four hands
holding a trident and some unidentifiable object to the left, and a rosary
of rudraaks a beads and a citron to the right. The lower part of the body is
clothed in the skin of a tiger, while the upper part of the body is shown
naked. An important point to be noted is that this image bears the sacred
thread Y a feature lacking in the Mathura image. It is represented with
the third eye in the forehead, with matted hair on the head, and the
membrum virile upraised. We do not know the precise date, but the

19
See above footnotes 12 and 13.
20
Svacchandatantra 11.295cdY297cd:
kaladyavaniparyantam gahaneśsadinaksaye jj
nanabhuvanaviny _
asaracanadivibh
usitam_ j
sagun a dha raparyantarudraks etraj~
n_asan_ kulam jj
 _ tikarane j
gahaneśe layam yati mulaprakr
_ _ _
ratryante jayate bhuyo gahaneśapracodana t jj
BWhen the day of Gahaneśa declines, [all of ] the tattvas, beginning from kala down
to the earth, [the whole] populated by Rudras and Kśetraj~nas of the dominion from the
gunas down to the base, [and] adorned with a pattern of various worlds, dissolve into
_
Gahaneśa, the cause of promordial matter. They emerge again when the night declines
at the instigation of Gahaneśa.[
Parakhyatantra 5.152cYd:
mayakhye gahaneśsanah sarvarudramaheśsvarah jj
BGahaneśana, the overlord_ _ in the land of Maya.[
of all Rudras, [reigns]
21
Bhandarkar 1931: 8.
214 DIWAKAR ACHARYA

original image of Chatracan deśvara

 in the Paśupati temple, installed by
the Paśupata teacher Pranardanapran a, seems to have been replaced by
the present image in mediaeval times. _ The original may have been
destroyed in some natural calamity or during the Muslim invasion. Still,
it is possible that the attributes of the original image were copied onto
the present image. Whatever the case, the present image must be regarded
as necessarily following a visualization of [Chatra]candeśvara.22 He is
represented in almost each and every Śiva temple of Nepal _ _ in this form.
No image of him is found in seated posture, and absence of a hatchet is
distinctive in Nepalese sculptures of Candeśvara. However, his asso-
__
ciation with a hatchet is found in earliest surving sources as Goodall and
23
others have pointed out.
The Svacchandatantra and Jayadrathayamala state that Can d a and
all other teachers in the gurupan_ kti are visualized as having three eyes
and being stainless as pure crystal, but they do not provide any detailed
description of their appearance and attributes. The Niśsvasatattvasam-
hita24 and Sv ayambhuvas utrasamgraha25 name Candeśvara a bearer of_ a
hatchet, though they do not offer _ a visualization. _ _
The Pratisth alaksanasarasamuccaya, a Śaiva text on the erection of
temples composed__ _ _ the twelfth century,26 ordains erecting a shrine
before
of Candeśvara near every temple of Śiva, and itself offers a visualization
of Can_ d_ eśvara. According to it, Candeśvara should be sculpted with four
heads _and
_ four hands, bearing in succession
__ a trident, rosary, water-pot
and hatchet. Further, the image bears the third eye in the forehead, and
matted hair and the crescent moon on the head, and features serpents as

22
It is noteworthy that the image of Chatracan d eśvara in the courtyard of the Paśupati
temple and the image of Canda on the Mathura pillar _ _ bear some common elements. Both
_ _
images are shown standing and holding some unidentifiable but similar object, the hair
on their head matted and the third eye in their forehead clearly visible.
23
Goodall 2005: 170, fn. 121.
24
Mulasutra 5.22cd: aparedyur yajed devam candlsam t an_ kadh arinam j
25
Sv
ayambhuvasutrasamgraha 15.7cd: tarpayed _ n_ a
_ _ yaj~ _sesena cand_ esam tan_ kadh
ari-
namj _ _ _ __ _ _
_ 26 See Bühnemann 2003: 10. Bühnemann writes, BThe date appears in the colophon as
yugm a castaśs
l ti = (N.S.) 288.^ However, the complete date is Friday, the 8th of the
bright-half_ _of A sadha, [Nepal Samvat] 280 (1160 AD). Here is the colophon verse which
gives the date of_ copying:
_
yate samvatsaraśatayugme c ast l tiyute śuciśukle j
aśs
astamy am atha bhargavavare _śs_astram hy etad alekhi samastam jj
__ _ _
 THE ROLE OF CANDA ETC. 215
__
ornaments and as a sacred thread.27 However, the same text in another
place recognizes Can d eśvara as one of the attendants of Śiva, and
prescribes that the deity be depicted white, with two hands, three eyes
and matted hair, and that it bear a terrifying staff or a hatchet and a
rosary.28
Recapitulating the above evidence: Can d a or Can d eśvara, one of the
ascetic forms of Śiva,29 was a particular sectarian deity of the Paśupatas,
at least of the Vaimalas of Mathura30 and Mun d as and Śr n_ khalikas of
Nepal, and was honoured by some of them as late as _ _ the seventh
_ century. By
that period, the cult of Can d eśvara was well advanced and the deity was
sculpted in different forms with different attributes. There is literary
evidence that there were more shrines in older times dedicated to
Can d eśsvara. Kalidasa knows of a candeśsvarasya dhaman (Meghaduta
_ _ an a mentions more than once
33) in Ujjayinl, while the original Skandapur
an abode of Śiva ( ayatana) called Dr mican d_ eśvara, most probably the
_ __

27
Pratisth alaksanas arasamuccaya 17.16Y18:
__ _ _
dhyayen nav ambudabh asam caturvaktram caturbhujam j
tryaksam candrajatajutam d _ lpt
asyam phan _ ikan_ kanam jj
_ _ _ _
sulatan_ kodyatam hastam kamandalvaks _ _ _ alikam_ j
am
_ _ _ _  _
mahoragopav lta m ca prapanna
 rtivinaśinam jj
krtv a pr anair yath atmastham cand am tatra niveśayet j
The_ same visualization
_ is found _ _ in _ the
_ Pa~ nc
avaranastava of Aghoraśiva (Goodall
et al. *2005: verse 102): _
krsnam sudamstracatur ananam induc ud am
_a_rd
ś _ ulacarmavasanam
_ _ __ jatilam trinetram_ j _
t an_ kam ca śulam atha_ kun _ d ikay
_ a_ksam alam
_cand evvaram _ smara karair _ _ _ caturbhih
dadhatam _ jj
28 Pratis_ _ th alaksanas arasamuccaya 6.183: _ _
śvetas _ _ tryakso_ _dvibahuś ca jat l tanka k sam alikah j
pracand adandadh arl ca k aryaś _cand e_ _v varo mah _ an jj
_ _ _ _ _
I follow the suggestion of Goodall et al. (*2005:172) and interpret ca in the sense of va.
29
It is possible that some other ascetic forms of Śiva were also worshipped.
Dindimunda, an ascetic form of Śiva with shaven head, is mentioned in the original
__
Skandapur _ _ana and various Śaiva texts. He looks forward to the Sv ayambhuxvasu
trasamgraha_ who appear in Nepalese inscriptions. Very noteworthy is the fact that the
_
Svayambhuvas utrasamgraha replaces Dindimund a with Candlva.
30
The names of Uditacarya’s teacher_ _and the _ _ latter’s teacher
__ given in the Mathura
inscription are Kapilavimala and Upamitavimala respectively. Bhandarkar records these
names without the last component, vimala, regarding this latter as an adjective. But Bakker
(2000: 6) has taken this word, commonly met with as the last component in the names of
ascetics, as an indication of a connection with the Vaimala school of the Paśupatas, one of
the Paśupata schools mentioned in the Niśv asaguhya and Svacchandatantra.
216 DIWAKAR ACHARYA

shrine of Can d eśvara in the Dr mitlrtha mentioned in the Mahabharata,31 a

place where some river meets_ the sea. Nityanathasiddha, the author of the
Rasaratn akara, mentions a shrine named Amaracandeśsvara on the bank of
the Arabian Sea.32 In all likelihood, Canda was a_ _pre-Lakullśa deity ac-
__
cepted in all schools of the Paśupatas. Afterwards he was downgraded to a
minor deity but remained in Śaiva temples as one of Śiva’s ganas. Ac-
cording to the Ś _ 34 it
Sivopanisad 33 and the Pratis thalaksanasarasamuccaya,
_ _ _ _ _
is necessary to erect a shrine of Candeśsvara of the same size as the inner
sanctum of the main temple and install _ _ an image of the deity alongside
almost each and every public temple of Śiva.35
This phenomenon can be taken as an attempt of Saiddhantikas to
subjugate the supreme deity of the Paśupatas in the subsequent phase
of Śaivism.36 In the school of Lakullśa Paśupata, Candeśsvara was identified
__
31
Bisschop 2006: 220. Another interpretation of the name could be the shrine of
Candesvara where he resides with the goddess Dr mi.
_ _ Rasaratn
32
akara I.7.38: paścimasamudrasya_ tate  amaracan deś
 svaro n
ama deva-
tayatanam tasyagre valukamadhye. . . .
33 _
ŚSivopanis ad 2.18ab:
sth _ d eśsam aiśsany
anam can am pusp ar amam tathottaram j
BThe place of _ Can _ _ deśsa is in the _north-east
_ and_ the garden in the north.[
34 
Pratis th
 a lak san
 _ asa rasamuccaya 16.213:
mukhalin_ ge sahasr ahve śatahve vyaktasamj~ nake j
avyaktakhye sthir arc ay
am can d akl ptiś
s ca _ yate jj
vaks
 __ _
BThe installation of an image of Can da is taught in_ connection with [the consecration of ]
a mukha-, sahasra-, śata-, vyakta- and _ _ avyaktalin_ ga, and with any fixed image.^
Pratis th
 alak sa
 nas
 arasamuccaya 17.13Y14:
prasadagarbhamanena k rtv  adau can daveś
 s ma ca j
candeś svaram  vyavasthapya aiśsany  jagatl 
am a (sic) bahih jj
BHaving first built a shrine for Canda of the same size as the inner sanctum of the main

temple in the outer north-east corner, one should install an image of the deity [in it].^
35
An image of Candeśsvara is still commonly found in Nepalese Śiva temples, even

though the general populace takes the image of Candeśsvara to be Kamadeva, not
recognizing the ithyphallic element as a symbol of yogic_ _achievement. This has deluded
even some art historians.
As Dominic Goodall has kindly informed me, Bevery Śaiva temple [in South India] has a
Candevvara shrine in the north-east and people today end their circumambulation of the

temple there^ (personal communication). In South Indian mythology, he is born a brahmin
boy who attained a divine status with his extraordinary devotion to Śiva. This myth is
represented by Candeśsanugraha images found in South Indian Śaiva temples. According to
Goodall, these images _ _ Blink the old Sanskritic figure who receives nirmalya with the old
South Indian super bhakta^ (personal communication). For further details on Candeśsvara in
South India, see Goodall et al *2005: 169Y172. __
36
A similar attempt can be seen in the Niśv asaguhya’s account (12.17Y18) of the
origination of the Śaivas from the top-most Iśana face and of the four divisions of
Paśupatas from the other four faces of Śiva.
 THE ROLE OF CANDA ETC. 217
__
with Lakullśa, or else some of the attributes of the former were transferred
to the latter. There is thus all reason to identify the sculptures or figures
previously assumed to be standing figures of Lakullśa37 as depicting rather
Canda or Can d eśsvara.
_There
_ still remain a few crucial questions. Why is Lakullśa not
mentioned in the Mathura pillar inscription or in the Chatracan de  svara
pillar inscription, even though Kuśika is mentioned in both.38 Why did the
Paśupatas derive their line of succession from Kuśika, not from Lakullśa, if
Lakullśa was the guru of Kuśika and propounded or systematized the
Paśupata system? This fact casts suspicion on the authenticity of the Puranic
account of the Paśupata teachers, which shows Kuśika as one of the direct
disciples of Lakullśa.39 As Bisschop writes, Bthe only probable historical
figure of the earliest stage of Paśupata history is Kuśika, while identification
of his teacher _ Fthe lord_ (bhagavat) according to Kaun d inya’s Bhas ya Y as
Lakullśa was established afterwards and is therefore a later invention.^40
The Puranas are not aware of three of the four major schools of the
_
Paśupatas mentioned in the Tantras, namely Vaimala, Pramān n a and Karuka.41
Musulendra, who was a direct disciple of Lakullśa, according to Ks emaraja,

37
The four squatting figures upon the rim of the Nand column or lin_ ga can also be
regarded as representing ascetic forms of Śiva. Maxwell (1988: 12Y13) unconvincingly
 The 
suggests taking these as r sis. urdhvaretas or ithyphallic appearance is a special
attribute of Śiva, one not applicable to other divinities or to the r sis.
 The utkutik  asana is
another of the postures associated with Can d eśsvara (see Liebert 1976: s.v. can d eśsvara).
Finally, the ascending sequence of divinities below each of these figures on the shaft of
the Nand column are more likely to symbolize the guardians in the various cosmological
spheres described in Śaiva cosmology than the progressive meditative states.
38
The Mathura pillar inscription refers to Bhagavat Kuśika, and while Kuśika is not
directly mentioned in the Chatracan deś
 svara pillar inscription, the Paśupata teacher Pranarda-
naprana,
 the donor of the inscription, is called a descendant of Kuśika (kauśika).
39
The Puranic account of Lakullśa and his four disciples seems to be the earliest such
account of the Lakullśa Paśupata-s. However, this account is not reliable, at least to judge
by Kaundinya’s Bhasya on the P aśupatasutra, which does not refer to this scheme.
_ _ gives the legend
Kaun dinya _ of the Lord’s incarnation in a Brahmin body and also refers

to his disciples, starting with Indra and Kauśika. Note that he mentions Indra in the first
place and Kauśika only in the second, whereas the former cannot be found in the Puranic
account, not even in the extended list of seventeen and eighteen t l rtheśas. For the
references to the Purana-s and subsequent sources, see Lorenzen 1991: 180Y181,
footnotes 39Y43; Bisschop _ 2006: 44Y50.
40
Bisschop 2006: 47Y48.
41
Niśvasaguhya 12.17Y18, Svacchandatantra 11.71Y72, Siddhayogeśvarl mata 1.18
(Karuka and Prama na only), Pratis th
 alak sa
 nas
 arasamuccaya 16.212 (Lakula and Karu-
ka only). Pratis tha _ sanasarasamuccaya
lak and Siddhayogeśvar l mata both read K araka
  
for Karuka.
218 DIWAKAR ACHARYA

and propounded an independent Paśupata school called Mausula,42 is not

included in the Puranic account either. Thus it is possible that the
Puranic account of the Paśupatas was composed to establish the su-
premacy of Lakullśa at a time when Lakullśa was at the peak of his
popularity, by placing him at the top of the list of succession and
excluding the records of other schools and sects of the Paśupata system.
Whether this is true or not, it is certain that Lakullśa was not popular
enough to be sculpted and deified till the sixth century.

42
Ksemaraja on Svacchandatantra 11.71.
_
 THE ROLE OF CANDA ETC. 219
__

REFERENCES

Acharya (Diwakar), 1998: FPraclna Nepalama Paśupata Mata_ [The Paśupata Sect in
Ancient Nepal]. Garima 191, pp. 85Y92.
Agrawal (R.C.), 1969: FCarved Pillar of Gupta Year 61_ Journal of the Oriental Institute
XIX, M. S. University of Baroda, p. 355.
Bakker (Hans), 1997: FTala Revisited’ in Maurizio Taddei and Giuseppe de Marco, South
Asian Archaeology 1997. Volume III. Rome: Instituto Italiano per l’Africa e l’Oriente,
pp. 1115Y1170.
Bakker (Hans), 2000: FSomaśarman, Somavamś  sa and Somasiddhanta. A Paśupata
tradition in seventh-century DaksinaKoala
 : Studies in the Skandapurana III_ in
Ryutaro Tsuchida and Albrecht Wezler, Har anandalaharl: Volume in honour of
Professor Minoru Hara on his Seventieth Birthday. Reinbek, pp. 1Y19.
220 DIWAKAR ACHARYA

Bhandarkar (D.R.), 1931: Mathura Pillar Inscription of Chandragupta II: G.E. 61.
Epigraphia Indica XXI, pp. 1Y9.
Bhas  savall Vol. I: ed. Nayanatha Paudela
 avamś  . Vol. II: ed. Devlprasada Lamsala.
Puratattvaprakaśanamala 21 and 38. Kathmandu: Nepala Rastriya  Pustakalaya, 1963
and 1966.
 al ðKrsnaprasa
Bhattara  adaÞ,1988: Skandapur anasya
 Ambikakhan da
 h . ed. Krsna
 Prasada
:
Bhattaral. MahendraYRatna Series, no. 2. D ang. (Nepal), Mahendra Sanskrit
University.
Bisschop (Peter), 2006: Early Śaivism and the Skandapur  : Sects and Centres.
ana
Groningen, Egbert Forster. Groningen Oriental Studies XXI.
Brahman dapur
  . ed. J. L. Shastri. Delhi: Motilal Banarsidass, 1973.
ana
Bühler (G.) and Indraji (Bh. L.), 1880: Twenty-three Inscriptions from Nepal. Indian
Antiquary 9, pp. 163Y194.
Bühnemann (Gudrun), 2003: The Hindu Pantheon in Nepalese Line Drawings: Two
Manuscripts of the Pr a tisth
alaksanas
  a rasamuccaya. Varanasi: Indica Books.
Dyczkowski (Mark S.G.), 1989: The Canons of the Śaiv agama and the Kubjik a Tantras
of the Western Kaula Tradition. Delhi, Motilal Banarsidass.
Ganakarika of Bhasarvajña with the Ratnat lka and four appendices. ed. Chimanlal D.
_
Dalal. Baroda 1920. Reprint 1966.
Gode (P.K.), 1953: Studies in Indian Literary History, Vol. I, Bombay, Bharatlya Vidya
Bhavan.
Goodall, Dominic et al.: *2005, See Pa~ nca varanastava
 .
Hazra (R.C.), 1975: Studies in the Puranic  Records on Hindu Rites and Customs. 2nd ed.
Delhi, Motilal Banarsidass.
Jayadrathayamala. The first satka  . NAK MS 3Y358, NGMPP Reel No. A 995/6.
PalmYleaf, Folios 394, Nagari script.
Kiranatantra
 . ed. and trans. Dominic GOODALL. Bhatta  _s Commentary on
 Ramakan tha
the Kiranatantra
 . Volume I: chapters 1Y6. Critical edition and annotated translation,
Publications du département d’indologie 86.1. Pondicherry, Institut français de
Pondichéry/ Ecole française d’Extrême-Orient, 1998.
Kramrisch (Stella), 1976: The Hindu Temple. 2 vols. Delhi, Motilal Banarsidass.
Kurmapur  . ed. Pañcanana Tarkaratna. Calcutta, Van_ gavasl Press, 1885.
ana
Liebert (Gösta), 1976: Iconographic Dictionary of the Indian Religions: Hinduism-
Buddhism-Jainism. Leiden, E.J. Brill.
Lin_ gapur ana
 . ed. Jlvananda Vidyasagara. Calcutta, Nutana Valmlki Press, 1885.
Lorenzen (David N.), 1991: The K apalikas and K al
amukhas: Two Lost Śaivite Sects. 2nd
ed. Delhi, Motilal Banarsidass Publishers.
Matan_ gap arameśsvar agama, Vidyapada, with commentary Y V RT  T I of Bhatta
 Ra makan tha
 .
ed. N. R. Bhatt. Publications de l’IFI No. 56. Pondicherry, IFI, 1977.
Matan_ gap arameśsvar agama, Kriyapada, Caryapada and Yogapada, with the commentary
ð-VRTTI
 Þ of Bhatta  Ra makan tha
 up to kriyapada 11:12b, ed. N. R. Bhatt. Publications
de l’IFI No. 65. Pondicherry, IFI, 1982.
Matsyapur ana. ed. Jivananda Vidyasagara, Calcutta: the editor himself, 1876.

Maxwell (T.S.), 1988: Viśvarupa. Oxford University South Asian Studies Series. Delhi,
Oxford University Press.
 THE ROLE OF CANDA ETC. 221
__
Mrgendratantra
 , Vidyapada and Yogapada, with the commentary ð-VRTTI  Þ of Bhatta
__
Ramakantha, ed. Madhusudan Kaul Śastra. KSTS 50. Srinagar, 1930.
__
Niśsv  ta. NAK MS 1Y227, NGMPP Reel No. A 41/14. Palm-leaf, Folios
asatattvasamhi
117, early Nepalese (Kuti  la) script. Described by Hara Prasad Śastrl (1905:lxxvii and
137Y140). There are two apographs available, both in Devanagari and on paper: NAK
MS 5Y2401, NGMPP Reel No. A 159/18, and Welcome Institute for the History of
medicine, London, Sanskrit MS I.33. The verse and Chapter numeration used in
footnotes is that of Goodall’s edition in progress.
Pa~nc avaranastava
 . The Pa~
nc avaranastava
 of Aghoraśivac
arya: a Twelfth-Century South
Indian prescription for the visualization of Sadaśiva and his retinue. edited and
traslated with annotation by Dominic GOODALL, Nibedita Rout, R. Sathyanarayanan,
S.A.S. Sharma, T. Ganesan and S. Sambandhaśivacarya. Draft. *2005.
Par..akhyatantra. ed. and trans. Dominic GOODALL. The Par akhyatantra, a scripture of the
Saiva siddhanta. A critical edition and annotated translation. Collection Indology 98.
Pondicherry, Institut français de Pondichéry/Ecole française d’Extrême-Orient, 2004.
Pasśupatasutra. With Kaun dinya_s
 Pa~nc
a rthabhasya Commentary. ed. R. Ananthak-
rishna Sastri. Trivandrum Sanskrit Series, no. 143. Trivandrum, University of
Travancore, 1940.
Pathak (V.S.), 1960: History of Śaiva Cults in Northern India from Inscriptions.
Varanasi, Ram Naresh Varma.
Pratisth
 alaksana
 sarasamuccaya. 2 vols. ed. Damodara Śarma and Babukr_s_na Śarma.
Kathmandu, National Archives, 1966Y68. Consulted Manuscript: National _ Archives
Kathmandu, Palm-leaf dated NS 280, Manuscript no. 5Y316, NGMPP reel no. A 57/29.
Regmi (D.R.), 1983: Inscriptions of Ancient Nepal. 3 vols. New Delhi, Abhinava
Publications.
Shah (U.P.), 1984: FLakullśa: Śaivite Saint_ Discourses on Śiva, ed. M.E. Meister, pp.
92Y102. Philadelphia, University of Pennsylvania Press.
Siddhayogeśvarlmata. 1-Asiatic Society of Bengal. Manuscript no. G 5465. 2- National
Archives Kathmandu. Manuscript no. 5Y2403.
Sivopanisat. Printed (pp. 324Y378) in Un-published Upanishads, ed. the Pandits of
Ś
Adyar Library under the supervision of C. Kunhan Raja. Adyar Library Series Vol. 14.
Madras: Adyar Library and Research Centre, 1933. Consulted Manuscript: National
Archives Kathmandu, Palm-leaf, Manuscript no. 5Y738, NGMPP reel no. A 11/3,
172Y189.
Skandapur ana
 . Volume I Adhyayas 1Y25 Critically Edited with Prolegomena and
English Synopsis. ed. Adriaensen, R., Bakker, H. T., and Isaacson, H. Groningen,
Egbert Forsten.
Svacchandatantra with the commentary (-UDDYOTA) of Ra janaka Ksemar a ja. ed.
Madhusudan Kaul Śastrl. KSTS 31, 38, 44, 48, 51, 53, and 56. Bombay 1921Y35.
Sv a yambhuvas u trasamgraha
 . National Archives Kathmandu, PalmYleaf, Folios 36,
Manuscript no. 5Y348, NGMPP reel no. A 30/6.
Tantrikabhidhanakośa II. 2004. A Dictionary of Technical Terms from Hindu Tantric
Literature. sous la direction de H. Brunner, G. Oberhammer et A. Padoux.
Östereichischen Akademie der Wissenschaften, Philosophisch-Historische Klasse,
Sitzungsberichte, 714. Band. Beiträge zur Kultur- und Geistesgeschichte Asiens 44.
Wien: Verlag der Östereichischen Akademie der Wissenschaften.
222 DIWAKAR ACHARYA

Vajracarya (Dhanavajra), 1973: Licchavik alak

a Abhilekha [Inscriptions of Licchavi
Period]. Kathmandu: Nepal and Asian Studies Institute, Tribhuvan University, 1973
(V.S. 2030).
V
a yupur ana 
 . ed. Hari Na ra yana
 Apte. Anandaśrama Sanskrit Series, no. 45. Poona:
Anandaśram Press, 1905.

Graduate School of Letters,

Kyoto University,
Yoshida-Honmachi, Sakyo-ku, 606-8501 Kyoto,
Japan
E-mail: acharyadiwakar@hotmail.com