Divinities

Identity and Divinity: Boundary-Crossing Goddesses in Medieval South India
Author(s): Leslie C. Orr

Source: Journal of the American Academy of Religion, Vol. 73, No. 1 (Mar., 2005), pp. 9-43
Published by: Oxford University Press
Stable URL: http://www.jstor.org/stable/4139876
Accessed: 15-01-2017 13:02 UTC
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact [email protected].
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms
Oxford University Press is collaborating with JSTOR to digitize, preserve and extend access to Journal
of the American Academy of Religion
This content downloaded from 14.139.224.146 on Sun, 15 Jan 2017 13:02:56 UTC
All use subject to http://about.jstor.org/terms
Identity and Divinity:

Boundary-crossing Goddesses
in Medieval South India

Leslie C. Orr
I seek in this article to consider the extent to which schemas of classification
current in popular and scholarly understandings of India's religions may

be misleading as guides to the situation in Tamil Nadu in the period of
the eighth to thirteenth century. By examining the myths, iconography,
inscriptions, and placement in ritual settings of female deities, we are led
to the realization that various distinctions among types of divinity that
seem to us self-evident may have had little meaning for those who
worshipped these goddesses a thousand years ago. I suggest that a more
correct understanding of the significance of these goddesses in their own
historical context must involve serious reconsideration of the salience of
sectarian categories-Hindu or Jain, Saiva or Vaisnava-and the organization of deities into central or peripheral objects of worship, local or
universal manifestations of the divine, and representatives of dangerous
or beneficent powers.
Leslie C. Orr is an associate professor in the Department of Religion at Concordia University,

Montreal, Quebec, Canada H3G 1M8.
Many thanks are due to Corinne Dempsey for organizing the panel for the annual meeting of the
American Academy of Religion, which first brought this group of articles together. The interchanges
with my fellow panelists that have resulted from our collaboration have been of great value to me,
and I am indebted to them all, not the least for their patience. Without the techno/iconographic
assistance of Tanisha Ramachandran none of this would have been possible. As always, my gratitude
to the director, M. D. Sampath, and the staff of the Office of the Chief Epigraphist in Mysore,
is immeasurable. It is also a great pleasure to acknowledge the insights, information, and
encouragement that I have received, which were of help to me in developing the ideas presented
here, from Sarah Caldwell, John Cort, Phyllis Granoff, Jon Kalina, Lois Martin, Anne Monius,
Charlotte Schmid, Martha Selby, Ranvir Shah, and Devesh Soneji. The epigraph is from Adam
Gopnik's Paris to the Moon (New York: Random House, 2000).
Journal of the American Academy of Religion March 2005, Vol. 73, No. 1, pp. 9-43
doi:10.1093/jaarel/lfi003
@ The Author 2005. Published by Oxford University Press, on behalf of the American Academy of
Religion. All rights reserved. For permissions, please e-mail: [email protected]
10 Journal of the American Academy of Religion

It was not just that you could not see the trees for the forest. It was that
you could not see the forest because it was covered by a map.
-Adam Gopnik
MAPS AND MODELS OF various kinds appear to be indispensable

for the religious studies scholar, in ordering the complexity and ambiguity that he or she so often confronts. These tools of systematization
allow us to make comparisons, to investigate interactions, and to chart
historical change. But although they are useful in providing points of
reference and orientations within our field of study, even the best map
cannot tell us what it is that we are looking at. I seek in this article to
consider the extent to which schemas of classification current in popular
and scholarly understandings of India's religions may be misleading as
guides to the situation in South India a thousand years ago. By examining
the myths, iconography, inscriptions, and placement in ritual settings
of female deities, we are led to the realization that various distinctions
among types of divinity that seem to us self-evident-even our catego-
ries of "Hindu" or "Jain"-lack salience from the perspective of those

who worshipped these goddesses. If we cannot depend upon our familiar
paradigms to help us discover the identities of deities, clearly we must
also raise the question of whether the maps we have been using to articulate the boundaries between religious communities can be regarded as
reliable.
The landscape of early medieval South India is filled with various

deities and religious activities. The dynamism and diversity of the
period of the eighth to thirteenth century in the region today known
as Tamil Nadu is attested by the material remains that have come

down to us. We have, for example, many beautiful bronze images of
deities and of Jain enlightened ones (Tirthafikaras) that were taken
out in procession on festival days. There are richly carved cave
temples from the early part of this period and, somewhat later, structural
stone temples dedicated to the gods Siva or Visnu, or to one of the

Tirthahikaras. In time, these shrines grew into large complexes, surrounded by concentric enclosure walls surmounted by tall, ornately
decorated gopuras over entrances at the four cardinal points, within
which subsidiary shrines, storage areas, kitchens, and pavilions were
built. Although these images and structures-which today we would
classify as "Hindu" (or, more precisely, "Saiva" and "Vaisnava") and
"Jain"-form by far the greatest part of the artistic inheritance from
the past and will be my primary concern in this article, there is also
evidence from this period of Buddhist and Muslim worship going on,
Orr: Identity and Divinity 11
particularly in settlements along the east coast of T

such as Kanchipuram, Kayalpattinam, Nagapattinam,
The art historical and inscriptional evidence o

much of the devotional literature, testify to a rit
shared in common among the worshippers of var
Whether an image or a shrine were Saiva or Vaisnava

set up, daily worship was offered-involving bathing
image and presenting food and flowers-and festiv

Schooled as we are in the important theological and
ences that set these religious traditions apart fro
rather puzzling to find so little that is distinctive in

cially when the inscriptional record makes it clear th
was of central importance to members of religiou

merely a concern of the "masses."2 One source of ou
an exaggerated sense of the coherence and fixity
Srivaisnavism, and South Indian Digambara Jainism a

systems, when in fact the fitting of fluid and varian
tices into ordered and consistent structures was a pr
from complete and was in the midst of being negotia
and place. The abundant outpouring of didactic, apol
tional, hagiographical, and commentarial literatur

Tamil and in Sanskrit, provided the charter docu
definition of these philosophical schools and these re

but it is only in retrospect that we can see that this
religious world was experienced by those who were p
offered worship in these past times, definitions of re
have been rather different. Indeed, I would like to su
number of those medieval worshippers there was not
ness in terms of notions of coherence within and ad
Vaisnava, or Jain "system" or community but a g

these categories, despite the fact that the contempo
ture exerted itself considerably to make these dema
ever well-known and useful this classificatory schem
of us who study the history of Indian religions or f
1 On the commonalities among medieval South Indian Hindu, Jain, and
see Orr 1999a and 1999c. See Monius for an excellent treatment of Buddhism in Tamil Nadu. For
Islam in Tamil Nadu, and the sharing of forms of worship, devotional idioms, and ritual space by
Muslims with Hindus, see Bayly; Richman; Saheb; Narayanan.
2 Schopen has made a persuasive case for the central role of monks and nuns in the historical
development of Buddhist worship practices. For the importance of devotion and image worship
within Jainism-for Jain ascetics and teachers, as well as for the laity-see Zydenbos; Orr 1999b;
Carrithers; Cort 2002.

participants in contemporary religious life, this way of mapping the diversity of religious beliefs and practices in fact may not have guided many of
those people of the past whose identities and activities we seek to describe.
In this article I will examine and compare a number of different types of
goddesses, who seem to have had an important place in early medieval

Tamil Nadu, including various goddesses whom we would now classify as
Hindu and those goddesses found in the Jain context, who resemble them
closely, known as yaksis. My effort will be-in response to Corinne Dempsey's
initial invitation-to see this world of the past through the eyes of yaksis.3
Would the yaksis recognize the categories that we employ as scholars to
analyze the nature and scope of divinity in their world-our sectarian
classifications, the division between local and universal manifestations of

the divine, the separation of dangerous from auspicious or salvific power,
or even the distinction between human and divine? Do the yaksis acknowledge the boundaries that we use to parcel out their religious landscape?
TALES OF ENCOUNTER
In medieval Tamil literature we find some evidence that these g

would have been familiar with sectarian divisions and that they
fact participants in religious turf wars. For example, in Nilakec
composed in the ninth or tenth century, we find the story of the co
of a bloodthirsty goddess to Jainism. This goddess, Nilakeci,
devote the rest of her life to propagating this religion. But Nilak
not pit herself against the followers of Siva and Visnu; her eff
almost entirely dedicated to the refutation of Buddhist beliefs an
tices. The association of goddesses with South Indian Jains and the
felt by Jains toward Buddhists are also brought out in the story of t
monk Akalafika. In this tale Akalafika is called upon to debate w

Buddhists in the court of the king of Kanchipuram in the northe
Tamil Nadu. With the aid of the yaksi Ambika of the Jain temple in
Karantai, Akalafika defeats the Buddhists, who are expelled from the
3 For the idea of considering deities-and their material forms-as subjects, I am in
Richard Davis's presentation in Lives of Indian Images.

4 That part of the city of Kanchipuram that is sacred to the Jains is known as Tiruparu
The goddess Ambika at Tiruparuttikunram, Karantai, and at many other places in the Ta
bears the name Dharmadevi. The Tamil version of the tale of Akalafika is found in the Mackenzie
manuscript collection, and there are local traditions at Jain temples in and around Kanchipuram
that celebrate Akalafika's victory over the Buddhists and his devotion to the goddess
(Ramachandran: 42; Ekambaranathan: 111). Meanwhile, another version of the story appears in the
Sanskrit Kathdkosa, composed in the eleventh century; in this telling, which is set in Kalifiga (Orissa),
Akalafika's yaksi ally is Cakresvari, while his Buddhist opponent is aided by the goddess Tdra
(Granoff 1985: 461-462).
The special relationship between Jains and ya
in one of the episodes of the twelfth-centu
Purdnam, although in this case the goddesses are

faith against the challenge of an opponent. The st
the city of the far south that was the capital of the P
Purdinam we see the Jain ascetics who inhabit th

city filled with foreboding, as the dream of one
"A sannyasini [kuratti, a Jain religious woman]
[Jain] pundits coming out of their caves rode the
after them ran bejewelled Iyakkis [yak.sis] how
(Vampardvariva.ntuccarukkam: 638, trans. by T. N

is an omen of the impending downfall of the Jain
be defeated by the Saiva saint Tirufianacampantar
the Pandya king, who orders them to be impa

hostility toward the Jains echoes what is expr
Tirufianacampantar himself five centuries earlier.
Perhaps the Saivas labored to differentiate them
because of the resemblances that seem to unite them. These affinities are
brought out, for example, in a comparison of the stories of the Saiva saint
Karaikkal Ammaiyar and the Jain yaksi Ambika, which brings to light the
great similarity in the value placed by Jains and Saivas on devotion, almsgiving, and asceticism. Both tales recount the tribulation and vindication
experienced by a virtuous housewife who had offered a mendicant food
that was meant for another purpose. In the story of Karaikkal Ammaiyar,
told in Periya Purdnam, a mango given to a devotee of Siva had to be

miraculously replaced by Lord Siva himself when our heroine's husband
asked for it to be served at his midday meal. Perceiving its divine origin,
he was seized by terror, and fled to another country. Years passed before
Karaikkal Ammaiyar learned that her husband no longer regarded her as
his wife but rather as a goddess. She then requested that Lord Siva release
her from her beautiful womanly body and give her the skeletal form of a
ghoul (pjy), so that she might become one of Siva's host and worship
him eternally. The story of Ambika begins with her offering to a Jain
mendicant the food meant for the Brahmins attending the ?rdiddha cere-
mony for her husband's ancestors. When he found out what she had
done, he forced her out of the house. Eventually, she, together with her
two sons and an attendant woman, found refuge in a place high up on
the hill in the shade of a tree. Because of her great spiritual merit, this tree
became a "wishing tree," which provided for all of their needs. Meanwhile, events transpired at home that proved her virtue and her husband,
overcome with remorse, came to find her. Seeing him approach and
certain that he wanted to punish her further, she climbed to the peak of

the hill and leapt to her death. In the very next instant she was reborn as a
yaksi named Ambika. Her husband, seeing her yaksi form, was at first
filled with fear, but then, realizing his wife was dead, he killed himself just
as she had done, and was reborn as a lion, who lay at the feet of the goddess.
Thus the lion became the "vehicle" of Ambika and the goddess herself
became the attendant of the Tirthafikara Neminatha.5
In both the Jain story of Ambika and the Saiva story of Karaikkal
Ammaiyar the "sectarian conflict" that provides the dramatic impetus
toward transformation is not, for example, between Jain and Buddhist or
between Saiva and Jain but rather between a woman's devotion and her
wifely duty. In both cases the resolution of this conflict involves getting
rid of a husband, by terrifying him and "converting" him into a worshipper. Although it is more overt in Ambika's story, for both of these tales
the competing religions (defined largely in terms of practice rather than
belief) are, on the one hand, a normative Brahmanical orthodoxy and, on
the other, a liberating spiritual path-represented by the heroine's singleminded service of the mendicant, the way of life of the mendicant himself, and the transformation of the heroine from a housewife to a divine
or non-human being privileged to eternally serve her Lord.
GODDESS IMAGES
Although Ambika's story culminates in her transformat

a-sana devatd, an attendant or guardian deity, in the religi
medieval Tamil Nadu she most often appears as an ind
goddess. In figure 1 we have such an image, carved in th
century on a rock face in the hills of Kalugumalai, some eigh

the south of the city of Madurai. Here we see several of the e
Ambikd's story, including the goddess's lion mount, her two
behind her, the "wishing tree," bearing mango fruit. The p
and autonomy of this figure are typical of the way yaksis are
in Tamil Nadu and at variance with the pattern generally f
where in India, in which yak.sis, paired with their male count
yaksas, most often appear as small attendant figures flan

Tirthafikaras. Images of yaksis of medieval Tamil Nadu a
' This version of Ambikd's story is based on T. N. Ramachandran's telling (157-159),

from a manuscript found at the Vardhamdna temple in Tiruparuttikunram. Ambik
most important objects of worship in this temple, and tradition has it that her image
from the temple dedicated to the goddess Kdmiksi in Kanchipuram (Soundara Raj
The Tamil story of Ambikd is very similar to that found in the fourteenth-centur
Vividhatirthakalpa (Granoff 1990: 182-184).
?'OP
tV
role
jrr
owe
rr/AP
.7)
f.s
.00;5
Ol
W,
Alpr
;r
?E-
rr
-10
vftv-
Figure
graphs
1.
Ambika,
are
by
the
au
S.
!- I
Alt
i,.
1 ' *
4(dD
6p
Figu
(Sou
de
-1
!_C~
rm
Figure 3. Pdrvati standing with att
(Tiruchirappalli district), tenth centur

the Government Museum, Chennai.
more independent of their Tirthaftkaras, and more prevalent than in

other Jain contexts.
The iconography of the South Indian yaksi also sets her apart from
the Jain goddesses of other parts of India-and suggests instead a connection with the Hindu goddesses of the South. The two bronze processional images from the Madras Museum reproduced in figures 2 and 3
show a striking resemblance to one another, despite the fact that they
were crafted three centuries apart. Of the two, the image of Ambika
(figure 2) is more recent, dating from the thirteenth century, and stands
just over twenty inches high. The figure of Parvati (figure 3) is slimmer
and taller, and the sculpture as a whole is larger, with a height of thirtysix inches. Both goddesses stand on a lotus base, and their clothing and
adornments are the same, with the exception that Ambika's crown carries
an image of a Tirthafikara. Their postures and gestures are identical, with
one hand raised as if delicately holding something (a lotus blossom?
a bunch of young mango fruits?) and the other resting lightly on the head
of a diminutive female attendant. Ambika is also accompanied by one of
her sons.
Jain yaksis appear not only in the form of graceful and gracious
goddesses but also as warriors. At Kilakkudi, just outside of Madurai,
is a cave sanctuary in which the images of several Tirthalikaras and two
goddesses are carved high up on the rock face. The image furthest
within the interior of the cave (figure 4), which dates from the eighth
or ninth century, has been variously identified as the Jain yaksi

Ambikd (Srinivasan 1975: 228) or Siddhayika (Desai: 58-59) or as the
Hindu goddess Durga (Shah: 287). The panel is worn, but we can
make out the form of a four-armed goddess mounted on a lion and

poised to release an arrow from her bow. Her opponent, who is placed
before and somewhat below her, is of smaller size than the goddess
but, like her, seems to wear a conical crown. He is mounted on an
elephant, which is being attacked by her lion; he has one arm upraised
with a sword in his hand and with his other hand holds a shield before
him to protect himself from the goddess and her weapons. When we
compare this panel with the famous image of Mahisdsuramardini-
Durga as the slayer of the buffalo-demon-from Mamallapuram
(figure 5), carved a century earlier and three hundred miles to the
northeast, we see a similar composition and especially compelling cor-
respondences in the posture and gestures of the two goddesses. The

Mamallapuram panel is more filled with figures and with movement,
and the depiction of the buffalo-headed demon is certainly very different from that of the opponent of the warrior goddess of Kilakkudi. But
the goddess of Mamallapuram herself, mounted on her lion with arms

Ih~-n?~L)
~? ;
rr?
:r r
R
'
i?,Fck?~,, I~
r
'
?'t? `i~~
?r
rYY
i.
.?
r?
.1
ZrJ
?r3
?r
ii
8
s;
?c.??. r
r
??
..
~??
r?
4iCB; r, G.
JC
`'"
"i~t?,
-r
C:?:
?C
~.~
4 ???~Ci
C ??
I
??
,:
r
~
i?
'
,*
::
''
?r??
?'
2~.5
ic
sL,`
l?i
C?
tr.
~,
rr
?,)
F:
P'
E r~ ?i~t r? -?r?
*
..
s
..
i.
;V
c.
-,?
?r.
k:
~j~j
?? z
?~J
tr
..
'1;
r?
i',
)t~Y~
;1
C? ); s -~4Y
i'
)
?r
Y. ~Y..
..
t.
?
n
't??~,
1
rr. I
?1
:'r
.??
;t,,
1 ?;~j~1
= -\?..
(r:
"
rl
..
`?
?t
Figure
century.
outstretched to let loose her arrow, seems to be closely related to her

Jain counterpart.6
Powerful, multi-armed goddesses--both Jain yaksis and goddesses with
Hindu associations, including MahiSasuramardini-are more often depicted
in medieval South Indian art in hieratic poses that evoke their status as
objects of worship than in narrative images that represent them in the midst
of battle. Figure 6 shows us the yak4i Padmdvati of Kalugumalai, where we
have already encountered Ambika (figure 1). Like Ambika, Padmivati's

image is placed amid carved stone panels of the Tirthafikaras. The four-
armed Padmivati is flanked by two female attendants, she is crowned, and

behind her head is a serpent's hood. Her right leg is drawn up on her seat and
6 Another type of image of the Hindu warrior goddess, which is more commonly met with in early
medieval South Indian art, is that of Durg5 as an armed goddess standing beside her lion mount, with
her leg raised and her foot planted on the lion's back (Tartakov and Dehejia). It is interesting to
discover that the Jain goddess Ambikd is similarly represented in this triumphant pose, for example, in
a seventh-century image at Anandamangalam (see Shah: figure 48) and in a twelfth-century image at
Tirumalai (see Sivaramamurti 1983: plate 89), both sites in the northern part of Tamil Nadu.

~ ivt.
44~
'Ron!
ir
Figure
seventh
her
and
left
cont
while
In
le
figu
legs
are
lower
h
"fear-n
noose-w
is
found
known
attenda
her
fee
Buddhis
is
not
c
7
Because
Padmdvati
from
the
became
his
interestin
description
..
o'
rd
' O r
41
,lie,
?
,.
Figure
6.
(Ramnad
Seated
distric
goddess
may
in
the
eight
context
in
Tiruparanku
stone
beds
ca
ascetics,
as
w
tional
literat
tance
of
thi
Murukan;
m
century
con
its
great
Su
8
An
eighth-cen
shrine
(as
well
as
This
Durgi
can
Subrahmanya
an
Subrahmanya
te
dr I
41 "P
oll,
.-. 6 L
"" -
'L9 ]..?LL~e ~ ~ ~ -'L9 -?~
"
_,,?e
WOO
Figure
7.
(Madurai
"Annapur
district),
frangaise
d'Extreme
complex,
cut
in
central
deities
ar
Whatever
"Ann
blance
to
Jain
g
Precisely
the
sam
of
the
former),
is
found
in
the
p
miles
from
Tiru
warrior
goddess
Siddhayika,
alt
The
of
presence
9
See
Devakunjari:
106
its
later
importance
as

z r-,
d
c;r
L'
r
?%
?'
~]
?;
??
;U~3~1 r?-~p-~P~~U' -r~ ~Zlf~Ti.f~~ ~trd, u~;"~

\`
~?
ill:
?'t~
1
~7df:
?.
;t
I,
ji
.X
~5r
,i
ill
r\ ..B ~??jl)
?~' r ~3ir'
r
I~
I~
`r' i'- ?)1
W'
I~ r' r?.rT
j r?C j:f 6J;iP] ''
'''
II
.,
?-
?*
?)
a~ ?.
$ ' .,'
".: ::
' 'i
fi
, 5
,
r
?t
.ai~?~
ii.
5f
tr
krY~
r
r
r
-,
t~lc;
,,
r
~t~qb
..
3
2)
(
'*~??-
:~ll~p
CP
~c
~;L~a;;
"?
r a, ?- E ~n?
'r
?+??!~L~J
~cr~,
r
*?r
:??
?
'r
r
p'~'l
r-v
;SI
.b
I li?g.re
r'
rb~A
r; ;J'?
~i~
~Y ~,' 43
~.ur~~5~
i;
:r,
1~~Ftr,)
r,
.?,
?~-?~,N~+
..
C~ilTllr-
"~
t
I
Figure 8. Seated goddess (Siddhayika? Padmavati?) with figure of a child below, ston
relief, Kilakkudi (Madurai district), eighth to ninth century.
which could be a child, suggests the possibility that this is, once again,
Ambika. Whoever she is, this Kilakkudi goddess, and the Padmdvati of
Kalugumalai and the Annapiirni of Tiruparankunram-goddesses
found in the caves and hills not far distant from Madurai-resemble not
only one another but another category of South Indian goddesses. These
are the saptamat.rks, the "seven mothers."
These goddesses seem originally to have been associated with child-
hood illness and death, although their malevolent character is not
evident even in the earliest images of these deities, from North India,
which date from the fourth and fifth centuries. The saptamdt.rks first
appear in the Tamil country in the eighth century. In Tamil Nadu there
are a few examples of temples that seem to have been dedicated to this
group of goddesses, but after the ninth century these goddesses found
their place as subsidiary deities in temples where Siva was the main object
of worship (Srinivasan 1960; Mahalingam; Shanmugan). All the images
from Tamil Nadu, whether sculpted in relief or in the round, show that,
iconographically, the goddesses have been "Brahmanized" and are associated with male deities. For example, the mdatrka we see on the left in
figure 9 is identifiable as Vaisnavi, because she bears in two of her four
hands the emblems of conch and discus that belong to Visnu. Each of the
other six goddesses in this group from the temple at Arakandanallur,
flanked by images of Siva and Ganesa, has attributes that link her to a
male counterpart. With respect to the other features of these images, the
goddesses are virtually indistinguishable from one another. And so, too,
do they bear a very close resemblance to other saptamTt.rkd images
throughout Tamil Nadu, as well as to the seated yaksis to whom we have

just been introduced.'0 What is distinctive about the Tamil iconography
of both saptamdtrkas and yaksis is the absence of children held in the
goddesses' arms. Outside of Tamil Nadu the saptamatrkds are very
frequently depicted with children in their laps.
If such iconographic indications of motherhood are entirely-and

surprisingly-lacking in images of the "seven mothers" in Tamil Nadu,
they are quite rare in representations of other goddesses from this region
as well. In contrast to Ambika's portrayal with one or two sons seated in her
lap, which is commonly met with throughout the rest of India and which is
prescribed in the South Indian Digambara texts (Ramachandran: 209),

we have seen that in Tamil Nadu the Jain yaksi's offspring are positioned
somewhat off to the side, small crowned figures standing stalwartly, if a
'o Mishra has remarked on the similarity between medieval matrka figures and those of Jain yaksis
in central India, as has Mitra for Orissa. In both cases, however, the iconography is very different
from what is seen in Tamil Nadu.
- ?ti -,
IM.
Figure
ninth
bit
TP7
9.
Three
century
of
(?).
awkwardly,
bronze
from
S
Parvati's
son
Sk
Somdskanda.
In
(Uma),
and
the
raised
to
hold
a
South
Indian
re
mother-and-ch
busy-they
hav
activities
wants
to
to
perf
sit
dow
"
The
earliest
Somask
at
the
Kailasandtha
t
which
were
frequentl
common
in
temples
b
images
with
Skanda
st
(L'Hernault
1978:
66-
26 Journal of the American Academy ofReligion
in Tamil Nadu with respect to maternal elements (or their lack) strongly
suggests that the relationships among the female deities with whom we
are concerned here must be considered within the framework of a particular local history of religions and iconographic conventions. Given the
distinctive, and largely childless, character of these images, the resem-
blances among Jain and Hindu goddesses cannot be accounted for by

some general, pan-Indian process of Jain "borrowing" from the Hindu
pantheon, as has often been suggested, or by postulating that the Tamil
mt.rkiis and yaksis have descended from the same distant "mother god-
dess" ancestresses.
One interesting image that we do find of a mother is that of Jyestha.
Her children-a bull-headed son and a daughter--do not sit on her lap
either, because they are grown up, but sit or stand on either side of her
(figure 10). But there is little that is typically maternal in this goddess.
Jyesthd is the older sister of the goddess Laksmi and is regarded as being her
opposite; if Laksmi brings gladness and abundance, Jyestha is the embodi-
ment of inauspiciousness. The crow, harbinger of misfortune, is her
emblem. Yet despite her apparently undesirable qualities, Jyestha was very
widely worshipped in Tamil Nadu in the period of the eighth to eleventh
century (Srinivasan 1960: 26-29; Mahalingam: 29-30). In fact, this popularity seems to have given rise to feelings of rivalry among the devotees of
other deities. For example, in his poem "Tirumalai," the ninth-century
Tamil Vaisnava saint Tontaratippoti decries the worship given to Jyestha; it
is interesting to note that in the same poem he also has harsh words for the
Jains and Buddhists.12 In the period when this poet-saint was writing, one
of the places where we know that Jyestha was worshipped is a site where
reverence was also offered to other goddesses, to Visnu, to Siva, to
Subrahmainya, and to the Tirthafikaras. This place is Tiruparankunram,
outside Madurai, where there is an eighth-century inscription recording the
establishment of a shrine for Jyesthd (see note 8 above), and where still
today we can see an image of this goddess with her characteristic form
(Pattaviramin: 55 and plate 152/1). It is unlikely that Tontaratippoti was
familiar with patterns of worship at Tiruparankunram, but perhaps the
kind of devotional eclecticism or inclusivism prevalent here was something
he discerned, and disapproved of, elsewhere in the Tamil country.13
12 Nladyira Tivviyappirapantam. That there were Saivas who were similarly hostile to Jyestha's
worship, and who may have classed her together with Jainrand Buddhist goddesses, is suggested by
the Sanskrit Saiva text Lifigapurina, which depicts Jyegha as being comfortable among Buddhists
and naked Jain monks, while she makes every effort to avoid true religious learning and sacred
objects and actions (Leslie: 113-123).

13 Michael Carrithers calls this type of religious cosmopolitanism "polytropy" and considers it "to
be the norm, rather than the exception, in South Asia" (834).
"c`C.
~r
Lb'~'
,,
?,
r?
~si~ i4r
3--a,?1
~;;;~
''
?~
?f;
i~
J~?"
ilr f
i*
rt
'''~?is
~g~ tL'i' fi ,- t
,
t
-;
?;
w P~I ,~?i ?.3?. 1
~t
r
~~
CI
.'
r.
~liz-?? ? i
:X?~
,r
~ii"-
-i
r'4
r. ?? '3r; h
Ir
b3
~'r-
Or''
rJ
~?
ylll
',v"
~D~C~rl~Z
,a-?~
+
C,
Irt?'
I r~PE~ic- L~bc: -r
~I?fj4~ 4a,
t~
~k~S
ii
t.
I
~?
.?
rP
ii
I?;
,?
~ ~1?,
?r~
~*.
F
~"?;
u:
,t
?9
II
cl
"r~
~
C
,.
?,
~Q?
i~ 9r*?
a~t~
-?;
,,?
a?a~';,i
r;r
?I~C1I~P~L-
:c-~i~L'
a?
~b
"'
i-
~s~
sl
't~
~t~'r:.7
;-'t." ' ap?r..rte~ ~Yr~b~';"t~,~?FIIR~~ =~t~l~~ ''
;r,
~B~M
a~i~L~
*1?
i,,
I-?
trl-:~sr*~
?;
I?
Figure 10. Jyegtha, with her children seated on either side, stone relief, Tiruvengaivasal
(Tiruchirappalli district), ninth century, Pudukkottai museum.
?I

NAMES AND PLACES
We have seen that there are some striking parallels in the
which Jain and Hindu goddesses were represented in early mediev

Nadu-although these similarities may not have been apparent, or
cant, from the point of view of the worshippers who gazed
images a thousand years ago. By turning to the evidence of the

tions, we are given a glimpse of how these goddesses were regar
how they were approached by the people of the past. The first th
we discover is that Jain and Hindu goddesses (and gods, Tirthafik
Buddhas) were worshipped in the same manner. The inscriptions
in contexts that we would today classify as Hindu or Jain, tell us that
goddesses were set up as stone or metal images, that they were giv
and ornaments, and offered worship (arccanai). The question the
If their images resemble one another and their ritual treatment wa
cal, were there any distinctions made between these two categories
and Hindu that to our modern eyes are so clearly separate? D
language of the inscriptions provide us with the means to differ

between these goddesses' identities and associations?
In the inscriptions the vast majority of the goddesses are name
are identified solely with reference to their locale. For exam

goddess at the Jain site of Chitaral-who was presented with
ornaments-is simply called the goddess of Tiruccaranam, and the
used is pa.tari, "Mistress," in its honorific plural form (TAS 1: 1
More often, Jain goddesses (i.e., goddesses mentioned in insc
from Jain sites) are referred to as yaksis (Tamil iyakkiyar), with no f
identification except in some cases, again, the connection with a p

place.14 I have come across only three named Jain goddesses: the
Avvai in a ninth-century inscription from Aivarmalai in Madurai
(SII 14.22); Varasundari in a tenth-century inscription from
(TAS 4.41); and the yaksiy-dr Malkaiyar-nayakiyar, who is ment

two inscriptions of the late twelfth century from Perumandur
Arcot district (SII 7.846 and SII 7.848). It is significant that none o
names correspond to those of the Jain yaksis as we know the
textual sources.
14 Using Jaina Inscriptions in Tamilnadu (JIT) as a guide, I have located nineteen Jain inscriptions
that refer to goddesses. Of these, eleven use some form of the term yaksi. Three of the yaksis are
identified with reference to their locale. There is also one reference to an image of the "golden yaks!"
(ponni yakkiyarpatimam) (El 4: 136-137). Four of the inscriptions use the term pa.tdriyar, pi.tdri, or a
similar term-in two cases in combination with yaksi-and two refer to the goddesses as bhagavati.
There are three inscriptions that simply refer to an "image" (tirumeni) or "shrine" (tevdram).
There are many more inscriptions that come

contexts than from Jain ones, and we have ther
number of references to goddesses whom we would

is important to note, however, that a large propo
inscriptional references to goddesses-those dating
tenth century-come from Jain rather than Hindu s
of references to goddesses in Hindu inscriptions dat
century onward and are particularly common in the
Like the Jain yaksis these deities are rarely referred
when they are, we are likely to find their names fam
(Tamil Tukkai) are among the more frequently
More common than such particular names are terms
or "goddess" such as pirdt.tti, or patadri and its varian
ized bhat.tari. The great majority of these godde
inscriptions of the thirteenth century, are referred to a
often with the specification of which divine Lord the

the face of it, this type of nomenclature would seem
establish the identities of these goddesses, according
described as the consorts of Siva or Visnu. Male deit
scarcely ever identified in straightforward sectarian

named in the inscriptions as the "Lord of such-andare hints in the inscriptions that may help us determ
in question is Siva or Visnu (or a Tirthafikara or a B
fact that the inscriptions do not highlight the sectarian
and that we have to do the sleuth work to figure it o

us that we are concerned about issues other than th
the authors of the inscriptions.
The inscriptions can tell us a few things about wh
Hindu goddesses were-that they were sometimes
sovereignty over a particular locale, or that they we

but it seems that the inscriptions are of little use in
boundaries and finding a place for the goddesses with
epigraphical evidence seems to aggravate boundar

help resolve them, when we consider that both Jain
were referred to by the term patdiri/pitadri and its var
title, meaning "Mistress" or "Lady," is also found
15 Given the enormous number of medieval inscriptions from Hindu

have used various sampling methods to learn something about these g
on my survey (for Orr 1999b) of inscriptions from several small areas
Nadu, based on the summaries found in A Topographical List (TL) and on
entire corpus of inscriptions from eight particular temples that are the f
inscriptions as part of the names of Tirthaflkaras, in the masculine form
patidrar-"Master" or "Lord." Jain inscriptions dating from the ninth

and tenth centuries refer to goddesses as pi.taris, as do inscriptions from
the Hindu context, particularly in the tenth and eleventh centuries. On
the Hindu side, the term is applied to a variety of goddesses, including
Uma, Durga, the saptamit.rks, and Kali. The fact that the term piptdri
means "village goddess" in contemporary Tamil usage has created some
misunderstandings about the character of the deities so designated in
medieval inscriptions. For example, Suresh Pillai (441-442) proposes
that the piptiris mentioned in the eleventh-century inscriptions of the

Rdjarajeivara temple at Thanjavur were local goddesses who were being
incorporated into the Saiva pantheon through assimilation to Siva's consort Uma. But at nearby Tiruvidaimarudur, in exactly the same period,
Uma herself was referred to as a pi.tdri (SII 23.278). It is clear that we are
not dealing here with a folk or village deity but with a goddess belonging
to the "great tradition."
The change in meaning of the term pitadri, over the last thousand
years, points toward the possibility that a process of "de-Sanskritization"
has been taking place, through which some of the goddesses we have
been discussing were transformed over time from great goddesses to
village deities (cf. Padma). It has been suggested that certain Jain yaksis
experienced such a diminution of divine status and authority: contempo-
rary South Indian village goddesses who-unusually-do not accept
sacrificed animals may be the descendents of vegetarian Jain goddesses
(Subramaniam: 95-96). The history of the saptamit.rkas of the Tamil

country seems to present another case of de-Sanskritization. I have
already referred to the fact that there were temples dedicated to the
saptamit.rkifs in the early medieval period in Tamil Nadu. But these
goddesses very soon became only parivdira devatis, "deities of the entour-
age," in a corner of the temple complex organized around the central

shrine dedicated to Siva. From the eleventh century onward some of
these goddesses were displaced even from this position; the images of
saptamit.rkis were removed and installed as guardian deities in small vil-
lage temples, becoming pitidris in the modern sense of the word.16 Similarly, when we first encounter Jyesthd in the Tamil country, she seems to
16 Saptamdt.rkds appear to have been the main focus for worship in temples at Alambakkam and
Nangavaram in Tiruchirappalli district, Velacheri, Tekkar, and Uttaramerur in Chingleput district,
Perunkanchi in North Arcot district, and Siddhalingamadam in South Arcot district. Evidence of
their placement as parivdra devatds in the ninth to eleventh century may be found at Kilur,
Tiruvisalur, and Tiruvannamalai, and many other temples. L'Hernault describes the removal of
saptam.trkas from Saiva temples and their installation as village deities, called pitadris. See
Mahalingam: 28-29; Baskaran; Shanmugam; Kaimal: 217-218; L'Hernault 1993: 49; Orr 2000.
be worshipped as an independent deity, as we
Tiruparankunram, near Madurai. In the next phase

a parivdra devatJ in Siva temples but then, after the
this role to her auspicious younger sister Laksmi. W
is suggested by the fate of one image of Jyestha, dati
ninth century, which is today found stuck into the

the southwest corner of the town of Uttarameru
village's temple for Jyestha should be located, as presc
the traditional manuals for artists and architec
tradition, she is the bearer of misfortune, and if t

upright, calamity would strike the village (Gros an
POSITIONING DIVINITY
In the case of Jyegsha, the saptamatrkas, and "vegetarian" v

desses, we see how important the matter of placement and pos
the understanding of a deity's significance, and even her ver
Where a goddess is may be the determining factor in establishi
is. Why do we classify a particular figure as Jain or Hindu? So
we have seen, the identification cannot be made on the basis
iconography or terminology. Then we look at location. A g

Jain temple is Jain ... A consort deity in a Siva temple is Par
so on. But how do we classify the temples themselves?
The Bhagavati temple in Chitaral in Kanyakumari distr
southern tip of the subcontinent, provides an example of the

we encounter in trying to make sectarian assignments. Carved
the rock face on the north side of the temple are a number of
tenth-century figures of Tirthafikaras along with images of
who are identifiable as Padmavati and Ambika.17 But from th

the thirteenth century the temple has been described in insc
being dedicated to Bhagavati, the name of a goddess usually r
belonging to the Hindu pantheon (TAS 1: 194; TAS 1: 297 /TA
the beginning of the twentieth century, when T. A. Gopinath
17 The goddess whom we may identify as Padmavati stands beside "her" Tirthaflkar
and Ambika is accompanied by her lion mount, two sons, and a female attendant (
in TAS, vol 2: 125-127; Sivaramamurti 1983: plate 95). The image of Ambika at Ch
very closely that at Kalugumalai (figure 1 above). There are several inscriptio
boulders a little distance away that provide some designation or name for goddesses at
is impossible to determine which images they are referring to. Two of these inscriptio
late ninth century, use the term patfdriyidr (TAS 1: 194-195; TAS 4.40), and a tenth-c
inscription describes the construction of a stone doorway for the shrine of the godd
(TAS 4.41). Beside the north rock face, where the Jain images are carved, is t
structural temple, which is built out from a cave facing toward the west.
the site, the images under worship within the Bhagavati temple were old
stone images of the Tirthafikaras Mahavira and Pargvanatha and a plaster
figure of a goddess; they were regarded as Hindu deities at that time, as
they continue to be (TAS vol. 1: 193-194; see also TAS vol. 4: 147). But
Gopinatha Rao considers that this temple's shift from being a "Jain" to a
"Hindu" sacred site is a relatively recent development and that in the

thirteenth and fourteenth centuries the object of worship referred to in
the inscriptions was, in fact, "a Jaina Bhagavati" (TAS vol. 2: 125).
Many of the other early medieval goddesses we have been considering
found themselves in contexts that, from our perspective, are similarly
problematic. Often these goddesses were worshipped in cave temples,

and cave temples have several features that make it difficult to determine
the sectarian character, or the central focus of worship, of a particular
sacred site. The cave temple layouts of early medieval Tamil Nadu
obstruct our efforts to identify the "main deity" because of the fact that
they frequently are made up of multiple shrines, or have reliefs of diverse
deities carved side by side on the rock face, and because there is often no
clear line of demarcation between the inside and the outside of the temple,
or an obvious center. Cave temples that were excavated and consecrated
in the seventh, eighth, and ninth centuries are spread throughout the
Tamil country. There are many sites where both Siva and Visnu were
honored; there are places where Jaina, Saiva, and Vaisnava deities co-
existed; and in many cave temples female deities, including the
saptamittrkds and Durga (or warrior goddesses who might have had this
name) as well as yaksis, are prominent. Tiruparankunram, near Madurai,
is an excellent example of such a site, and there are many other places of
worship that have a similarly mixed character (see Sivaramamurti 1961;
Pattaviramin). I think we tend to assume that the peculiarities of cave
temples are due to the challenges involved in working with solid rock, in
excavating the side of a hill. We have the idea that were it not for these
difficulties, the people who built these temples would really have preferred to construct a "regular" temple-that is, the type of temple that is
characteristic of the later period in South India, with a small enclosed
sanctum at its heart surrounded by symmetrically arrayed enclosure walls
with towering gopuras over the entrances piercing the walls.
But I would like to suggest that these cave temples may actually represent what was a "regular" temple for the early medieval period. Structural
temples from this time have been subject to renovations and expansions
in subsequent centuries that have radically re-shaped the space of the
temple. Those temples that have remained more or less intact are very
simple structures, and much of what we now consider part of the standard lay-out of the South Indian temple, including the demarcation of
the sacred space with enclosing walls, is the resul

took place in the twelfth, thirteenth, or later cen
suggests elsewhere in this set of articles that w
around the eighth-century Kailasanatha temple is
and Harle believes that the gopuras that surmoun
the typical temple in Tamil Nadu were in the first
than gateways (Harle: 10-16). Brunner (21) con
Indian Saiva temple as it has taken shape (and even

the Agamas) shows signs of being a composite, wit
the character of the central sanctuary and that of
in niches on the temple walls and in the shrines ar
Even in the case of such a venerable institution as
Chidambaram, there are scholarly disagreements ab
inal" center of the complex, and a number of othe
country show signs of having been made up of se
were at one time independent of one another (see
Not only are we deterred from making definit
about which deity is the main one in the case of mult
we also frequently find that medieval South Indian

multi-templed, just like many of the cave sites. Ins
and thirteenth centuries at the Nataraja temple of
the names of at least twenty-nine other temples in
(Orr 2000). And the author of an old chronicle of U
at the time that Rajendra Chola re-organized this s
village, it contained thirty-eight temples-twelve d
to Visnu, and fourteen to various goddesses (Gros a
It seems likely, therefore, that the cave temple
arrangement, while the more centralized mode

temple, and the South Indian temple town, may
times. When multiple shrines came to be ordered i

there was a singular central focus of worship and o
formed into attendants or consorts or parivdra d
possible to define whose temple this was.
DISCUSSION AND CONCLUSION
Are the goddesses we have been considering here "subsidiary

devotion? Is there actually anything about the ways they are
iconographically, the ways they are named in inscriptions, o
they are positioned at sacred sites that indicates that they w
to the worship focused primarily on the Tirthafikaras, o
Visnu? There seems to be little evidence that this is the case fo
these goddesses. But, in fact, does our ability to assign these goddesses
to a particular sectarian milieu-as Hindu or Jain goddesses, Saiva or

Vaisnava-not depend on our classifying them as subordinate deities?
I believe that if we really attempt to regard these goddesses with the eyes of
those who worshipped them a millennium ago, we will be forced to abandon
several of the categorization schemes that we commonly use to understand
India's pantheon of divinities and systems of religious practice. The division
into sectarian categories and the organization of groups of deities into central
and peripheral objects of worship appear "natural" to us today-we who are
scholars and teachers, or contemporary Hindus or Jains-but these conceptual structures may be less useful than we would have thought in making
sense of India's religions in early medieval Tamil Nadu.
It is evident, when we look at the array of texts produced in Tamil
and Sanskrit in this time and place, that there were indeed spiritual aspirants and devotees whose religious focus was fixed on the great saints and
gods, the Tirthafikaras, Siva, and Visnu. But the literature that they
composed and transmitted raises certain questions about what precisely
sectarian identity, or sectarian community, meant to them. Some of their
poems and stories indeed reveal a sense of rivalry or competition with
those adhering to other religious viewpoints. Yet when we put these

kinds of texts side by side, it becomes evident that the sectarian hostility
is not reciprocated. The Saivas are engaged in struggle against the Jains,
while the Jains ignore the Saivas and instead seek to defeat the Buddhists.
What is revealed here about the construction of sectarian boundaries is
that it was a unilateral project.

It is hard to avoid the conclusion, therefore, that if we have a map
marked out with divisions between Jain, Saiva, and Vaisnava territories,
this is mostly our map and not one that represents something that very
many of the Jains, Saivas, and Vaisnavas of early medieval Tamil Nadu
would have agreed upon. When we look specifically at the character of
the goddesses who were worshipped in the past, it is particularly difficult
to try to find a place for them on such a map. The resemblance among
the images of goddesses whom we would today categorize as Jain or

Hindu, the vocabulary used in the inscriptions to name them, their
appearance in diverse ritual settings, and the stories told about them
make the task of assigning them to their "proper" place immensely com-
plicated. The inscriptional nomenclature eschews sectarian categories,

and the iconography does little to clarify matters. If the "Pdrvati" we see
in figure 3 had been moved to a Vaisnava or Jain temple, instead of to the
Madras Museum, would she have been out of place as an object of
worship or a participant in festival processions? Images of goddesses that

cannot be moved, because they are carved into the stone of cave temples,
are no easier to pin down with regard to their sectar

is the Chitaral Bhagavati? Is the Kilakkudi warrior go
deity? Does Mahisdsuramardini of Mamallapuram
constellation of deities, or might she be a Vai
"Annaptirn~" of Tiruparankunram Saiva, Vaisnava, o

tions are impossible to answer-and probably pointle
I believe that we have to be open to the idea that t
regarded by their worshippers "simply" as goddes
Jain, or Vaisnava goddesses. Those who paid homage
a particular locale in the Tamil country may have
male deities or Tirthafikaras who had a place in her
or in the stories told about her, and they may also
I have done here, whenever I have referred to "Ambi
"Durga") to goddesses elsewhere with whom they we
attributes, names, and myths were similar. Or they
her primarily as a goddess whose unique character
dent of such associations.
These various possibilities bring into question not only the salience of
sectarian classifications in medieval Tamil Nadu but also the categorization
of gods, and especially goddesses, as either "local"/ "village" or "universal"/
"pan-Indian" deities. We owe this differentiation not so much to historical
study as to relatively recent observations and analysis of religious practice
by missionaries and ethnographers in the early part of the twentieth

century.19 They have described village religion as being quite distinct
from orthodox Hinduism and have speculated that it reflects a "Dravidian"
or pre-Aryan legacy. This archaic religion is held to be dominated by the
worship of goddesses, who were associated with agricultural fertility and
who were unpredictable or dangerous and demanded animal sacrifices.

Such ideas about primordial goddess worship in South India, and about
the "essential" nature of South Indian goddesses, have been widely
accepted by later scholars. For example, David Shulman's work on late
medieval Tamil temple myths highlights the theme of the marriage of
a local goddess to the great god Siva; his analysis reinforces the image
of the South Indian goddess as indigenous, local, popular, and ruralin contrast to male objects of worship who are Aryan, Sanskritic, or
"8 Although Mahi~ssuramardini, like other forms of the goddess Durga, is usually considered to be
a member of the Saiva pantheon, Charlotte Schmid has recently suggested that the image at
Mamallapuram and other early representations of Mahisdsuramardini have much stronger Vaisnava
mythic and ritual associations than Saiva ones (Schmid forthcoming).
19 Among others, Bishop Henry Whitehead, a Protestant missionary in South India in the early
twentieth century, has been influential in the creation of this polarized model of religious practice
and belief (Whitehead). See Orr 1999b for further discussion.
pan-Indian. The temple myths describe the process by which the independent local goddess is "domesticated" or "tamed."
But in the context of early medieval Tamil Nadu is the distinction
between "local" goddesses and "universal" gods a legitimate one? The

answer to this question is a definite "no." The inscriptions make it clear
that male deities, as much or even more than their female counterparts,
were regarded as belonging to a particular place and that their identities
as manifestations of pan-Indian deities were of very little concern to their
worshippers. It seems likely that both gods and goddesses were (and still
are) understood as being both local and translocal.20 The next question is
whether the Tamil temple myths reflect an actual historical process in
which local goddesses were "tamed" and admitted into a Brahmanical

pantheon as consorts of the great male deity. Although it is probable that
the shift in the dominant perception of a goddess as being "local" or

"pan-Indian" did in some cases move in the direction indicated by Shulman,
we see precisely the opposite development in the cases of Jyestha and of

the saptamat.rkds-where pan-Indian goddesses were transformed in the
course of time into village deities.
The simultaneity and bi-directionality that characterize apprehensions of goddesses as local or universal lead us also into a reconsideration
of the value of categorizing the goddesses of medieval Tamil Nadu in

terms of the polarities of benevolent and dangerous. The "split-goddess"
model-in which goddesses of a beneficent nature tend to be depicted as
the consorts of male deities, whereas goddesses with destructive powers
are independent of male divinity-has been recently subject to a good

deal of critical scrutiny (see e. g., Erndl). The material presented here
adds to the evidence that such distinctions are frequently quite meaningless. Most of the goddesses whose images we have seen are not consort
deities, yet none of them conform either to the stereotype of the wild,
terrifying, bloodthirsty independent goddess. Ambika is depicted both as
a gracious and beautiful young mother (although without a husband!)

and as a formidable warrior. The powerful and independent Padmavati
of Kalugumalai and "Annapiirna" of Tiruparankunram, and even the

saptamit.rkais, are represented as benevolent, auspicious deities, bestowing
their blessings on their worshippers.
20 Richard Cohen's analysis of the character of the Buddhist yaksi Hdriti at Ajanta shows how a
goddess might be regarded simultaneously as "a local yak.sini, a translocal Buddhist protector, or
even a great bodhisattva" (382-383). A number of recent studies of Hindu goddesses, like those by
Erndl and Humes, show how worshippers in India today understand the object of their devotion as
"Local Goddess Yet Great Goddess," as Humes's article is entitled.
Incongruity between the image of a deity and

tives associated with that god or goddess is a well
throughout India.21 The presence of two Tirthafik
in the Bhagavati temple at Chitaral is an exam
across in Tamil Nadu, and it seems likely that ther

that relate directly to the (other?) goddesses who
stories we have considered here. Sometimes such a situation is said to be
the result of historical processes that are termed assimilation, accultu-
ration, incorporation, containment, or conversion-implying that a
deity is brought into the sphere of influence of a new, or more powerful,

god. But it is not clear how useful these words are. What little explanatory
significance they have seems to rest on the assumption of continuity and
conscious strategy, although more often what we observe is, in Meister's
words, a "nonlinearity of use and of interpretations" (133-134). The

other problem is that these terms imply the abandonment of the previous
character and affiliations of a deity in favor of a reconfigured identity; the
goddess who is "tamed," for example, is assumed to have been wholly

transformed. This seems highly unlikely, in light of the fact that iconographic, ritual, and mythic elements associated with the deity's earlier
persona are conserved and continue to define her.
What is required is the recognition that at every moment in history the
yak.sis and other goddesses have had a history and the appreciation that in
any particular time and place they are composite beings.22 If we are ever to
succeed with a mapping project, it is, first of all, necessary to look at specific
local and historical contexts and to try to understand what categories were
used to make sense of those contexts. Was the question "is she Hindu or is
she Jain" a meaningful one in medieval Tamil Nadu? It is not, and so we
must come up with better questions. Further, as we seek to define the identities of deities and of their devotees, we must embrace complexity and multi-
plicity. This means we should be drawing various maps and be willing to use
them in alternation with one another, and to lay them aside altogether when
they are not helping us to get anywhere. A goddess might be many things at
once, to different worshippers at one and the same time.23 But she may also
have multiple meanings for a single worshipper, whose vision of the goddess
21 See, for example, the cases of the Didarganj yaksi (Davis: 3-6) and the goddess Sacciya-mata at
Osian (Meister; Cort 2000).
22 Carrithers urges scholars to avoid analyses that depend on the concept of syncretism and instead
to regard any religious context, past or present, as "already presenting an array of holy persons and
of worshippers moving between them, already fecund with elaborations" (836).
23 As John Cort (2000) says: "Who a deity is, whose a temple is, who worships a particular deity,
and why that worship is performed - these are all questions that rarely if ever admit of single
answers.

would be colored by awareness of the mythic, ritual, nominative, and contextual aspects of this deity. Through this worshipper's eyes the divine presents a
richly and variously hued appearance, which, predictably, is not reducible to
a single name or meaning. So, too, should we, as students of the history of
religions, seek to glimpse the significance of religious practice and experience
outside the bounds of our accustomed categories. But in order to see, it is

necessary to look up from the map.
REFERENCES
Primary Sources and Reference Works:

Tamil Inscriptions and Tamil Literature
ARE
Annual Reports on Indian Epigraphy. 1905-1978.
Delhi: Manager of Publications. Transcripts of

the inscriptions abstracted in the ARE were
graciously made available to me at the Office of
the Chief Epigraphist, Archaeological Survey of
India, Mysore.
El
Epigraphia Indica. Calcutta/Delhi: Director

General, Archaeological Survey of India, 1892.
JIT
Ndildyira
Jaina Inscriptions in Tamilnadu (A Topographical

List), A. Ekambaranathan and C. K. Sivaprakasam.
Madras: Research Foundation for Jainology, 1987.
Tamil text. Chennai: Lifco, 1984.
Tivviyappirapantam
Nilakici
Tamil text. Tirunelveli: South India Saiva
Siddhanta Works Publishing Society, 1964.

Neelakesi
The Original Text and the Commentary of Samaya-
Divakara-Vamana-Muni. A summary-translation
by A. Chakravarti. Kumbakonam: author, 1936.
Periya Pura-nam
(Tirutton.tar Mdkkatai)
Tamil text. Tirunelveli: South India Saiva
Siddhanta Works Publishing Society, 1977.
Periya Puranam Translated by T. N. Ramachandran. 2 volumes

Thanjavur: Tamil University, 1990 and 1995.
SII
South Indian Inscriptions. Volumes 2-24. Delhi:
Director-General, Archaeological Survey of
India, 1891-1983.

TAS
Travancore Archaeological Series. Volumes 1-8.
Trivandrum, 1910-1938.
Tevaram
Tamil text. Edited by T. V. Gopal Iyer. 3

volumes. Pondichery: Institut frangais d'Indol-
ogie, 1984-1991.
TL
A Topographical List of Inscriptions in the Tamil
Nadu and Kerala States, ed. T. V. Mahalingam,
volumes II, V, and VI. New Delhi: Indian Council

of Historical Research, 1988-1991.
Secondary Sources
Baskaran, S. T.
1974
"The Saptamat.rikis of Perunkanchi." Lalit Kala

17: 14-17.
Bayly, Susan Saints, Goddesses and Kings: Muslims and

1989
Christians in South Indian Society 1700-1900.
Cambridge: Cambridge University Press.

Brunner, H1dene
1990
"L'image divine dans le culte agamique de Siva:

Rapport entre l'image mentale et le support
concret du culte." In L'Image Divine: Culte et
mediation dans l'hindouisme, 9-29. Ed. by Andre
Padoux. Paris: CNRS.
Carrithers, Michael "On Polytropy: Or the Natural Condition of Spiri2000

tual Cosmopolitanism in India: The Digambar Jain
Case." Modern Asian Studies 34/4: 831-861
Cohen, Richard S.
1998
"Ndga, Yaksini, Buddha: Local Deities and Local

Buddhism at Ajanta." History of Religions 37/4:
360-400.
Cort, John E.
2000
2002
"Multiple Visions: Experiences of Osian." Paper

presented at the University of California, Santa
Barbara's conference on "Vital, Mutual, and
Fatal Attractions: The Sharing and Contesting of
Sacred Space in South Asia."
"Bhakti in the Early Jain Tradition: Understanding
Devotional Religion in South Asia." History of

Religions 42/1: 59-86.
Davis, Richard H.
1997
Lives of Indian Images. Princeton, NJ: Princeton

University Press.
Desai, Pandurang B. Jainism in South India. Sholapur: Jaina Samskrti

1957
Samrakshaka Sangha.

Devakunjari, D.
1979
Ekambaranathan, A.
1996
Erndl, Kathleen M.
1993
Madurai through the Ages: From the Earliest

Times to 1801 A.D. Madras: Society for Archaeological, Historical and Epigraphical Research.
Jainism in Tamilnadu: Art and Archaeology. Port
Credit, ON: Jain Humanities Press.
Victory to the Mother: The Hindu Goddess of

Northwest India in Myth, Ritual, and Symbol.
New York: Oxford University Press.
Granoff, Phyllis
1985
1990
"Scholars and Wonder-Workers: Some Remarks
on the Role of the Supernatural in Philosophical

Contests in Vedanta Hagiographies." Journal of
the American Oriental Society 105: 459-467.
The Clever Adultress and Other Stories. Oakville:
Mosaic.
Gros, Frangois, Uttaramerur: Legendes, Histoire, Monuments.

and R. Nagaswamy Pondichery: Institut frangais d'Indologie.
1970
Harle, James C.
1995
Temple Gateways in South India: The Architecture
and Iconography of the Cidambaram Gopuras.
New Delhi: Munshiram Manoharlal.
Humes, Cynthia Ann "Vindyavasini: Local Goddess Yet Great

1996
Goddess." In Devi: Goddesses of India, 49-76.
Ed. by John Stratton Hawley and Donna
Marie Wulff. Berkeley: University of Califor-
nia Press.
Kaimal, Padma
1988
"Stone Portrait Sculpture at Pallava and Early

Cola Temples: Kings, Patrons, and Individual
Identity." Ph.D. diss., University of California,
Berkeley.
Leslie, Julia
1991
"Sri and Jyestha: Ambivalent Role Models for

Women." In Roles and Rituals for Hindu Women,
107-127. Ed. by Julia Leslie. Rutherford, NJ:

Fairleigh Dickinson University Press.
L'Hernault, Frangoise
L'Iconographie de Subrahmatnya au Tamilnad.
1978
Pondichery: Institut frangais d'Indologie.
1993
"Tiruvannamalai." In Temple Towns of Tamil

Nadu, 40-57. Ed. by George Michell. Bombay:
Marg.

Mahalingam, T.
"TheV.
Cult of Sakti in Tamilnad." In The Sakti
1967
Cult and Tdrd, 17-33. Ed. by D. C. Sircar.
Calcutta: University of Calcutta.
Meister, Michael W.
1998
"Sweetmeats or Corpses?" In Open Boundaries:

Jain Communities and Cultures in Indian History.
Ed. by John E. Cort. Albany: State University of
New York Press.
Mishra, R. N.
1975
Mitra, Debala
1975
Monius, Anne E.
2001
"Yakshini Images and Matrka Tradition in

Central India." Prachya Pratibha 3/1: 29-34.
Udayagiri and Khandagiri. New Delhi: Archaeological Survey of India.
Imagining a Place for Buddhism: Literary Culture
and Religious Community in Tamil-Speaking
South India. New York: Oxford University Press.
Narayanan, Vasudha "Religious Vocabulary and Regional Identity: A

2000
Study of the Tamil Cirappuranam." In Beyond
Turk and Hindu: Rethinking Religious Identities
in Islamicate South Asia, 74-97. Ed. by David
Gilmartin and Bruce B. Lawrence. Gainesville:
University Press of Florida.

Orr, Leslie C.
1999a
"Garlands, Lamps, and Drums: Temple Worship
and Medieval South Indian Society." Paper
presented at the Annual Meeting of the Canadian
Asian Studies Society, South Asia Council,
Montreal.
1999b
"Jain Worship in Medieval Tamilnadu." In

Approaches to Jaina Studies: Philosophy, Logic,
Rituals and Symbols, 250-274. Ed. by N. K.

Wagle and Olle Qvarnstrom. Toronto: Centre
for South Asian Studies, University of Toronto.

1999c
"Women's Wealth and Worship: Female

Patronage of Hinduism, Jainism, and Buddhism
in Medieval Tamilnadu." In Faces of the Feminine in Ancient, Medieval, and Modern India,
124-147. Ed. by Mandakranta Bose. New York:

Oxford University Press.
2000
"Mapping Movement through Social and Sacred
Space in Medieval Tamilnadu." Paper presented

at the 36th International Congress of Asian and
North African Studies, Montreal.

Padma, Sree
2001
"From Village to City: Transforming Goddesses

in Urban Andhra Pradesh." In Seeking Mahadevi:
Constructing the Identities of the Hindu Great
Goddess, 115-143. Ed. by Tracy Pintchman.
Albany: State University of New York Press.

Pattaviramin, P. Z.
1975
Ramachandran, T. N.
1934
Sanctuaires Ruprestres de l'Inde du Sud.

Pondichery: Institut frangais d'Indologie.
"Tiruparuttikunram and Its Temples." Bulletin

of the Madras Government Museum n.s.-general section, vol. I, part 3.
Richman, Paula
1997
Extraordinary Child: Poems from a South Indian
Devotional Genre. Honolulu: University of
Hawai'i Press.
Saheb, S. A. A.
1998
"A 'Festival of Flags': Hindu-Muslim Devotion and

the Sacralising of Localism at the Shrine of Nagore-
e-Sharif in Tamil Nadu." In Embodying Charisma:

Modernity, Locality and the Performance of Emotion
in Sufi Cults, 55-76. Ed. by P. Werbner and
H. Basu. London: Routledge.

Schmid, Charlotte
Forthcoming
"Mahisasuramardini: A Vaisnava Goddess?" In
Chidambaram Seminar Papers. Ed. by R.
Nagaswamy.
Schopen, Gregory
1988
"On Monks, Nun and 'Vulgar' Practices: The

Introduction of the Image Cult into Indian
Buddhism." Artibus Asiae 49/1-2: 153-168.
Shah, Umakant P.
1987
Shanmugan, P.
1983
Jaina-RiTpa-Mandana, vol. 1. New Delhi:
Abhinav.
"The Saptamatri Cult during Chola Times." In

Srinidhidih: Perspectives in Indian Archaeology,
Art and Culture, 257-260. Ed. by K.V. Raman,

et al. Madras: New Era Publications.
Shulman, David Dean

1980
Tamil Temple Myths: Sacrifice and Divine
Marriage in the South Indian Saiva Tradition.

Princeton: Princeton University Press.
Sivaramamurti, C.
Kalugumalai and Early Pandyan Rock-Cut
1961
Shrines. Bombay: N. M. Tripathi.
1983
Panorama of fain Art. New Delhi: Times of India.

"Deccan
Soundara Rajan, K.
V. and South India." In Jaina Art And
1975
Architecture, vol. 2, 310-332. Ed. by A. Ghosh.
New Delhi: Bharatiya Jiianapitha.
Srinivasan, K. R.
1960
1975
Some Aspects of Religion as Revealed by Early

Monuments and Literature of the South. Madras:
University of Madras.
"South India." In Jaina Art And Architecture, vol. 2,
207-230. Ed. by A. Ghosh. New Delhi: Bharatiya

Jfidnapitha.
Subramaniam, T. N.
1958-59
"Pallankovil Framentary Jaina Copper-Plate

Grant of Early Chala Period." Transactions of the
Archaeological Society of South India, 84-110.
Suresh Pillai, B. "The Raajaraajeesvaram at Taficaavuur." In

1968
Proceedings of the First International Conference Sem-
inar of Tamil Studies, vol. 1, 438-447. Kuala Lampur:

International Association of Tamil Research.
Tartakov, Gary Michael,
and Vidya Dehejia

1984
Whitehead, Henry
[1921] 1983
Zydenbos, Robert J.
1993
"Sharing, Intrusion, Influence: The

Mahisdsuramardini Imagery of the Calukyas and
the Pallavas." Artibus Asiae 45: 287-345.
The Village Gods of South India. 2d ed. New
Delhi: Cosmo Publications.
"The Concept of Divinity in Jainism." The 1992
Roop Lal Jain Lecture. Toronto: Centre for
South Asian Studies, University of Toronto.

Divinities

Uploaded by

Copyright:

Available Formats

Divinities

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Divinities

Uploaded by

Copyright:

Available Formats

Identity and Divinity: Boundary-Crossing Goddesses in Medieval South India

Author(s): Leslie C. Orr

Identity and Divinity:

in Medieval South India

I seek in this article to consider the extent to which schemas of classification

current in popular and scholarly understandings of India's religions may

historical context must involve serious reconsideration of the salience of

Leslie C. Orr is an associate professor in the Department of Religion at Concordia University,

is immeasurable. It is also a great pleasure to acknowledge the insights, information, and

Religion. All rights reserved. For permissions, please e-mail: [email protected]

10 Journal of the American Academy of Religion

MAPS AND MODELS OF various kinds appear to be indispensable

ries of "Hindu" or "Jain"-lack salience from the perspective of those

The landscape of early medieval South India is filled with various

as Tamil Nadu is attested by the material remains that have come

out in procession on festival days. There are richly carved cave

stone temples dedicated to the gods Siva or Visnu, or to one of the

Orr: Identity and Divinity 11

particularly in settlements along the east coast of T

The art historical and inscriptional evidence o

Whether an image or a shrine were Saiva or Vaisnava

image and presenting food and flowers-and festiv

ences that set these religious traditions apart fro

rather puzzling to find so little that is distinctive in

was of central importance to members of religiou

an exaggerated sense of the coherence and fixity

Srivaisnavism, and South Indian Digambara Jainism a

tional, hagiographical, and commentarial literatur

definition of these philosophical schools and these re

Vaisnava, or Jain "system" or community but a g

1 On the commonalities among medieval South Indian Hindu, Jain, and

12 Journal of the American Academy of Religion

goddesses, who seem to have had an important place in early medieval

analyze the nature and scope of divinity in their world-our sectarian

classifications, the division between local and universal manifestations of

In medieval Tamil literature we find some evidence that these g

composed in the ninth or tenth century, we find the story of the co

of a bloodthirsty goddess to Jainism. This goddess, Nilakeci,

monk Akalafika. In this tale Akalafika is called upon to debate w

3 For the idea of considering deities-and their material forms-as subjects, I am in

Richard Davis's presentation in Lives of Indian Images.

Orr: Identity and Divinity 13

The special relationship between Jains and ya

in one of the episodes of the twelfth-centu

Purdnam, although in this case the goddesses are

Purdinam we see the Jain ascetics who inhabit th

after them ran bejewelled Iyakkis [yak.sis] how

(Vampardvariva.ntuccarukkam: 638, trans. by T. N

the Pandya king, who orders them to be impa

Perhaps the Saivas labored to differentiate them

told in Periya Purdnam, a mango given to a devotee of Siva had to be

14 Journal of the American Academy of Religion

Although Ambika's story culminates in her transformat

medieval Tamil Nadu she most often appears as an ind

goddess. In figure 1 we have such an image, carved in th

century on a rock face in the hills of Kalugumalai, some eigh

in Tamil Nadu and at variance with the pattern generally f

where in India, in which yak.sis, paired with their male count

yaksas, most often appear as small attendant figures flan