Friday, 28 October 2011

Shiva as ‘cosmic dancer’: on Pallava origins for the Nataraja bronzea

Orion the hunter as per western astronomy is known to Ancient Egyptians, native South Americans and red Indians. They built several magnificent structures in praise of Orion. South Indian Dravidian style temples were based on the formations in the constellation. Indian Astronomy and mythology speaks a lot on the constellation and its adjacent constellations like Canis major, minor, Lepus, Auriga, Eridanus.


This paper explores the origins of the Nataraja cult in the state of Tamil Nadu in southern India. The trajectory of the dancing Shiva is traced: from the processional worship of metal icons outside the sanctum to the cultic elevation of the Nataraja bronze into the sanctum at Chidambaram. Archaeometallurgical, iconographic and literary evidence discussed shows that the Nataraja bronze, depicting Shiva’s anandatandava or ‘dance of bliss’, was a Pallava innovation (seventh to mid-ninth century), rather than tenth-century Chola as widely believed. That this formulation was informed of ‘cosmic’ or metaphysical connotations is also argued on the basis of the testimony of the hymns of Tamil saints.


More than other divinities of Hinduism’s trinity (including Brahma, the creator, and Vishnu, the preserver), the icons of Shiva, the great destroyer, span a range of paradoxical interpretations: from the erotic to the ascetic, the dreadful to the sublime and the human to the cosmic. Temple practice in late medieval Tamil Nadu followed the canon of Saiva Siddhanta worship, the agamas (Davis 2000: 10). Under the Cholas (tenth – thirteenth century), a dual form of Saivite worship thrived. Inside the sanctum, a static stone pillar or lingam was worshipped (which had both aniconic and phallic connotations). Portable metal images (utsava murtis) were taken out of the sanctum in procession, including kinetic depictions of Shiva’s awesome dance of destruction.

Nataraja, the subject of this paper, di ers from other images of dancing Shiva in his posture with the left leg extended across at hip level known as bhujangatrasita karana. At the extant twelfth – thirteenth century Nataraja temple complex at Chidambaram a unique

dual form of worship reaches culmination in the innermost sanctum: of a three-foot-high metal icon of Nataraja alongside an empty curtained space representing him as akasa or ether. This depiction of Shiva as the akasa lingam or pillar of space/cosmos is the Chidambaram Rahasya or revelation of Chidambaram. This golden-roofed inner sanctum is called Chit-Sabha or Hall of Consciousness.

This paper tackles three unresolved issues concerning the Nataraja bronze. The first is the date of emergence of the Nataraja bronze and its worship at Chidambaram. This has been uncertain since Hindu icons were rarely inscribed and are mainly dated with reference to stone sculpture at inscribed temples. The earliest properly three-dimensional stone Natarajas are found in temples built by the great Chola patroness, queen Sembiyan Mahadevi, such as the Kailasanathaswami temple, c. AD 940, at the place named after her. Hence, Dehejia (1990: 45 – 6) speculated that the Nataraja bronze developed under her. However, this rules out the existence of earlier images.

The second question concerns the original significance of the Nataraja bronze. Ananda Coomaraswamy (1924) put forth a lyrical interpretation drawn from thirteenth-century Tamil Saiva Siddhantic texts composed around Chidambaram such as the Tirumantiram and Unmaivilakam. He interpreted the icon as Shiva’s anandatandava or cosmic dance of creation and destruction, which is also the dance of bliss after annihilating the ego. In this form, the four-armed Nataraja exhibits five primordial acts or panchakritya: creation symbolized by the drum in the rear right-hand, protection by the front right arm, dispelling of ignorance and ego by trampling the demon muyalagan (Tamil) or apasmara (Sanskrit) with his right foot, granting of solace by the crossed left arm, destruction by the fire in the rear left arm, while the ring of fire symbolizes perpetual cosmic cycles (Nagaswamy 1983: 62). Shivakami as consort inspires this cosmic activity.

Coomaraswamy’s philosophical musings on the ‘cosmic dance of Shiva’ had wide appeal, even providing a post-modernist metaphor for savouring the implications of modern physics with Fritjof Capra (1976: 258) writing that ‘Shiva’s dance is the dance of sub-atomic particles’. However, several early Tamil hymns depict Shiva with a touch of morbidity as the lord of destruction, wandering and dancing at cremation grounds, albeit with a starkly moving appeal. In a thought-provoking paper, Kaimal (1999) countered that there is a disconnection between Coomaraswamy’s metaphysical interpretations based on thirteenth-century texts and Nataraja’s original significance for the Cholas. She argued that the icon was propagated for its martial implications, traceable to the worship of the Tamil goddess of war, Kottravai, with an ecstatic victory dance with lifted leg.

The third question is whether the Nataraja icon was originally significantly understood in terms of a cosmic or transcendental dimension. In the hymns of Saivite Nayanmar saints of the seventh – eighth centuries, Chidambaram was known as Tillai with a temple to the dancing Shiva known as Cirrambalam or Little Hall (Younger 1995: 14, 84). Younger (ibid.) suggests that it was the later Sanskritization of this word to ‘Chidambaram’ (chit: consciousness; ambaram: cosmos) which led to Nataraja’s identification as the cosmic dancer/consciousness, symbolically depicted as the akasa lingam or space/cosmic pillar.

However, the studies cited above do not significantly take into account pre-Chola Tamil evidence, which forms the focus of this paper. My purpose here is to show that the Nataraja bronze emerged under the Pallavas (c. seventh – mid-ninth century) from

iconographic evidence and archaeometallurgical studies on 130 south Indian metal icons (Srinivasan 1996, 1999, 2001). On the basis of the writings of Appar, Manikkavachakar and poetess Karaikkal Ammaiyar (sixth – ninth century), it also argues that a metaphorical interpretation of the icon, linked to the cosmic dance and the consciousness, was already in vogue prior to a more beatific phase of thirteenth-century worship to which it is usually attributed.

Early material evidence in relation to Saivism

Aspects that later came to be associated with Saivite worship are among the most archaic of South Asian cultic expressions. In the context of Harappan sites (c. 2600 – 1800 BC), for instance, several objects resembling ‘lingam’ motifs have been found such as a 4-foot-high, broken, bulbous-ended column from Dholavira, while a seated trident-headed ‘ithyphallic’ figure on Indus seals has been compared to Shiva as meditating ascetic (Kenoyer 1998: 112; O’Flaherty 1997: 8 ; Settar 1973: 38). However, in peninsular India, the earliest known overtly ‘Hindu’ stone icon may be a vital 5-foot-high Satavahana depiction from Gudimallam in Andhra Pradesh which dates to c. the second – first century BC. It is of a standing male within a bulbous-ended linga and strikingly exemplifies the merging of aniconic and anthropomorphic manifestations of Shiva. His club identifies him as Lakulisa, Shiva as ascetic, seen in later (sixth – ninth-century) peninsular Indian sculpture, whose ithyphallic aspect connotes asceticism and conserved procreative potentialities, rather than mere eroticism (O’Flaherty 1973: 80, 8). The grimacing dwarf demon under Shiva’s foot too is a precursor to seventh-century Pallava sculpture.

Other early icons of Shiva come from Bactrian and Kushan coinage (c. first century BC– third century AD) from Afghanistan and Pakistan. A copper seal from Sirkap has a legend Sivarakshita in Brahmi and Kharoshti beside a figure with club and trident (Sivaramamurti 1974: 168). A gold double stater of Kushan King Kanishka III, c. 240 – 50 BC, in the Fitzwilliam Museum, Cambridge (Errington and Cribb 1992: 66, 70) has an ithyphallic deity identified as Shiva beside a bull, with trident, water pot and lion skin (of Greek god Heracles), inscribed ‘Wesho’ in Bactrian after the Zoroastrian wind god.

While some early copper alloy figurines are known from the Tamil region, such as a tiny ‘mother goddess’ from Adichanallur, c. 800 BC, and a Buddha from Kaveripattinam, c. first – third century AD, it was the ascendancy of Puranic Hindu worship over the heterodox faiths of Buddhism and Jainism by the latter first millennium which resulted in a dramatic spurt in casting solid religious icons.

Exploring the presence of Nataraja in the Pallava period

On Pallava stone sculpture relating to Nataraja

The Pallavas, originally governors of Andhra dynasties, ruled in the Tamil region from about AD 550 until their overthrow by Aditya Chola around AD 890. Whereas a few earlier

Jain caves are known, it was with Mahendravarman Pallavan (who died in AD 630) switching from the Jain to Saivite faith that Hindu iconography in stone came to prominence. At Mamallapuram/Mahabalipuram, free-standing monolithic stone shrines were sculpted whose novelty is indicated by the Mandagapattu inscription praising Mahendravarman for his conceptions of ‘neither brick, timber nor mortar’ (Srinivasan 1975).

The Nataraja bronze has no exact iconographic precedents outside the Tamil region in stone or metal. Early stone sculptures elsewhere rather depict Shiva dancing in the chatura tandava or rhombus pose without the dwarf demon underfoot, such as an eight-armed fifth-century Gupta sculpture, from Sirpur in central India, and a dramatic sixteen-armed sixth-century Chalukyan sculpture from Badami in south-western India (Sivaramamurti 1974: 172).

In fact, the earliest stone approximation to the Nataraja bronze is found in a seventh century Pallava cave temple at Siyamangalam (Plate 1) (ibid.: 193). Here, a four-armed Shiva dances in bhujangatrasita karana with right leg extended, front right hand in the protective gesture and rear right hand holding a lamp with fire; however, the front left hand is not crossed and the demon is missing. Executed on a pilaster, it suggests a creative portrayal inspired by a dancer rather than a worshipped icon. There are few iconographic parallels elsewhere save for a fine eighth-century Rashtrakuta cave painting from Ellora, Maharashtra (Vatsyayan 1982: 39). Additionally, Nagaswamy (1995: 113) claims to have identified a Pallava stone Nataraja in the Tirukkadaimudi Mahadeva temple, Tirucchi-nampundi.

On processional icons and Nataraja in wood or metal: Appar’s seventh-century verses

The empty shrine cells in many of Mahendravarman’s cave temples (early seventh century) indicate that images worshipped inside the sanctum were not yet of stone. Finds of a sunken chase in the wall suggests that the deity was a carved wooden panel, while traces of lime plaster indicate the worship of stucco figures (Dayalan 1992: 19).

As such, Pallava epigraphic evidence about Chidambaram, or for Nataraja bronzes, does not seem to have surfaced, although a foundation inscription on the Kailasanatha temple in Kanchipuram confirms that Narasimhavarman Pallavan II (AD 695 – 728) was a follower of Saiva Siddhanta agamas (Davis 2000: 12). However, Tamil hymns of Appar (seventh century) and Sambandar (eighth century) clearly mention the ‘cittar’, or ascetic followers of Saiva Siddhanta, living in Tillai where ‘the Lord dances in the Little Hall (Cirrambalam)’ (Younger 1995: 14). The earliest structure at Chidambaram, now called the Chit-Sabha, is a small rectangular wooden building with a thatch-styled roof (said to have been gilded to mark Aditya Chola’s overthrow of the Pallavas) which resembles the seventh-century stone Pallava Draupadi Ratha at Mamallapuram (ibid.: 84). Thus it seems likely that the earliest icons of Nataraja worshipped at Chidambaram in Pallava times also were of wood.

Appar’s seventh-century verses also mention an image of dancing Shiva taken in procession during the festival of Ashtami and another clad in ceremonial dress (Dehejia 2003: 103). Dehejia (ibid.) surmises that practices of taking metal images out in procession,

elaborately clad in silks, jewellery and garlands, originally derived from e orts to enhance duller wooden icons.

Intriguingly, Appar also mentions Shiva’s ‘sweet golden foot raised in dance’ (Prentiss 2003: 68). Perhaps then, the worship of a metal Nataraja icon in his time cannot be ruled out. That the image worshipped at Tillai by Appar approximated to the Nataraja bronze is suggested by the verse below (Younger 1995: 203):

He is well-versed in the Secret Knowledge: he holds a drum in his hand; The moon shines over his head;

and he is the Lord who removes di culties of those who worship him. In the Little Hall at Tillai he dances with fire

And makes noise with his tinkling anklets.

Metallurgical investigations on south Indian metal icons

Although there is no absolute method of dating solid metal artefacts, lead isotope ratio analysis can assist in exploring typological similarities on the basis of shared metal sources because these remain unchanged from ore to extracted metal, and are found to be distinctive for di erent ore sources due to geological factors related to the age of the

deposit and original concentration of uranium and thorium which decides the proportion of daughter isotopes of Pb206, Pb207 and Pb208 in relation to primeval lead, Pb204 (Gale and
Stos-Gale 1982).

In an archaeometallurgical finger-printing exercise undertaken by the author, 130 representative south Indian metal icons sampled by micro-drilling from the Government Museum, Chennai, Victoria and Albert Museum, London, and British Museum, London, were analysed for eighteen elements, of which sixty were analysed for lead isotope ratios (Srinivasan 1996, 1999, 2001). Whereas lead, tin and zinc were found to be intentionally alloyed at random, systematic di erences were found related to dynastic chronology in the trace elements (i.e. 51 per cent), especially the chalcolithic ones of nickel, cobalt, arsenic, bismuth and antimony, perhaps due to di erences in sources or metal processing. The lead isotope ratios and trace elements of a control group of images of well-dated or inscribed artefacts (including some coins and charters) were thus calibrated against stylistic or inscriptional criteria, yielding characteristic finger-prints for stylistic groups of images.

Against these trends, the images including those of uncertain attributions, were re-classified as Pre-Pallava (c. AD 200 – 600) [8], Pallava (c. AD 600 – 875) [17], Vijayalaya

Chola  (c.  AD  850 – 1070)
[31],  Early
Chalukya-Chola  (c.  AD  1070 – 1125)  [12],
Chalukya-Chola   (c.   AD
1125 – 1279)
[17],   Later   Pandya   (c.   AD   1279 – 1336)

Vijayanagara and Early Nayaka (c. AD 1336 – 1565) [20] or Later Nayaka and Maratha (c. AD 1565 – 1800) [12]. Table 1 gives the salient lead isotope ratio trends of Pb 207/206, Pb 208/206, Pb 206/204 for major stylistic groups of images analysed by thermal ionization mass spectrometry at ISOTRACE laboratory, Oxford. Figure 1 gives the lead isotope ratio plot of Pb 208/Pb 206 vs. Pb 207/206 for most sampled south Indian images. The ellipses circumscribe those artefacts with a spread of +0.3 per cent, suggesting that they represent lead from discrete ore sources or zones within a larger mining region. Isolated ellipses with less than three artefacts might represent those alloyed with the same

Table 1    Lead isotope ratio trends for groups of South Indian images

Vijayalaya Chola (i.e. Imperial Chola) (c. AD  850 – 1070)

Ellipse 1a
17 images

Mean Pb 207/206: 0.84505;
Spread: +0.26%

Mean Pb 208/206: 2.08923;
Spread: +0.2%

Mean Pb 206/204: 18.6003;
Spread: +0.35%

Vijayananagara (c. AD  1336 – 1565)

Ellipse 2a
11 images

Mean Pb 207/206: 0.85537;
Spread: +0.1%

Mean Pb 208/206: 2.11668;
Spread: +0.18%

Mean Pb 206/204: 18.3656;
Spread: +0.19%

Chalukya-Chola (i.e. Late Chola) (c. AD  1070 – 1279)

74 images

Mean Pb 207/206: 0.85385;
Spread: +0.09%

Mean Pb 208/206: 2.10344;
Spread: +0.21%

Mean Pb 206/204: 18.3743;
Spread: +0.18%

Pallava (c. AD 600 – 875)

Line 1
6 images

Lead isotope ratios fall on or close to line of equation

y = mx + c; m (slope) = 1.3042; (constant) = 0.98336

* Analysed by thermal ionization mass spectrometry at ISOTRACE laboratory, Oxford.

batch of lead mixed from more than one source. Significantly, these ellipses also show internal stylistic consistencies. Table 2 gives the averages of major and selected trace elements for groups of images, analysed by inductively coupled plasma atomic emission at Royal Holloway and Bedford New College, Egham.

On archaeometallurgical evidence for a Pallava school of bronzes (c. AD  600 – 850)

As such, Hindu bronzes have seldom been attributed to the Pallavas due to the lack of inscribed images (Balasubrahmanyam 1971: 279; Barrett 1965: 1 – 5; Nagaswamy 1995). However, the archaeometallurgical investigations suggest that a shift from wooden or terracotta processional images to metal icons was made within the Pallava period itself. Figure 1 indicates that lead isotope ratios for Chola bronzes (c. AD 850 – 1070) cluster into the ellipse 1a because their lead came from a discrete source. On the other hand, lead isotope ratios for Pallava artefacts (c. AD 600 – 850) follow a reasonably discrete linear trend lying close to the Line 1, which can be explained if their lead was randomly mixed from two other sources. The trace element profile of Pallava bronzes could also be distinguished from Chola, with the average Ni/Co ratios of 18.88 and 3.75 respectively and Fe/S ratios of 16.09 and 5.21 respectively.

Also, a copper-plate grant of the Pallava ruler Paramesvaravarman I (c. AD 630 – 68) from a temple in Kuram clearly mentions metal images (Srinivasan 1963: 68 – 70). Indeed,

Figure 1 Lead isotope ratio plot of Pb 208/Pb 206 vs. Pb 207/206 for sampled south Indian bronzes with markers indicating dynastic chronology. The ellipses indicate groups of objects with similar sources of lead; the ellipse 1a consists mostly of early and high Chola bronzes (late ninth to mid-eleventh century). The objects lying close to the line would have lead mixed from two sources; these are found to be Pallava artefacts.

a bronze of Shiva dancing in urdhvajanu pose, of Natesa (not Nataraja), from Kuram (Plate 2), in Government Museum, Chennai/Madras, followed the ‘Pallava’ metallurgical profile, making it datable to Paramesvaravarman I. Although Barrett (1965: 40) argued it was Chola, Nagaswamy (1988: 142) and Rajeshwari (1988: 20) supported a Pallava attribution. Its forward facing demon is found in Pallava stone sculpture, rather than Chola, such as a Shiva in urdhvajanu pose in the seventh-century Dharmaraja ratha at Mahabalipuram.

Archaeometallurgical evidence for Pallava Nataraja bronzes (c. AD  800 – 50)

For two fine Nataraja bronzes (with leg extended in bhujangatrasita karana), which were previously classified as Chola, the lead isotope ratios rather followed Pallava trends. One is a charming Nataraja image from Kunniyur in Government Museum, Chenai (Plate 3) dated as tenth-century Chola by Nagaswamy (1983: 71). However, on re-examination, it resembles the Tripurantaka from Ponvilaintanpatti attributed to the Later Pallava period by Nagaswamy (1983: 80), with both standing on a crouching sideways facing dwarf. Chola Natarajas di er in usually having flying locks, with the dwarf’s face and body twisted more upwards. Hence, a later Pallava attribution prior to AD 850 seems more reasonable.

The other, a petit Nataraja from the British Museum (Plate 4) shows features found in Pallava stone sculpture and in the seventh-century Natesa bronze from Kuram (Plate 2) of

Table 2    For South Indian icons, average weight percentage of major and selected trace elements

Stylistic groups

Pre – Pallava and Andhra

(c. AD  200 – 600)

Pallava (c. AD  600 – 875)
Early Vijayalaya Chola

(c. AD  850 – 940)

Later Vijayalaya Chola

(c. AD  940 – 1070)

Early Chalukya – Chola

(c. AD  1070 – 1125)

Later Chalukya – Chola

(c. AD  1125 – 1279)

Later Pandya

(c. AD  1279 – 1336)

Vijayanagara & Early Nayaka

(c. 1336 – 1565)

Later Nayaka, Maratha

(c. 1565 – 1800)


Analysed by inductively coupled plasma atomic emission spectrometry at Royal Holloway and Bedford New College, Egham.

‘Vijayalaya Chola’ is synonymous with Imperial Chola, after the first ruler of the dynasty while ‘Chalukya-Chola’ is better known as ‘Late Cholas'

Shiva as‘cosmicdancer’

Plate 1 Early Pallava stone sculpture of Nataraja, c, seventh century, Siyamangalam (photograph credits: French Institute of Pondicherry and Ecole Franc¸aise d’Extreˆme Orient, Pondicherry).

the looped waist robe, tassels and forward facing dwarf. Although this suggests an early date, the fact that this image had lead isotope ratios very similar to the Kunniyur Nataraja, both falling in the ellipse 6a close to the Line 1, suggests that its event of manufacture may be linked closer to this latter image. Thus, an intermediate date of about AD 800 seems reasonable, making it the earliest known Nataraja bronze of Pallava a liation.

Postulated iconographic transformations between wood, metal and stone

That these two ‘Pallava’ Nataraja images evolved from wooden prototypes is suggested by their resemblance to modern wooden Natarajas from south India in their compactness, reflecting the low tensile strength of wood, exemplified in the close-set limbs with the sash

Plate 2 Natesa (acc. no. 53/38), Kuram, attributed to c. AD 650 by the author, Government Museum, Chennai (photograph credit: Government Museum, Chennai).

hanging down, in one case enclosed in an elliptical aureole or prabha. It seems that only with an appreciation of the greater tensile strength of metal compared to wood were the limbs, locks and sash flared out more later in tenth-century Chola Natarajas towards a 2 circular shape. This is seen in a Nataraja (Plate 5) from Kankoduvanithavam in the Government Museum, Chennai, analysed by the author, for which metallurgical trends corroborate a Chola date (c. AD 1040).

The poor tensile strength of stone compared to metal could explain why three-dimensional Nataraja stone sculptures emerged later than metal icons. Kaimal (1999) points to a tiny shallow relief of a Nataraja on an arch, c. AD 920 in the Sadaiyur temple in Tirucchenampundi, circumscribed by a circle, which looks like it was copied from bronze. Well-rounded Nataraja sculptures first emerged under Chola queen Sembiyan Mahadevi; however, even these demonstrate the problems of essaying the

Plate 3 Nataraja, Kunniyur, Tanjavur district (acc. no. 752/75), attributed to Pallava period, c. AD 850 by author, Government Museum, Chennai (photograph credit: Government Museum, Chennai).

lifted leg of the icon in stone, such as a fine Nataraja (Plate 6) in the Manavalesvarar temple, Tiruvelvikudi (c. AD 949 – 57). Here, a strut joins the crossed left hand to lifted foot, albeit aesthetically disguised as clothing. Likewise, the lifted leg of a stone Nataraja on the eleventh-century Chola temple of Gangaikondachalapuram is propped by a rough basal strut; in a few other examples the lifted leg is broken. Such stone struts may have also been inspired by runners for facilitating metal casting which were broken o to finish icons.

Indirect evidence of Pallava enthusiasm in depicting Shiva’s dance and of transforma-tions from wood to stone may come from far-o Prambanan in Indonesia. Here, at the Shiva temple dated to the mid-ninth century AD, sixty-two karanas or dance movements of Shiva are sculpted in stone (Iyer 1997), predating the well-known 108 karanas series sculpted inside the late tenth – early eleventh-century Brhadisvara temple of Rajaraja

Plate 4 Nataraja, attributed to Pallava period, c. AD 800 by author, British Museum (acc. no. OA – 1969 – 12 – 16 – 1) (photograph credit: Trustees of the British Museum).

Chola in Tanjavur and on the twelfth – thirteenth-century Chidambaram temple towers. However, the possibility of such sculpture being of Pallava antiquity is suggested by the karanas sculpted in a little known temple of Dantivarman Pallava (c. AD 800 – 50) at Bahur, Tamil Nadu (Sivaramamurti 1974: 348). Indeed, the triple-shrined model of Prambanan dedicated to Brahma, Shiva (central shrine) and Vishnu (ibid.) follows the Pallava Trimurti cave temples at Mamallapuram. Pallava-inspired Hindu bronzes are found in Indonesia (Scheurleer and Klokke 1988: 26 – 7) while Saiva Siddhanta philosophy had spread to Southeast Asia (Davis 2000: 12). One may even speculate that a wooden Pallava karana series from Chidambaram was copied elsewhere. Legend has it that the stone Nritta Sabha or Dance Hall at Chidambaram, built on a wooden prototype, replaced a shrine to the goddess whom Shiva defeated in a dance contest and delighted in watching dance from his opposite shrine (Younger 1995: 90).

Plate 5 Nataraja (acc. no. 238), Kankoduvanithavam, c. AD 1040, Government Museum, Chennai (photograph credit: Sharada Srinivasan).

Understanding the symbolism of Nataraja in Pallava times

Poetic allegories for Nataraja’s cosmic dance and destruction of the ego

That Nataraja was worshipped at Chidambaram by the Pallava period with underlying philosophical concepts of cosmic cycles of creation and destruction is found in Tamil saint Manikkavachakar’s Tiruvachakam: ‘ let us praise the Dancer (kuttan) who in good Tillai’s hall dances with fire, who sports (vilaiyatu), creating, destroying, this heaven and earth and all else’ (Dehejia 2003: 103). Manikkavachakar, who gave up a government job to become a devotee of Nataraja at Chidambaram, is dated by many scholars to the ninth century, although tradition dates him to the fifth century (Younger 1995: 30).

Another of his verses (Mowry 1983: 53) indicates how the lingam stood for an abstraction of the infinite and provided a humbling experience, predating thirteenth-century texts describing Shiva as destroyer of the ego.

Plate 6 Nataraja, Manavaleshvarar temple, Tiruvelvikudi, c. AD 949 – 57, in style of Chola queen Sembiyan Mahadevi (photograph credits: French Institute of Pondicherry and Ecole Franc¸aise d’Extreˆme Orient, Pondicherry).

What is the meaning of your Lord, standing as a pillar of fire,

reaching from the nether regions to the sky, that Brahma and Vishnu may not know Him? reaching from the nether regions to the skies, had he not stood so,

both of them on account of their altercation between them, would never have shed their egotism. . ..

He who creates, protects,

and destroys the verdant world, The primeval One. . .

The c. sixth-century AD Tamil poetess and saint Karaikkal Ammaiyar, who is said to have prayed to Shiva to be turned from a beautiful woman into a hag (Prentiss 2003: 28), also provides a striking allegory for release from vanity and egotism.

Shiva as ‘akasa’: cosmic connections in the time of the Pallavas

A Pallava sculpture reported by Rajeshwari (1988: 43) depicts Shiva as ashtamurti, his eight dancing manifestations of immanence including the element akasa. A seventh-century verse by Appar (Younger 1995: 203) describes Nataraja at Chidambaram as the ‘holder of Secret Knowledge’, echoing the Chidambaram Rahasya, or ‘revelation of Chidambaram’ representing Shiva as the akasa lingam or space. There have been conjectures about the astro-archaeological underpinnings of some Pallava sculpture at the Kailasanatha temple and of Nataraja’s iconographic links to the constellation, Orion (Srinivasan 2003). The festival of Margayi Tiruvadirai held at the winter solstice, when Nataraja is taken out in procession on a chariot from the Chidambaram temple, is related to ardra, the star Betelgeuse in Orion.

The following hymn of Manikkavachakar testifies that the Chidambaram Nataraja had, by the pre-Chola period, an abstract or ‘cosmic’ symbolism linked to the five elements including ether (Yocum 1983: 20). It also expresses his longing for mystical union with the transcendental, akin to the lover-beloved model of Bhakti worship:

Ah, When will I get to gaze upon the unique One to whom no other compares

Him who is fire, water, wind, earth and ether, Him whom others cannot understand. . .

With voice stammering,

a cataract of tears gushing forth, hands joined in worship

When will I adorn Him with fragrant flowers? When will I be united with my uncut Gem

The dance of bliss, dance of the lord and devotees

The Pallava Nataraja icons and stone and metal discussed (especially Plate 1 and Plate 3), suggest the notion of Nataraja’s blissful dance with an unmistakable attitude or smile of puckish delight that contrasts with the gravitas of Shiva dancing as Natesa (in chatura pose) from Badami or Ellora (Sivaramamurti 1974: figs 12, 17). This could suggest that the anandatandava or dance of bliss represented not just a di erent mode of dance but a philosophy unique to the Pallava domain. Zvelebil (1985) and Yocum (1983: 31) respectively speculated that the Nataraja cult and Manikkavachakar’s devotional experience as hysterical bliss drew from older practices reflected in Tamil Sangam literature (c. fifth century BC to fifth century AD), such as the ecstatic kavadi dances and possession (veriyatal) linked to the cult of Murugan/Velan, and the Tamil folk deity with the trident or vel. Rajaraja Chola’s Brhadisvara temple (c. AD 985 – 1014) had hundreds of dancing girls or devadasis attached to it with the 108 sculpted karana reliefs indicating their

role in invoking Shiva’s anandatandava. Interestingly, an earlier verse by Manikkavacha-kar (Yocum 1983: 30) reveals that, even for the Saiva Siddhanta ascetic and devotee, such bliss could be experienced by breaking into dance:

He. . .revealed His foot which is like a tender flower, caused me to dance

entered my innermost part (akam) became my Lord.

‘Or unarve’, the one consciousness: views of the mystic and the monist

The thirteenth-century Saiva Siddhantic text, Kunchitangrim Bhaje by Umapati of Chidambaram, describes Nataraja as sacchidananda or ‘Being, Consciousness and Bliss’ (Smith 1998: 21). This approaches the doctrine of Advaita, or abstract monism, expounded by celebrated south Indian philosopher-saint Sankaracharya, which holds the individual (jivatman) and supreme soul/consciousness (paramatman) to be one. Although Sankara, who lived around AD 788 – 820, set up a monastery at the Pallava bastion of Kanchipuram and knew of Saiva Siddhanta worshippers of Chidambaram, there is no evidence for his influence at Chidambaram before the thirteenth – fourteenth century (Davis 2000: 14; Younger 1995: 112, 223). However, an earlier hymn to Nataraja by Manikkavachakar (Yocum 1983: 24), indicated below, clearly identifies him with the unitary supreme consciousness, by using the Tamil word ‘Or unarve’ rather than the Sanskrit ‘chit’.

O unique consciousness (or unarve),

which is realised (unarvatu) as standing firm,

transcending words and (ordinary) consciousness (unarvu), O let me know a way to tell of You. (22: 3)

It is thus tempting to speculate about an osmosis of ideas, in whichever direction, between the mystic Manikkavachakar and monist Sankara or their followers.


Processional icons of dancing Shiva and the Nataraja worshipped at Chidambaram until Mahendravarman Pallavan’s time (c. AD 600 – 30) were probably of wood. A switch to metal processional icons was made by at least Paramesvaravarman I’s time (c. AD 650) from the Kuram copper plate and dates from archaeometallurgical finger-printing for the Shiva Natesa from Kuram (Plate 2), while Appar’s seventh-century verses leave open the possibility of the worship of a Shiva Nataraja bronze at Chidambaram. Finger-printing supports a Pallava attribution, c. AD 800 – 50, for two Nataraja bronzes (Plates 3 and 4), fitting dates for Dantivarman Pallava under whom karanas of Shiva’s anandatandava were sculpted at Bahur. This coincides not only with dates for Manikkavachakar’s mystical hymns, evoking Nataraja as the elements and the one consciousness (Tamil: ‘or unarve’),

but also for Sankara, c. AD 800, whose monistic doctrines might also have had a bearing. Setting aside the odd flattish stone frieze, sizeable well-rounded stone Natarajas emerged later than the metal icon, and specifically under the reign of Chola queen Sembiyan Mahadevi (c. AD 950).


The author would like to acknowledge the generous guidance of the late Dr Nigel Seeley as past MA tutor in the archaeometallurgy of South Asia at University College London Institute of Archaeology.

National Institute of Advanced Studies, Indian Institute of Science Campus, Bangalore


Balasubrahmanyam, S. R. 1971. Early Chola temples Parantaka I to Rajaraja I (A.D. 907 – 985). New Delhi: Orient Longman.

Barrett, D. 1965. Early Chola Bronzes. Bombay: Bhulabhai Memorial Institute, Motilal Banarsidass and Indira Gandhi National Centre for the Arts.

Capra, F. 1976. The Tao of Physics: An Exploration of the Parallels between Modern Physics and Eastern Mysticism. London: Fontana.

Clothey, F. and Long, B. (eds) 1983. Experiencing Siva: Encounters with a Hindu Deity. New Delhi: Manohar.

Coomaraswamy, A. K. 1924. The Dance of Siva: Fourteen Indian Essays. New York: Sunwise Turn.

Davis, R. 2000. Worshipping Siva in Medieval India: Ritual in an Oscillating Universe. New Delhi: Motilal Banarsidass.

Dayalan, D. 1992. Early Temples of Tamilnadu: Their Role in Socio-Economic Life (c. A.D. 550 – 925). New Delhi: Harman.

Dehejia, V. 1990. Art of the Imperial Cholas. New York: Columbia University Press.

Dehejia, V. 2003. Shiva – his temples, his images. In The Sensuous and the Sacred: Chola Bronzes from Southern India (ed. V. Dehejia). New York: American Federation of Arts, pp. 90 – 150.

Errington, E. and Cribb, J. (eds) 1992. The Crossroads of Asia: Transformation in Image and Symbol. Cambridge: The Ancient India and Iran Trust.

 Gale, N. H. and Stos-Gale, Z. A. 1982. Bronze age copper sources in the Mediterranean: a new approach. Science, 216(4541): 2 – 19.

Iyer, A. 1997. Prambanan: Sculpture and Dance in Ancient Java. Bangkok: White Lotus.

Kaimal, P. 1999. Shiva Nataraja: shifting meanings of an icon. Art Bulletin, 81(3): 390 – 420.

Kenoyer, M. 1998. Ancient Cities of the Indus Valley Civilization. Karachi: Oxford University Press.

Marr, J. R. 1985. The Eight Anthologies: A Study in Early Tamil Literature. Madras: Institute of
 Asian Studies.

Mowry,  L. 1983. The  theory of the  phenomenal world in Manikkavachakar’s Tiruvachakam.  In

Experiencing Siva: Encounters with a Hindu Deity (eds F. Clothey and B. Long). New Delhi: Manohar, pp. 37 – 59.

Nagaswamy, R. 1983. Masterpieces of Early South Indian Bronzes. New Delhi: National Museum.

Nagaswamy, R. 1988. South Indian bronzes. In The Great Tradition: Indian Bronze Masterpieces (ed. K. Khandalavala). New Delhi: Festival of India, pp. 142 – 79.

Nagaswamy, R. 1995. On dating south Indian bronzes. In Indian Art and Connoisseurship: Essays in Honour of Douglas Barrett (ed. J. Guy). Ahmedabad: Indira Gandhi National Centre for the Arts in association with Mapin, pp. 101 – 29.

O’Flaherty, W. 1973. Siva: The Erotic Ascetic. Oxford: Oxford University Press.

Parker, S. 1992. The matter of value inside and out: aesthetic categories in contemporary Hindu temple arts. Ars Orientalis, 22: 97 – 109.

Prentiss, K. 2003. Joyous encounters: Tamil bhakti poets and images of the divine. In The Sensuous and the Sacred: Chola Bronzes from Southern India (ed. V. Dehejia). New York: American Federation of Arts, pp. 64 – 80.

Rajeshwari, D. R. 1988. The Pallava Sculpture. New Delhi: Intellectual Publishing.

Ramanujan, A. K. 1985. Poems of Love and War. New York: Columbia University Press.

Scheurleer, P. A. and Klokke, M. J. 1988. Divine Bronze: Ancient Indonesian Bronzes from A.D. 600 to 1600. Leiden: E. J. Brill.

Settar, S. 1973. World History. Madras: Macmillan India.

Sivaramamurti, C. 1974. Nataraja in Art, Thought and Literature. New Delhi: National Museum.

Smith, D. 1998. The Dance of Siva: Religion, Art and Poetry in South India. Cambridge: Cambridge University Press.

Srinivasan, K. R. 1975. Temples of the later Pallavas. In Studies in Indian Temple Architecture (ed. P. Chandra). New Delhi: American Institute of Asian Studies, pp. 197 – 239.

Srinivasan, P. R. 1963. Bronzes of South India. Bulletin of the Madras Government Museum. New Series, Vol. 8. Madras: Government Museum.

Srinivasan, S. 1996. The enigma of the dancing ‘pancha-loha’ (five-metalled) icons: archaeome-tallurgical and art historical investigations of south Indian bronzes. PhD thesis. Institute of Archaeology, University of London.

Srinivasan, S. 1999. Lead isotope and trace element analysis in the study of over a hundred south Indian metal icons. Archaeometry, 41(1): 91 – 116.

Srinivasan, S. 2001. Dating the Nataraja dance icon: technical insights. Marg: A Magazine of the Arts, 52(4): 54 – 69.

Srinivasan, S. 2003. Heritage: the Nataraja catapulted onto the global stage from sacred environs. The Week, 21(29): 60 – 2.

Vatsyayan, K. 1982. Dance in Indian Painting. New Delhi: Abhinav.

Yocum, G. 1983. ‘Madness’ and devotion in Manikkavachakar’s Tiruvacakam. In Experiencing Siva: Encounters with a Hindu Deity (eds F. Clothey and B. Long). New Delhi: Manohar, pp. 19 – 37.
Younger,  P.  1995.  The  Home  of  Dancing  Sivan:  Traditions  of  the  Hindu  Temple  in  Citamparam.

Oxford: Oxford University Press.

Zvelebil, K. 1985. Ananda-Tandava of Siva-Sadanrttamurti. Madras: Institute of Asian Studies.

Sharada Srinivasan did a B.Tech in engineering physics from the Indian Institute of Technology, Mumbai (1987), followed by a Master’s in art and archaeology from School of Oriental and African Studies, London (1989). She got her PhD degree in 1996 from the Institute of Archaeology, University College London, in the archaeometallurgy of South Indian metal icons. She is currently Fellow at the National Institute of Advanced Studies, Bangalore, India. Her research interests are in archaeometallurgy, archaeomaterials, art history and also, as an exponent of the classical Indian dance form of Bharata Natyam, Indian dance traditions.


QUERIES: to be answered by A

No comments: