Current local: April 18, 2014, 3:35 p.m.

Madurai and Chidambaram:

The Tamil Cities that Created Important Sanskrit Myths

by S. Palaniappan

May 1999


In his paper, "The Changing Notion of ziSTa from pataJjali to bhartRhari", (in "Bhartrhari: Philosopher and Grammarian" edited by Saroja Bhate and Johannes Bronkhorst, first published in 1993) M. M. Deshpande writes,

K. V. Abhyankar (1954:352) suggests that the southern grammarians knew the account of the transmission of the mahAbhASya and its recovery by candrAcArya from the south, and then they created the elaborate myths about pataJjali appearing in Chidambaram. However, this does not explain why pataJjali should have been regarded as an incarnation of zeSa, rather than some other divinity." (p.113 n.7)

In his later paper, "Who Inspired pANini? Reconstructing the Hindu and Buddhist Counter-Claims", JAOS, 117.3, 1997, Deshpande writes,

It would thus seem that the notion of pANini having been inspired by ziva may have developed in certain zaivite communities around the middle of the first millennium A.D. It was originally, in all likelihood, independent of the Chidambaram mythology associating ziva naTarAja with pataJjali incarnated as AdizeSa. David Dean Shulman (1980: 122ff.) discusses extensively the complex relationships of AdizeSa with various different south Indian zaivite shrines, most of which have little to do with his character as the grammarian pataJjali. In fact, I have a suspicion that bhartRhari, calling pataJjali AdiziSTa, may have been later [mis]understood as referring to AdizeSa." (p. 453)

Deshpande also acknowledges the interaction between the Tamil and Kashmir zaiva traditions in the spread of these myths. However, he seems to prefer Kashmir as the source of these myths.

In the Indology query, initiated by M. M. Deshpande to get early textual references to identifying pataJjali with AdizeSa, P.Magnone wrote on 20 Mar 95:

> As for YS II,47, there indeed is a varia lectio ananta/aanantya; among
> the commentators reading ananta, Vaacaspatimizra, VijJaanabhikzu,
> Raghavaananda Sarasvatii, BhaavaagaNeza, NaagojiibhaTTa, Raamaananda and
> Sadaazivendra Sarasvatii refer this to ZeSa, but they are all later than
> KaiyaTa. ... they are all later than KaiyaTa except for Vaacaspati, of course.

Enrica Garzilli wrote,

Even though is likely that Patanjali of the YS is not the Patanjali of the Mahabhasya, (See e.g. A.B. Keith, Jacobi, onwards), the YS II, 47 explicitely mentions Ananta. It was interpreted by Vacaspatimisra (IX cent. A.D.) as Vasuki, teh mithological snake (brother of Sesa). (However Ananta is also the name of Sesa himself, the snake-god) . King Bhoja (IX cent.) read it as 'endeless', 'eternal'.

M. M. Deshpande concluded the discussion saying,

Thanks for the reference to Yoga-suutra II 47 which refers to Ananta. Though it is doubtful that the word refers to the snake divinity in the original suutra, one can see how a later commentator could take it as a proper name referring to the snake divinity. However, it is not clear how even this reference could have led to the identification of Patanjali as an incarnation of Shesha.

All these efforts to analyze the myths regarding pANini and pataJjali have not been fully successful. However, if one looks at the Tamil materials, literary and epigraphical, one can arrive at a better understanding of the history of these myths. There are different elements which have contributed to these myths. These include competition between zaivism and Jainism, competition between Pandyas of Madurai and the Pallavas/Cholas associated with Chidambaram, and competition between zaivism and vaiSNavism at Chidambaram, the telescoping of pataJjali of yogasUtra and pataJjali of mahAbhASya. But the fundamental factor seems to be the political and religious competition between Madurai and Chidambaram. This competition seems to be the motivating force behind the adoption of the myths linking pANini and pataJjali with ziva at Chidambaram. These myths were carried to the northern Sanskritic tradition through the medium of zaivism. In effect, the myth linking ziva and pANini was not independent of the Chidambaram mythology.

dakshiNAmUrti of Potiyil/Madurai:

As I have discussed in earlier postings, potiyil in the Pandyan kingdom (capital- Madurai) was the locus of the daKSiNAmUrti cult. As the dakSiNAmUrti cult established itself in Madurai at least by 5th century, it created myths of ziva as the teacher of Tamil grammar which was taught to agastya first. This helped to put the real grammar of tolkAppiyar (2nd century BC), probably a Jain, as a work based on the earlier one by agastya (akattiyam, a mythical text). Considering all the Tamil Brahmi epigraphic evidence attesting to the strong presence of Jainism around Madurai, one can understand the motivations of the zaivites.

Due to the efforts of the followers of dakSiNAmUrti, ziva becomes a member of the Tamil academy at Madurai and also authors the authoritative text dealing with the grammar of Tamil love poetry (as mentioned in a text believed to be from 8th century.). Since the CT poets considered the love poetry as equivalent to/essence of Tamil, ziva becomes the authority on Tamil. From this point onwards zaivism and Tamil are tightly linked.

Ziva also marries the Pandyan princess and inherits the Pandyan kingdom. (In short, he becomes a Pandya king. One should note that even in pre-zaivite days of Pandyan history, a Pandyan king was customarily called the "southern one" which happens to fit well with "dakSiNAmUrti". In fact, the CT text maturaikkAJci notes this equivalence. ziva comes to be praised as "ten2n2AtuTaiya civan2E pORRi" (praises to ziva of the southern country) by mANikkavAcakar of 9th century AD.

Dancing god of Chidambaram

Around the time Pandyas of Madurai recovered their kingdom from the Kalabhras, the Pallavas were the rulers of northern Tamil country. By eighth century AD, we have evidence that they began to patronize the dancing ziva cult at Chidambram. As Kamil Zvelebil notes in "Ananda-tANTava of ziva-sadAnRttamUrti", p. 25, cuntaramUrtti, says in tEvAram 90.4: "The god of puliyUr ciRRampalam will cause trouble to kings who refuse to pay tribute to the Pallavas who guard the world." Zvelebil remarks, "The dancing ziva of dabhra sabhA is thus connected with ruling dynasty as its patron and protector - a fact of no small importance." Cuntarar should be placed in the late 7th or early 8th century based on his mentioning a ruling Pallava king named cigKan2 (Skt. simha) which probably refers to Pallava Rajasimha aka Narasimhavarman II who ruled between 695 and 728 AD. (Zvelebil's dating of of cuntarar in the period 780-820 AD is not correct since there was no Pallava king named Simha after Rajasimha. (See A History of South India, by K. A. N. Sastri, 1987, p. 171. Another clue to the identity of the ruling monarch is given by the name of the chieftain, naracigka munaiyadaraiyan2, who brought up cuntarar. Medieval inscriptions reveal that often chieftains adopted the name of the ruling monarch as part of their own name. This suggests that if the chieftain's name followed this practice, the ruling monarch would have had the name narasimha, which in this case will match Narasimhavarman II.)

Zvelebil also says,

One or two elements of the iconography of dancing ziva may in fact go as far back as Harappan culture, although the full-fledged concept seems to be absent. The Ananda-tANDava dance of ziva seems to be, though, a specific South Indian (mainly Tamil?) development. Even before a more precise iconologic-iconographic image of ziva naTarAja developed, the concept of a dancing god must have obviously been inherent in the Indo-Dravidian religious milieu since very ancient, possibly pre-historic or proto-historic times. (p.72-73).

Based on Hermann Kulke's work on cidambaramAhAtmya, Zvelebil traces the naTarAja cult at Chidambaram to three important cults one of which was the cult of a dancing god.

Competition leading to myth-making

As Zvelebil notes (p.71), "The zeSa-pataJjali legend is the central part of the mAhAtmya of Chidambaram." But researchers have till now not answered the basic question as to why pataJjali was inducted to be a devotee of ziva. To get an answer, we have to look at the political history of the Tamil country from 6th to 10th century AD. The Pandyas ruled at Madurai in the south while first Pallavas and later Cholas ruled in the north. There was bitter competition between the Pandyas on the one hand and Pallava/Cholas on the other. As Cholas took political control of northern Tamil country away from the Pallavas, they also adopted the cult of naTarAja of Chidambaram as their own. Thus it will be in the interest of the Pallavas and Cholas to ensure that the ziva of Chidambaram could be no less than the ziva of Madurai. It was also in the interest of the priestly community (which was closely allied with the kings) to create the legends to satisfy the royal needs for prestige on par with the Pandyas.

This is the background against which one has to analyze the Chidambaram myths. In this context, if the creators of the myth wanted a grammarian, they could have chosen pANini instead of pataJjali. The fact that they chose pataJjali means that they were not looking for a grammarian, but rather an authority on yoga. This will make sense if one compares the dancing god of Chidambaram with dakSiNAmUrti of Madurai. dakSiNAmUrt is the ultimate teacher-yogi. By recruiting pataJjali as naTarAja's devotee, the dancing god becomes a master of yoga too.

If pataJjali, the yoga exponent, was to be made a devotee of ziva, where was the need to make him an incarnation of the snake, AdizeSa? As discussed by Paul Younger in his "The Home of the Dancing S'ivan, 1995, p. 172), Chidambaram was also a vaiSNavite center sung by the Azvars. By making pataJjali an incarnation of AdizeSa, ziva is made superior to viSNu.

Researchers have for the most part linked the pataJjali of Chidambaram with the author of mahAbhASya. But if one considers the pataJjali of yogasUtra, the association of snake with pataJjali will make sense because kuNDalini is thought of as a snake. Apart from these, there is an important etymological connection of AdizeSa with dakSiNAmUrti as well. zeSa means "remainder, that which remains or is left". The equivalent of Skt. zeSa is Tamil "eccam". The word "eccam" means remainder, children, legacy, bequest, heirs, etc. In Tamil, children and disciples are referred to by the same terms such as putalvar and maintar. Children are left behind by the parents. Disciples are left behind by teachers. In fact, if children are one's physiological heirs, disciples/students are one's spiritual/intellectual heirs. Given this Tamil viewpoint, adoption of AdizeSa as a devotee of naTarAja is a fitting counterpoint to dakSiNAmUrti, the teacher-yogi, who is often accompanied by disciples.

But dakSiNAmUrti was also a teacher of vedas and vedaGgas which includes grammar. According to the Madurai mythology, it was ziva-dakSiNAmUrti who first taught Tamil grammar to agastya. So the dancing god of Chidambaram has to be vested with grammatical expertise also. Unlike the Tamil grammarian, tolkAppiyar, pANini was an adherent of Vedic religion. So, there was no necessity of an intermediary between ziva and pANini. The myth of ziva teaching akSarasamAmnAya to pANini directly was created. That is why the words of Haradatta (9th century) from the Chola country of Tamilnadu, the first author in the pANinian tradition to mention the ziva-pANini connection, seem to indicate a local tradition as its source. (See Deshpande, 1997, p.449) The next author to mention this story, dharmakIrti, was also from the Chola country.

Thus the legend of the dancing god at Chidambaram who originally could not have had the yogic and grammatical wherewithal to compete against the teacher-yogi dakSiNAmUrti of Madurai was complemented by the addition of a grammarian and a yogi as disciples.

Tracing the history of the myths

Probably the earliest mention of pataJjali at Chidambaram is by tirumUlar who was at least a contemporary of cuntarar (late 7th or early 8th century) if not older. tirumUlar says,

nanti aruL peRRa nAtarai nATiTin2
nantikaL nAlvar civayOka mAmun2i
man2Ru tozuta patajncali viyAk ramar
en2Ru ivar en2n2OTu eNmarum AmE. (tir. 67)

In this poem, tirumUlar lists those who obtained ziva's grace. This includes the four traditionally understood to be sanaka, sanandana, sanAtana and sanatkumAra who were taught by ziva dakSiNAmUrti. In addition, he includes civayOkamAmun2I (zivayoga mahAmuni), pataJjali, and vyAghrapAda who worshipped in the hall, and himself. One can see in this poem evidence for an attempt to link the traditions of the dancing god with that of dakSiNAmUrti. tirumUlar was probably basing this poem on existing priestly tradition.

pataJjali and vyAghrapAda are also indicated in the following poem where they are referred to simply as "iruvar" (the two) who saw ziva's dance in Chidambaram.

iruvarum kANa ezil ampalattE
uruvOTu aruvOTu urupara rUpamAyt
tiru aruL cattikkuL cittan2 An2antan2
aruL uru Aka nin2Ru ATal uRRAn2 E. (tir. 2790)

Is there any basis for interpreting iruvar as pataJjali and vyAghrapAda? Consider the following poem from Tamil kantapurANam (14th century).

virikan2al vELvitan2n2il viyan2 talai arintu vITTip
peru varu tavattai ARRum pataJcali pulikkAl aNNal
iruvarum uNarvAl kANa ellaiyil aruLAl Ican2
tiru naTa iyaRkaik kATTum tillai mUtUraik kaNTAn2 (kan. 1642)

It is obvious that we are correct in interpreting "iruvar" as "pataJjali and vyAghrapAda".

tirumUlar also says in another verse:

egkum tirumEn2i egkum civa catti
egkum citamparam egkum tiru nATTam
egkum civamAy iruttalAl egku egkum
tagkum civan2 aruL tan2 viLaiyATTu atE. (tir. 2722)

Regarding this, Zvelebil says,

The tendency expressed in this stanza is, of course, in perfect agreement (hence, this stanza should not be regarded as an interpolation!) with the over-all views of tirumantiram which are 'zaktic' and so-to-say of the anti-establishment - in this case going against the claim of the priesthood of Chidambaram as being the only and exclusive place of ziva's divine macrocosmic/microcosmic dance." (p.50)

Thus by late 7th/early 8th century AD, the pataJjali is associated with Chidambaram and interestingly, as we see, it is attested first in the earliest extant work of tantra/yoga in Tamil. While pataJjali is mentioned by tirumUlar, the association of pataJjali with a snake is first found in the works of mANikkavAcakar of ninth century AD. In his work tiruvAcakam, he praises ziva as

pataJcalikku aruLiya parama nATaka... (tiruvAcakam 2. 138).

the supreme dancer who bestowed grace on pataJjali

In his tirukkOvaiyAr, he says,

nAkam toza ezil ampalam naNNi naTam navilvOn2 (tirukkOvaiyAr 171)

the one who dances in the beautiful hall (Chidambaram) with the snake worshipping

Thus we get attestation of a linkage between pataJjali and snake through mANikkavAcakar in the 9th century AD.

In kallATam (10th century AD), we find a very short summary of the pataJjali myth of citambaram as follows:

mattiyantaNan2 varal coli viTuppat
tillai kaNTa pulikkAl mun2ivan2um
cUyai kaiviTap pataJcali Akiya
Ayiram paNATavi aruntavattu oruvan2um
kaNNAl vAgki neJcaRai niRaippa
tirunaTam navin2Ra ulakuyirp perumAn2 (kallATam 42.4-9)

A rough translation is:

the Lord who is the life of the world who performed the holy dance with the sage with tiger-feet who saw Chidambaram because madhyandina asked him to go there, and the thousand-hooded one of rare austerities (AdizeSa) who became pataJjali because anasuyA dropped him, taking in the (view) through the eyes and filling their hearts...

Another text, tiruppallANTu, has the following lines,

pAlukkup pAlakan2 vENTi yazutiTap
pARkaTal IntapirAn2

A rough translation is:

the Lord who gave the ocean of milk when the baby cried for milk.

It is obvious that the story of upamanyu is indicated here.

Tamil to Sanskrit tradition

Regarding the Sanskrit version of the Chidambaram myth, cidambaramAhAtmya, Paul Younger writes that it

is clearly a pilgrim guidebook written for the use of North Indian pilgrims to Citamparam, in Sanskrit the language of all-Indian religious discourse. The text is not dated, but it reflects the situation of the twelfth and thirteenth centuries when large numbers of devout worshippers were fleeing to the famous temples of the South as the armies of the Delhi Sultanate swept across North India destroying many temples and using the stones of others to build mosques.

Based on the evidence presented here, it is obvious that while the Sanskrit text could have been composed in the 12th/13th century AD, but the tradition is at least 4/5 centuries older, if not more. The motivation of the tradition had nothing to do with North Indian pilgrims fleeing muslim invasion. It is an entirely local Tamil zaivite tradition of Chidambaram.

Can this history of the myth explain the varied and non-existent associations between zeSa and pataJjali in the Sanskrit texts? According to the Indology list discussion of March 20, 1995, only after Kaiyata (11th century), we get clear identification of pataJjali with zeSa. vAcaspatimizra associates pataJjali of yogasUtra with a snake but not zeSa. He probably got the association from his teacher trilocana of karnAta. (I thank Vidyasankar Sundaresan for information on vAcaspatinizra.) Probably in the transmission from Tamilnadu to Karnataka, some distortion crept into the information. However, there is a clear identification of pataJjali of yogasUtra with zESa before Kaiyata.

Abhinavagupta identifies pataJjali of yogasUtra with zeSa. In his book, "Abhinavagupta and His Works", V. Raghavan says on p.81,

On pp. 190 and 240 of the parAtrimzikA (Kashmir Series), there are two references to one and the same yogasUtra, namely III.17, on p. 190 as said by zeSamuni and on p. 240 as yogasUtra. In his tantrAloka (Vol. VIII., p. 95), in ch. 13, zl. 146, Abhinavagupta cites yogasUtra III.33 and mentions the author as zeSa mahAmuni: ...

This indicates two things. It confirms the position that originally the pataJjali of Chidambaram was indeed the author of yogasUtra as argued by me. Moreover, Abhinavagupta was very well familiar with the Tamil zaivite traditions of Chidambaram and has indeed accepted them and used them in his writings. Thus the impact of Tamil traditions on Kashmir zaivism needs to be revisited.

pANIni and Chidambaram

In his JAOS article Deshpande writes, "Kaiyata, coming from Kashmir shows no awareness of that pANini received his akSarasamAmnAya or his grammar from ziva. However, he refers once to pataJjali by the term nAganAtha... and this shows his awareness of the motif of pataJjali being an incarnation of zeSa...Haradatta shows his awareness of pANini being inspired by ziva, but not that pataJjali is an incarnation of zeSa. Haradatta being familiar with the local tradition's assocition with the author of yogasUtra need not say anything when he is discussing the grammarian. On the other hand Kaiyata who comes two centuries removed from Haradatta merges both pataJjalis which happens later in Chidambaram also. Being a vaiSNava, Kaiyata is amenable to accepting the pataJjali-zeSa myth but does not accept the myth of pANini and ziva.

Another text which mentions the zaivite origin of pANini's akSarasamAmnAya from ziva is nandikezvara kArika. In a posting dated Tue, 8 Dec 1998 entitled Where was PANini inspired? (Part II), N. Ganesan wrote about nandikezvara kArika thus: "This text is most likely Southern and not Kashmiri as the name, naTarAja never occurs in Kashmiri 'Saivaite texts. Also, Nandikezvara Kaarikaa is never quoted in any Kashmiri text. Furthermore, Upamanyu, the commentator on this work refers to Chidambaram Nataraja's dance and, myths of RSis like vyAgrapAda and pataJjali. Upamanyu bhakta vilAsam is a popular Sanskrit text retelling the PeriyapurANam stories of Tamil 'Saivaite saints (nAyanmAr). Prof. Raffaele Torella informed that "This is strongly my impression" when I queried whether Nandikezvara KarikA is a Southern text. Tamil kuutta nUl, a dance treatise, (13-14th century?) says: "From one side of 'Siva's drum comes Sanskrit letters and the other side gives birth to Tamil letters."..."

This means that it is but natural for nandikezvarakArika from Tamilnadu to mention ziva as the source for pANini's grammar in the same way agastya received the Tamil grammar from ziva. That takes us then to kathAsaritsAgara which links pANini and ziva. Somadeva,a Kashmir zaivite lived in the 11th century AD could have received the tradition from Tamilnadu where it had been mentioned in the 9th century. Finally, maJjuzrImUlakalpa which has been said to belong to 8th century mentioning pANini and avalokitezvara could also be inspired by dakSiNAmUrti/avalokitezvara syncretism located at potiyil as discussed in earlier posts.

Two-way competition

It was not that Chidambaram alone tried to emulate Madurai. Madurai did the reverse too. The ziva myths of Madurai include stories of the Pandyan king learning to dance and begging ziva to switch his legs so that ziva will be more comfortable. The result is of course, the right leg is lifted and left leg is planted , just the opposite of the pose in Chidambaram. In ziva's wedding at Madurai, pataJjali and vyAghrapAda participate and ziva dances for their sake. Of course, it so happens that the author of the first text of Madurai stories is perumparRRap puliyUr nampi named after Chidambaram.


The naTarAja cult at Chidambaram in competition with the dakSiNAmURti cult at Pandyan Madurai has been the source of the myths linking pataJjali with zeSa and pANini receiving the basis of his grammar from ziva. This investigation also reveals that Abhinavagupta of Kashmir was familiar with some of the traditions prevailing at the naTarAja temple Chidambaram and he used them in his writings.

Last modified: Mon Dec 19 10:41:03 2011

[Publisher & contact for INDOLOGY site: Dominik Wujastyk]   Dominik Wujastyk email address

Valid [XHTML | CSS] · [Site Standards]

RSS Feed icon