The Twelve Jyotirlingas

The pillar no one could fathom

The story the tradition tells about Shiva does not begin with a temple. It begins with an argument.

In the Shiva Purana, in the section called the Vidyeshvara Samhita, Brahma and Vishnu fall to quarrelling over which of them is supreme. While they argue, something appears between them. A column of fire. It has no visible top and no visible bottom. They strike a bargain. Whoever finds where it ends is the greater god. Vishnu takes the form of a boar, varaha, and burrows downward to find its base. Brahma takes the form of a swan, hamsa, and flies upward to find its crown. They go for a long time. Neither reaches the end.

The chapter titles of the Vidyeshvara Samhita carry the whole sequence in plain order: chapter 6, “Battle between Brahma and Vishnu”; chapter 7, “Shiva manifests himself as a column of fire in the battlefield”; chapter 8, “Shiva’s forgiveness of Brahma”; chapter 9, “The Proclamation of Shiva as Maheshvara (the great lord)” (J.L. Shastri trans., Motilal Banarsidass). Brahma, unwilling to lose, claims he reached the top, and produces a fallen ketaki flower as a false witness. Shiva exposes the lie, curses the ketaki out of his worship, and is proclaimed Maheshvara, the great lord, the one beyond measure and beyond form.

That episode has a name. The Lingodbhava, “the arising of the linga.” It is the theological root of the word jyotirlinga. The point the story makes is not subtle and not flattering to anyone except Shiva: the absolute cannot be reached by climbing or by digging. It has no end to find.

Now hold that beside what a jyotirlinga claims to be.

The word breaks into two. Jyotis, light or radiance. Linga, the “mark” or “sign,” the aniconic emblem of Shiva, which the tradition reads as the sign of the formless and not as a literal object. A jyotirlinga is a “linga of light.” And the twelve great ones are held to be places where that same fathomless pillar of fire from the Lingodbhava broke the surface of the earth and stayed. Self-manifest. Svayambhu, self-arisen, as against pratishthita, a linga shaped by a sculptor and installed by a priest (Shiva Purana, Kotirudra Samhita 1, “The greatness of Jyotirlingas and their Upalingas”). That last distinction is the tradition’s own self-understanding, and it is worth flagging as such. No archaeologist can confirm that a stone arose by itself. What the texts assert is that these twelve were not made.

So the whole subject sits on a tension that never resolves, and is not meant to. The deity who has no top and no bottom, whom two gods could not measure, is approached at twelve fixed addresses you can walk to. The formless, made locatable. The light is the reminder that the located god is not contained by the location.

The twelve and their single source

The number twelve and the canonical names come from the Shiva Purana, in the Kotirudra Samhita, which gives each of the twelve its own chapter of origin and mahatmya, “greatness.” (Chapter numbers in this article follow the J.L. Shastri / Motilal Banarsidass edition as indexed on wisdomlib; chapter divisions vary between printed recensions of the Shiva Purana, so the numbers are a guide, not a fixed coordinate.) On that edition the section runs chapters 14 through 33, and it opens by framing these twelve as the apex of linga-worship, each with subsidiary upalingas beneath it.

But most people do not learn the twelve from the Purana. They learn them from a short hymn, four verses long, the Dvadasa Jyotirlinga Stotra, which functions as a mnemonic that holds the whole network in the memory. It is traditionally attributed to Adi Shankara, the philosopher usually dated to the eighth century. That attribution is traditional, and it is worth being honest about it. Many stotras are ascribed to Shankara by convention, and this one is not text-critically established as his. The tradition says Shankara. The manuscript record does not prove it.

There is a reason the mistake is so easy to make, and it is worth setting down. A second hymn exists, also called the Dvadasa Jyotirlinga Stotra, and that one does carry a colophon naming Shankara: iti shrimad-shankaracharya-virachitam dvadasha-jyotirlinga-stotram sampurnam, “thus ends the twelve-jyotirlinga hymn composed by Shankaracharya.” It is a different composition altogether — twelve verses rather than four, one full verse for each shrine, in a longer metre, and it opens Saurashtra-deshe vishade’tiramye, “in the bright and lovely land of Saurashtra,” before naming Somanatha. Our quatrain opens Saurashtre Somanatham cha. Two hymns of the same name, both beginning in Saurashtra with the same god: a conflation waiting to happen. The ascription that properly belongs to the long hymn appears to have slid sideways onto the short one, which the manuscript record files simply as traditional, with no author and no colophon at all. The quatrain is the one everybody recites. It is also the one nobody can put a name to.

Here is the famous opening verse, in Devanagari and transliteration, after greenmesg.org:

सौराष्ट्रे सोमनाथं च श्रीशैले मल्लिकार्जुनम् । उज्जयिन्यां महाकालम् ॐकारममलेश्वरम् ॥१॥

Saurashtre Somanatham ca Shrishaile Mallikarjunam, Ujjayinyam Mahakalam Omkaram Amaleshvaram.

Dvādaśa Jyotirliṅga Stotra 1 #

The verse is a list of places and the lords who dwell in them. At Saurashtra is Somanatha, and at Shrishaila is Mallikarjuna; at Ujjayini is Mahakala, at Omkara is Amaleshvara. The next two verses run through the rest:

परल्यां वैद्यनाथं च डाकिन्यां भीमाशंकरम् । सेतुबंधे तु रामेशं नागेशं दारुकावने ॥२॥

Paralyam Vaidyanatham ca Dakinyam Bhimashankaram, Setubandhe tu Ramesham Nagesham Darukavane.

Dvādaśa Jyotirliṅga Stotra 2 #

वाराणस्यां तु विश्वेशं त्र्यंबकं गौतमीतटे । हिमालये तु केदारं घुश्मेशं च शिवालये ॥३॥

Varanasyam tu Vishvesham Tryambakam Gautamitate, Himalaye tu Kedaram Ghushmesham ca Shivalaye.

Dvādaśa Jyotirliṅga Stotra 3 #

At Parali is Vaidyanatha and at Dakini is Bhimashankara; at Setubandha is Ramesha, and Nagesha is in the Darukavana. At Varanasi is Vishvesha, and on the bank of the Gautami (the Godavari) is Tryambaka; in the Himalaya is Kedara, and at Shivalaya is Ghushmesha. And the closing verse turns the list into a practice:

एतानि ज्योतिर्लिङ्गानि सायं प्रातः पठेन्नरः । सप्तजन्मकृतं पापं स्मरणेन विनश्यति ॥४॥

Etani Jyotirlingani sayam pratah pathen-narah, Sapta-janma-kritam papam smaranena vinashyati.

Whoever recites these jyotirlingas evening and morning, the sins of seven births are destroyed by the very remembrance.

Dvādaśa Jyotirliṅga Stotra 4 · the phalaśruti #

Note what that last verse does. You do not have to walk the circuit. You can say it. The hymn carries the whole geography of the subcontinent in four short verses, and promises that the remembering is itself the pilgrimage.

A walk through the twelve

The twelve are not a flat list, and reading them as one misses the point. Each carries a legend, a form, a place. Some of those legends are vivid enough to have shaped the daily life of a town for a thousand years. What follows is each site’s story as the tradition tells it; the narrative voice below is the tradition’s, the myth of the texts, and not a historical claim. Here they are, grouped roughly as the land arranges them.

The west, and the moon’s reprieve. Somnath stands at Prabhasa, on the coast of Gujarat, in the present Gir Somnath district (Kotirudra Samhita 14). Its legend is among the gentlest. Candra, the Moon, also called Soma, had married the twenty-seven daughters of Daksha but loved only Rohini. The slighted Daksha cursed him to waste away, which is the tradition’s reading of why the moon thins to nothing. On the gods’ advice, Soma worshipped Shiva here at Prabhasa, and Shiva softened the curse so that the moon would wane and wax rather than die outright, and remained as the linga. Soma-natha. The Lord of the Moon. The waxing you watch in the night sky is, in this telling, a curse already commuted.

The south of the Deccan, and a son’s sulk. Mallikarjuna sits on the hill of Shrishailam in Andhra Pradesh, in the present Nandyal district (Kotirudra Samhita 15). The story turns on a family quarrel. Kumara, also called Kartikeya, lost a contest to his brother Ganesha and withdrew in hurt pride to Mount Krauncha. Shiva and Parvati, missing their son, followed him to the hill to be near him. The name Mallikarjuna is read two ways in the tradition: as mallika, jasmine, the flower Parvati offered, joined to Arjuna, a name of Shiva, and also as Mallika for Parvati joined to Arjuna for Shiva. Parvati here is worshipped as Bhramaramba. The regional development of the Krauncha episode draws also on Skanda Purana material.

Ujjain, and the face that turns to death. Mahakaleshwar rises at Ujjain in Madhya Pradesh, the old Avanti (Kotirudra Samhita 16 to 17). The brahmins of Avanti, in the telling, were tormented by a demon, and Shiva is said to have burst from the earth as a fiery form, Mahakala, “Great Time” or “Great Death,” to incinerate it, then stayed as the self-manifest linga. The demon’s name and the devotee’s name vary between retellings; the manifestation is the secure core. Tradition singles Mahakaleshwar out for its orientation, honouring the deity as dakshinamukhi, south-facing, south being the direction of Yama, the lord of death. (This is a point of devotional emphasis about the ritual facing of the deity, often phrased as the only south-facing jyotirlinga; it is not a claim the Kotirudra Samhita makes.) The linga that conquered death looks death in the face. The temple’s most famous rite is the bhasma-arati, the dawn worship in which the deity is honoured with sacred ash, bhasma, a daily reminder of what time reduces everything to.

The island shaped like a syllable. Omkareshwar stands on Mandhata island in the Narmada, in Khandwa district, Madhya Pradesh (Kotirudra Samhita 18). Two legends braid here: the penance of the Vindhya mountain, which wished to outrank its rival Meru, and the austerities of the Ikshvaku king Mandhata, at whose request Shiva is said to have settled on the island. The island is held to be shaped like the syllable Om. The deity is held in two parts, Omkareshwar on the island and, on the south bank, a shrine variously called Mamaleshwar, Amareshwar, or Amaleshwar. The stotra’s Amaleshvaram is usually read as this second aspect, though the verse reads Omkaram Amaleshvaram, which some take as a single name for the one jyotirlinga. Which of the two is “the” jyotirlinga is itself a question the tradition has not settled.

The Himalaya. Kedarnath sits high in the mountains of Rudraprayag district, Uttarakhand, near the snowline (Kotirudra Samhita 19). Here it matters which book you are holding. The Shiva Purana version is austere: the sages Nara and Narayana, forms of Vishnu, performed long austerities worshipping an earthen linga, and Shiva, pleased, manifested at Kedara as a jyotirlinga “for the welfare of the world.” The far more famous story, the one most pilgrims tell, of the Pandavas pursuing Shiva, of Shiva diving into the earth as a bull and leaving his hump behind at Kedarnath, and of the Panch Kedar circuit of five shrines, belongs to a different stream. That is Mahabharata-adjacent material developed in the Skanda Purana, in the Kedara Khanda, as local sthala tradition. It is a real and ancient story. It is simply not the Shiva Purana’s, and the two should not be conflated.

The Sahyadri, and a river born of heat. Bhimashankar lies in the Western Ghats, the Sahyadri, in Khed taluka of Pune district, Maharashtra (Kotirudra Samhita 20 to 21). A demon named Bhima, born in the Shiva Purana telling of Kumbhakarna’s line, persecuted a Shiva-devotee king (named Sudakshina in some tellings). When Bhima struck at the linga itself, Shiva is said to have erupted and reduced him to ash, and the Bhima river is said to spring from the heat of that battle. A few alternative locales have been claimed over the years, but the Sahyadri identification near Pune is the mainstream one.

The city of light. Kashi Vishvanath stands at Varanasi in Uttar Pradesh (Kotirudra Samhita 22 to 23). Of all the twelve this is the one where the light-theme becomes self-referential, because Kashi itself means “the city of light.” The tradition holds it to be Shiva’s own city, which he never forsakes, which is why one of its names is Avimukta, “the never-abandoned.” He abides there as Vishveshvara, “Lord of All,” and grants moksha, liberation, to those who die within the city’s bounds. Kashi is said to rest on Shiva’s trident, lifted clear of cosmic dissolution, pralaya, so that when the rest of creation is unmade, the city of light remains. The enormous expansion of this material is in the Skanda Purana, in the Kashi Khanda. The light that the tradition glosses as consciousness, and the city where to die is to merge with it. The two ideas are, in its telling, the same idea.

Nashik, and three faces. Tryambakeshwar stands in Nashik district, Maharashtra, at the source of the Godavari (Kotirudra Samhita 24 to 27). The sage Gautama, falsely accused of cow-slaughter, propitiated Shiva and asked the Ganga to descend and cleanse the place. Shiva released the Gautami Ganga, which is the Godavari, at Brahmagiri, and remained as Tryambaka, “the three-eyed.” The linga is unusual in showing three faces, glossed in the tradition as Brahma, Vishnu, and Rudra, and in some tellings as Brahma, Vishnu, and Maheshvara.

The physician. Vaidyanath, “the divine physician,” is the ninth (Kotirudra Samhita 28). The legend is one of the most striking. Ravana undertook a ferocious penance, offering his own heads one by one into the fire to win Shiva’s power. Shiva restored him and acted as his vaidya, his physician, hence Vaidyanatha. In the attached story, Ravana was carrying the linga toward Lanka on the condition that it not touch the ground; the gods tricked him into setting it down, and it became fixed where it stands. The legend is secure. The ground is not, and I will come to that.

The lord of serpents. Nageshwar, “Lord of the Nagas,” dwells in the Darukavana, the forest of the legend (Kotirudra Samhita 29 to 30). A demon named Daruka ruled a sea-forest and imprisoned a steadfast devotee named Supriya. Supriya led the other captives in chanting Om Namah Shivaya, and Shiva manifested, gave him the Pashupata weapon, destroyed the demons, and remained as Nageshwar, lord of serpents and, in popular devotion, a protector against poison. Where the Darukavana actually is, the text does not say, and that silence is the whole of the dispute over this one.

The bridge to the sea. Rameshwaram, the lord worshipped by Rama, stands on Rameswaram island in Ramanathapuram district, Tamil Nadu, at the far southern reach of the set (Kotirudra Samhita 31). The Shiva Purana core is simple: Rama worshipped Shiva to expiate the sin of the coming war and to seek victory over Ravana. The vivid episode that most pilgrims tell, of Hanuman sent to fetch a linga from Kailasa or Varanasi, of his delay, and of Sita fashioning a linga from sand so the auspicious hour would not pass, giving the shrine its famous two lingas, belongs mainly to the Ramayana tradition and the Setu-mahatmya of the Skanda Purana. (Most tellings credit Sita with the sand linga during Hanuman’s delay; a few credit Rama.) It is a beloved story, and it is sthala tradition rather than the Kotirudra Samhita’s own.

The devotee’s name. Grishneshwar, also Ghushneshwar, stands at Verul near the Ellora caves in Maharashtra, in the Chhatrapati Sambhajinagar district, formerly Aurangabad, the twelfth and last (Kotirudra Samhita 32 to 33). It is the only one of the twelve named for an ordinary woman. The devout Ghushma, called Kusuma in some tellings, co-wife of Sudeha, daily made and worshipped a hundred and one clay lingas. When the jealous Sudeha murdered Ghushma’s son and threw the body into the tank, Ghushma’s faith did not break; she continued her worship. Shiva, moved, revived her son and appeared as the jyotirlinga, taking her name. The supreme list of self-manifest lights closes not with a god’s quarrel or a demon’s defeat, but with the steadiness of one woman who would not stop praying.

The honest complications

Two of the twelve do not sit cleanly on the map, and a serious account has to say so.

Vaidyanath, the ninth, is claimed by three places, and the Government of India has not officially notified any one of them as the jyotirlinga. The best-known is Baidyanath Dham at Deoghar in Jharkhand, where the ancient name of the locale is argued to match the cremation-ground setting of the legend. But Parli Vaijnath in Beed district, Maharashtra, reads the stotra’s own word Paralyam, “at Parali,” as naming its own Parali, and points to its position among local rivers. And Baijnath in the Kangra valley of Himachal Pradesh is a third claimant. The stotra’s own wording at this point varies between manuscripts, which is part of why the dispute persists. There is no verifiable winner. The three coexist.

Nageshwar, the tenth, has the same problem, rooted in that unfixed Darukavana. Nageshwar near Dwarka in Gujarat is the most popular today, reading Darukavana as the forest near Dwarka. Aundha Nagnath in Hingoli district, Maharashtra, makes the case for a different forest. And Jageshwar in Almora, Uttarakhand, argues that daru means the deodar, devadaru, the cedar forests of the Himalaya, which would put the shrine in the mountains. Because the text never pins Darukavana to a coordinate, all three readings are textually possible. Popular practice counts Dwarka. The ambiguity is real, and it is honest to leave it open.

There is a quieter point underneath both disputes. The count of twelve is stable, and the Kotirudra Samhita’s names are the standard list, though tradition preserves minor variant enumerations too. What is not stable is the on-the-ground identification of a few of them, and the legends themselves often live in more than one book. The Shiva Purana gives one Kedarnath and one Rameshwaram; the epic and sthala traditions give another, richer set. Reading the record carefully, the cult of the jyotirlingas appears older and more layered than any single one of its temples.

History, where it is solid

Two of the twelve carry documented histories of destruction and rebuilding, and both deserve to be told plainly, without heat.

Somnath was plundered by Mahmud of Ghazni in around 1025 to 1026 of the Common Era, during the campaign dated to 416 AH; the linga was broken and an enormous treasure carried off, with Turko-Persian chronicles reporting figures in the millions of dinars, the kind of inflated and inconsistent number that the historiography below treats with caution. It was rebuilt repeatedly across the eleventh to thirteenth centuries under the Chalukya and Solanki kings, attacked again under the Delhi Sultanate, and rebuilt again. In November 1947, Sardar Vallabhbhai Patel visited the ruins and pledged reconstruction; on Gandhi’s advice the rebuilding was funded publicly rather than by the state, and a Somnath Trust was formed under K.M. Munshi, whose own memoir, Somanatha: The Shrine Eternal, records the effort from the inside. The ruins were cleared by 1950, and the new temple was consecrated by President Rajendra Prasad in May 1951. That consecration was itself debated. Prime Minister Nehru questioned whether the head of a secular state should inaugurate a temple, and President Prasad attended in what he framed as a personal capacity, which is part of the same making of meaning the historiography traces.

That is the documented frame. The historian Romila Thapar, in Somanatha: The Many Voices of a History (2004), adds a caution that any careful reader should carry. Her argument is that the meaning of the 1026 raid was constructed long after the event. She shows that the reading of it as a collective Hindu trauma crystallised in 1843, in a debate in the British House of Commons over the so-called gates of Somnath, and that the Turko-Persian chronicles glorifying the raid, which became the main source for later historians, contradict one another even on basic facts, such as the nature of the idol. Against those chronicles she sets contemporary local Sanskrit and Jain sources, inscriptions, and merchant records, which show a more layered picture, including ongoing trade and patronage rather than one uniform civilisational rupture. The plunder and the rebuilding are real history. So is the later making of their meaning. Both belong in an honest account, and the modern politics is not ours to adjudicate.

Kashi Vishvanath carries a parallel history. The shrine was demolished on the order of Aurangzeb in 1669, and the Gyanvapi mosque was raised on or beside the site, taking its name from the Gyan-vapi, the “well of knowledge.” The present temple was built just south of the mosque by Ahalyabai Holkar of Indore around 1780, and Maharaja Ranjit Singh gifted the gold plating for its spire in 1835, which is why Kashi has its “golden” temple. The Gyanvapi question is a live legal and archaeological dispute. The factual statement, and the only one worth making here, is that a mosque stands adjacent to the temple on the site of the pre-1669 structure, and that the matter is in litigation. The outcome is not for this page to characterise.

One body across a subcontinent

Step back from the twelve and a pattern appears that no single temple shows on its own.

They span the whole land. Kedarnath at the Himalayan snowline and Rameshwaram at the southern sea mark the vertical reach; Somnath on the western coast and Kashi on the Gangetic plain mark the horizontal. To pilgrimage the full set, the dvadasa-jyotirlinga yatra, is, in the tradition’s imagination, to trace Shiva’s presence across the entire breadth of the land.

The scholar Diana Eck, in India: A Sacred Geography (2012), reads this as something more than a coincidence of shrines. Her thesis is that networks of pilgrimage places, the four dhams, the shakti-pithas, the seven sacred rivers, and the twelve jyotirlingas among them, have over centuries composed India’s sense of region and belonging, and even its sense of itself as a connected land. On her reading, the grouping of twelve functions less as twelve isolated temples than as a single trans-regional network, a device that knit independent regional Shiva cults into one imagined sacred whole. This is Eck’s scholarly reading, and it should be held as that, alongside, and not in place of, the tradition’s own claim that the twelve are simply where the light arose. Her earlier Banaras: City of Light (1982) does for Kashi what the whole set does for the country, treating the city as the paradigm of light, which is the thread that ties it back to jyotis.

The stotra is what holds the network together for a person who will never make the journey. Recited morning and evening, it walks all twelve in under a minute, and tells the reciter that the sins of seven births fall away in the remembering.

The formless, in a fixed place

So we end where the Lingodbhava began. With a pillar of light no one could measure.

The twelve jyotirlingas are the boldest answer the tradition gives to a problem every embodied devotee has. The absolute is everywhere, which means it is nowhere a person can go. You cannot make a pilgrimage to “everywhere.” And so the same teaching that says Shiva has no top and no bottom also says that the limitless consented to break the surface of the earth at twelve points and stay.

The light is the safeguard against the obvious mistake. It is there, in the tradition’s reading, to say that the deity at the named place is not shrunk to fit the place. Jyotis is light, which the tradition glosses as the light of consciousness, and consciousness is not held in a stone any more than the column of fire had a crown for Brahma to find.

The tradition does not need this article to keep the twelve lights burning. They have been lit, in its telling, since before the first temple was raised over any of them. We have only walked the circuit on paper, named the places, marked where the record is firm and where the ground is still contested. The pillar has no end to find.

That was the whole point of it.