Nagara, Dravida and Vesara: The Styles of the Hindu Temple

Stand in front of an old Hindu temple and the first thing that reaches you is the tower. In one place it curves upward like a tall beehive and tapers to a stone disc. In another it climbs in clean horizontal storeys, a stone pyramid stepped like a ziggurat. The shrine inside may be small and dark. The tower above it is the signature, and from the tower alone a trained eye can tell you roughly where in India you are standing.

That difference is the heart of the classical scheme. Hindu temple architecture is traditionally sorted into three families: Nagara, Dravida and Vesara. Nagara belongs mainly to the north, Dravida mainly to the south, and Vesara to the Deccan in between. That is the short answer. The longer answer is more interesting, because the categories did not begin as a map.

Where the three names come from

The threefold classification is set down in the Shilpa and Vastu texts, the technical manuals of building, among them the Mayamata and the Manasara (both medieval South Indian treatises on architecture, cited by chapter). The point worth holding onto is that these texts did not first define the three styles by region. They defined them, in part, by the shape of the temple’s plan.

In that older register, Nagara meant a building on a square plan. Dravida meant an octagonal or polygonal plan. Vesara meant a round or apsidal one. The north, south and Deccan associations came later, as a secondary convention layered over the original geometry. Both registers are real, and both are worth knowing, because the regional reading is so common today that the plan-shape reading has almost vanished from popular accounts. Keep both in mind. The tradition that gives us these words is older and stranger than the tidy north-south-Deccan summary suggests.

There is a reason the plan-shape reading matters beyond pedantry. It tells us that the early theorists were thinking in geometry first and geography second. A square, an octagon, a circle. Those are abstract forms, and the men who wrote the manuals treated the temple as a worked-out diagram before they treated it as a regional product. When the labels later attached themselves to the map of India, they did so unevenly, and the fit was never perfect. That looseness at the root is part of why the third category, in particular, has never sat still.

Tradition holds that these forms are laid down by the shastras, the revealed canon, so that the builder is following a given pattern. The historical record reads them differently. To the historian the styles are empirical developments, regional and chronological, reconstructed from the monuments themselves and from inscriptions. Both accounts are presented here. They do not have to be reconciled to be understood.

Nagara: the curving tower of the north

The Nagara temple is built around a square sanctum, and over that sanctum rises a curvilinear spire called the shikhara (literally “mountain peak,” the whole tower in northern usage). The line of the shikhara curves inward as it rises, giving the beehive profile. At the summit sits the amalaka, a ribbed stone disc named for the amalaka fruit, and above that a kalasha, a pot-shaped finial. Sanctum, curving spire, ribbed disc, finial. That vertical stack is the north’s basic grammar.

Within Nagara the spire develops into recognised sub-types. Adam Hardy, in The Temple Architecture of India (2007), describes the latina (a single spire), the shekhari (a central spire clustered with subsidiary half-spires, the urushringa) and the bhumija (a lattice of miniature spirelets running up the tower). The shekhari is what gives the great central Indian temples their mountainous, many-peaked look.

The examples run across centuries. The Dashavatara temple at Deogarh dates to around the early 6th century, in the Gupta period, and is among the earliest stone temples to carry a true northern spire. The Kandariya Mahadeva at Khajuraho, built by the Chandela dynasty around 1025 to 1050 CE, is the high point of the shekhari mode, its central spire surrounded by a rising swarm of half-spires. The builder-king is sometimes named, but the attribution is not certain, so it is left open here. The Lingaraja at Bhubaneswar belongs to the 11th century and shows the eastern, Kalinga branch of the Nagara family. The Sun Temple at Konark, around 1250, is the dramatic late flowering of that same eastern line, conceived as the chariot of the sun god with carved stone wheels along its base. Read across these four buildings and you watch a single idea grow. The spire begins as one clean curve at Deogarh, gathers a crowd of half-spires at Khajuraho, and by Konark has become a whole architectural drama. The grammar held for seven hundred years. What changed was the ambition.

Dravida: the stepped pyramid of the south

The southern temple climbs differently. Its tower, the vimana, is a stepped pyramid that rises in clearly marked storeys called talas, each storey a shrunken tier of the one below. There is a crucial terminological trap here. In the south the word shikhara does not name the whole tower. It names only the domed crowning member that sits at the very top of the vimana. So a sentence about a “shikhara” means one thing in Bhubaneswar and another in Thanjavur. The same word, two structures. Hold the distinction.

The Dravida temple is also defined by what surrounds the shrine. Enclosure walls, the prakara, ring the sanctum in concentric rectangles. Cut into those walls are the gopuram, the great gateway towers, which in the largest southern temples grow taller than the vimana itself.

The chronology is anchored by named dynasties and, in the best cases, by inscriptions. The Shore Temple at Mahabalipuram was built under the Pallava king Narasimhavarman II, also called Rajasimha, around 700 to 728. The Kailasanatha at Kanchipuram belongs to the early 8th century, also Pallava. The Brihadishvara temple at Thanjavur, also called the Rajarajeshvara, was raised by the Chola emperor Rajaraja I and consecrated in 1010 CE, a date the inscriptions on the temple itself record. Its vimana is one of the tallest in India. At Madurai, the Meenakshi temple is an old foundation, but its famous towering gopurams are largely the work of the Nayaka rulers in the 16th and 17th centuries. The shrine is ancient. The skyline is comparatively late. This matters for anyone trying to date a southern temple by eye, because the most visible parts, the towering gateways, are often the youngest. A great southern temple is rarely the work of one reign. It is a precinct that accumulated over centuries, each dynasty adding an outer wall and a taller gateway, so that the oldest shrine ends up buried deepest at the centre and the tallest tower stands farthest out.

Vesara: the contested third

Then there is the third name, and here the ground gets soft. Vesara is described in the texts as a hybrid: a Dravidian plan dressed in Nagara detailing, associated with the Deccan. The dynasties usually placed under this heading are the Badami Chalukyas, the Rashtrakutas, the Western Chalukyas and, above all, the Hoysalas. The Chennakeshava temple at Belur was dedicated in 1117 CE. The Hoysaleshvara at Halebidu was built across a longer span, roughly 1121 to 1160. The two are often spoken of together, but the dates are not the same, and the Halebidu temple should not be collapsed into the Belur one.

Now the difficulty. The term Vesara is contested. Adam Hardy avoids it, preferring Karnata Dravida for the Deccan tradition, on the grounds that these buildings are a regional development of the Dravidian system rather than a true blend. George Michell, in The Hindu Temple (1977), also avoids the word. And the Shilpa texts themselves disagree among one another about what Vesara actually denotes, whether a plan shape, a regional manner, or a mixture. This is a live debate among serious scholars. It is not settled here. The honest position is that Vesara is a real word in the classical literature and a contested construct in the modern study of these monuments, and a reader should know both.

What is not in dispute is what the Deccan builders actually achieved. The Hoysala temples are among the most densely carved structures in India, built on star-shaped plans from a soft soapstone that hardens with age, their walls covered in horizontal bands of figures so fine they look turned on a lathe. Whether you file them under Vesara or under Karnata Dravida, the buildings remain. The argument is about the name we hang on them, and that argument is a useful reminder that the threefold scheme is a tool of classification, applied after the fact, rather than a law the builders consciously obeyed.

The broad shape of the history

Step back, and a rough timeline emerges. The earliest surviving free-standing stone Hindu temples belong to the Gupta era, roughly the 4th to 6th centuries CE, and they grew out of an older tradition of rock-cut shrines carved into living cliffs. From there the building tradition accelerates. The great age runs from about the 7th to the 13th century, carried by the Pallavas and Cholas in the south, the Chandelas in the centre, the Hoysalas in the Deccan, and the Eastern Gangas on the Kalinga coast.

The scholarship that reconstructed this history is worth naming. Stella Kramrisch’s The Hindu Temple (1946) read the building as a cosmic diagram and a symbol of the manifested universe. Percy Brown’s Indian Architecture (1942) laid out the survey of styles for a general readership. George Michell’s The Hindu Temple (1977) became the standard introduction to the meaning and the forms together. Adam Hardy’s Indian Temple Architecture: Form and Transformation (1995) and The Temple Architecture of India (2007) reworked the typology from the ground up, which is why his vocabulary, latina, shekhari, bhumija, Karnata Dravida, now shapes how the field talks.

Reading a temple now

The threefold scheme is a starting point, not a cage. Most temples sit cleanly inside one family. Some do not, and the awkward cases are often the most instructive, because they show a living tradition borrowing and adapting rather than copying a fixed canon. If the symbolic logic of the ground plan interests you, the companion piece on the Vastu-Purusha-Mandala explains the diagram beneath the stone. For the wider question of what a temple is for, see what is a Hindu temple. The eastern Nagara line reaches its climax at Konark, the Sun Temple, and the same Dravida and Nagara forms house the great Shiva shrines gathered in the Twelve Jyotirlingas. The texts that frame all of this, the Puranas alongside the Shilpa manuals, are introduced in what are the Puranas.

Next time a tower rises in front of you, look at how it climbs. If it curves, you are likely in the north. If it steps, you are likely in the south. And if a guide tells you with great confidence that the building before you is pure Vesara, you now know enough to ask, gently, which scholar they have in mind.