English only · Odia translation in progress

Chapter 2: The First Machine


The strange thing about the earliest surviving temples of Odisha is that they do not look like beginnings.

They are small. They are squat. Many are damaged. Their sculpture is fragmentary, their fallen pieces mixed into compound walls, their original settings partly lost. But the grammar is already there. The first visible Odishan temple is not a rough trial. It is a working machine in early form.

This is the problem Donaldson places at the start of the architectural story. Bhubaneswar’s earliest surviving shrines already possess the basic components of the later tradition: the rekha sanctum, the curving spire, the divided wall, the raha niche, the subsidiary pagas, the base mouldings, the baranda, the crowning system, the miniature shrine motifs, the relation between architectural frame and sculptural occupant.

The beginning we can see is not the beginning that happened.

Something earlier has disappeared: timber prototypes, brick shrines, ruined stone temples, lost workshops, small sanctuaries whose forms were absorbed into the surviving line. The earliest temples of Bhubaneswar are therefore not primitive. They are survivors from a system whose previous stages are mostly missing.


The Lakshmanesvara Group

The crucial early group is the cluster known as the Lakshmanesvara, Bharatesvara, and Satrughnesvara temples.

They stand near the old sacred geography of Bhubaneswar, tied to the growth of the Tribhuvanesvara-Lingaraja field. They are local, compact, and heavily damaged, but architecturally decisive. They show that by around the late 6th and early 7th century, Odisha already possessed a recognizable temple body.

These shrines are of the rekha order. The sanctum is square. The spire rises curvilinearly. The wall is tri-ratha: a central projection, the raha, flanked by subsidiary pagas. The bada has the early division of pabhaga, jangha, and baranda. The gandi is already more complex than the wall beneath it, using a pancha-ratha organization above a tri-ratha bada. This mismatch is important. It shows a system still working out how to align the tower’s vertical rhythm with the wall below.

The early machine is not clumsy. But it is not yet fully synchronized.

The raha niche in these temples cuts through the base mouldings in a way that simulates a doorway. This is one of the most revealing details. The temple wall remembers the idea of a four-door shrine. The niche is not yet merely a window cut into a wall; it behaves like a blocked sacred entrance. Later, the niche will rise above the pabhaga and become more clearly framed as a sculptural station. In these earliest temples, the memory of doorway and niche still overlap.

That overlap tells us something about origin. The Odishan deul may have emerged from primitive four-door shrines, not from cave architecture or pillared hall traditions. Its wall system preserves that older architectural memory.


The Vajra-Mastaka Logic

The most important early surface motif is the vajra-mastaka.

It appears as a crowning architectural form above niches and miniature shrine designs. In the early temples it is often made of two superimposed chaitya-medallions, sometimes with figures inside them. The subsidiary pagas are shaped as vajra-mundis, miniature shrines crowned by this form.

This is the early temple’s way of turning the wall into a sacred grid. Each side projection is not just a strip of masonry. It is a small shrine. Each small shrine has its own body, niche, and head. The great shrine is covered with little shrines.

On the Satrughnesvara, this system is especially important because it is preserved well enough to show the early logic. The subsidiary pagas terminate beneath the baranda. The raha continues upward and is visually crowned by a large vajra-mastaka rising into the gandi. The gandi itself is divided into bhumis, and the anuraha recesses are filled with superimposed miniature vajra-mundis.

In other words, the wall does not end when the spire begins. Motifs climb. The shrine repeats itself upward. The surface is bound together by miniature architecture.

This is one of the deepest features of the Odishan temple: the building is a hierarchy of repeated building forms.


The Wall Learns Its Images

The early temples are also important because their iconographic program is still becoming stable.

The major raha niches already receive important deities. The Shaiva program is visible: Ganesha, Kartikeya, Parvati, Mahishamardini, Lakulisa, various forms of Shiva, and related Brahmanical figures. But the placement is not yet as standardized as it will become later.

The subsidiary niches contain a wide range of images. Some are small forms of Shiva. Some are related deities. Some are narrative or auspicious figures. The temple is already full of sacred population, but the exact rules of occupation are not fully fixed.

This is what makes the earliest phase valuable. It shows a living system before it becomes fully codified.

Later, the Odishan temple will become more predictable. The parsvadevatas will occupy the major side and rear niches. Dikpalas will find their directional locations. Alasa-kanyas, mithunas, virala motifs, and other decorative figures will assume increasingly standard positions. In the earliest temples, the machine is working but still negotiating its occupants.

This is how architecture and theology mature together. The wall provides positions. The religious program fills them. Over time, repeated use turns placement into law.


Parasuramesvara: The First Complete Statement

If the Lakshmanesvara group shows the early grammar, the Parasuramesvara temple shows its first major surviving statement.

It is the best preserved specimen of this early phase and crucial because it preserves both deul and jagamohana. Donaldson treats it as a key monument not because it is the largest, but because it reveals how the early system had already learned to coordinate architecture, sculpture, and surface rhythm.

The Parasuramesvara is still an early temple. The deul is compact. The wall remains tri-ratha. The pabhaga has three mouldings. The subsidiary pagas are vajra-mundis. The raha is treated as a truncated rekha form. The gandi carries the major vajra-mastaka program. But the carving is more confident, the surface more varied, and the integration more complete than in the earlier group.

The jagamohana matters enormously.

In later Odishan architecture, the jagamohana becomes the standard partner of the deul. In the Parasuramesvara, we see an early surviving hall attached to the sanctum. It is not yet the fully mature pidha hall of later temples, but it proves that the two-part temple machine was already developing. The hall is not a casual porch. It is an architectural partner, a space of approach, transition, and additional surface program.

The relation between deul and jagamohana would remain one of the central questions of Odishan architecture. Sometimes a hall was added later. Sometimes bonding between the two was imperfect because of construction methods. But the mature tradition depends on their pairing.

Parasuramesvara helps show that the pairing was already present early, even if not yet in its later perfected form.


The Early Temple Is Dense, Not Large

The early Odishan temple does not achieve power through size.

Its scale is intimate. Its sanctum is small. Its wall is not yet tall enough to carry the later two-story jangha. The gandi has fewer bhumis. The jagamohana is modest. Compared with Lingaraja or Jagannath, these temples seem almost miniature.

But the density is extraordinary.

Every part of the surface is activated. The pabhaga carries mouldings. The jangha receives niches and miniature shrines. The baranda becomes a line of separation and narrative possibility. The gandi rises with repeated motifs. The doorframe becomes a major sculptural and ritual threshold. The dvarapalas, river goddesses, grahas, chaitya forms, scrollwork, ganas, and divine figures all take their places.

This is not minimal architecture. It is compressed architecture.

The later Odishan temple will expand this compression. It will stretch the wall upward, multiply the pagas, split the jangha, regularize the iconography, enlarge the jagamohana, and increase the gandi’s height. But the logic is already there in small form.

The first machine is a seed, not a sketch.


The Question of Influence

The earliest temples also raise the question of influence.

There are traces of Gupta idiom, connections with broader North Indian forms, and possible external contacts. But Donaldson’s larger argument resists treating Odisha as a passive receiver of styles from elsewhere. The surviving temples show an indigenous system already strongly formed.

The point is not that Odisha developed in isolation. No temple tradition does. Motifs, craftsmen, religious ideas, and technical solutions moved across regions. But the Odishan temple did not become itself by simply copying another school. It absorbed and reorganized.

This distinction matters.

A borrowed motif placed into an Odishan wall becomes Odishan because the wall gives it a specific architectural address. A form from elsewhere is flattened, contained, aligned, and made to serve the local surface system. The temple is not a museum of influences. It is a machine that processes influence.

That is already visible in the earliest phase.


The Compression

The earliest surviving temples of Bhubaneswar establish the central paradox of Odishan architecture:

The record begins late enough that the system is already mature, but early enough that we can still watch it becoming disciplined.

The Lakshmanesvara group shows the compact tri-ratha shrine, the early relationship between wall and tower, the dominance of the vajra-mastaka, and a still-fluid iconographic program. Parasuramesvara shows the system becoming more assured, more integrated, more sculpturally rich, and more capable of joining deul and jagamohana into a coherent sacred machine.

What comes next is not invention from nothing.

It is learning.

Bhubaneswar spends the 8th and 9th centuries testing the grammar: raising niches, adding pilasters, experimenting with khakhara forms, refining the jagamohana, and turning an early sacred machine into a city-wide architectural language.