
II. PARALLEL GENESIS
The earliest extant stone and brick shrines in both India and Cambodia are strikingly similar, standing at the inception of their respective traditions and dating from the “contact horizon” between South and Southeast Asian civilizations, roughly from the 2nd to 4th Centuries CE. This cultural exchange, largely from India to the east, was characterized by selective assimilation by local elites rather than direct diffusion through the general population which retained its animist beliefs. Indian merchants appear to have been the primary agents of transfer since they were referred to as pasupatis in recognition of the religious articles they brought. Pasupati (Sans.> pasu animals + pati lord of) was a pre-Vedic deity appearing on Harappan seals dating from 2000 to 1300 B.C.E., later though to have been syncretized with Rudra/Shiva.
After this initial, direct contact, the region’s Austronesian-speaking, thalassocratic states acted as the primary intermediaries in transmitting Indic beliefs along the trade routes through the Straits of Malacca and across the isthmus of Kra. They left irrefutable evidence at the Saliendras’ unprecedented, five-terrace, Buddhist stupa at Borobudur, the Hindu Sanjaya’s answer to it, the equally ambitious Saivite precinct at Prambanan, both in central Java, both 9th Century, as well as the Cham sanctuary at My Son in present-day central Vietnam, (active between the 4th – 13th Centuries.) Jayavarman II (r.802 – c.850,) founder of the “Second Khmer Empire” at Angkor, may have spent his youth at a Javan court, possibly as a hostage, while Champa was to become the principal adversary of his empire.

The earliest Gupta and Khmer prasats (Sans.> prasada) were usually square, single-celled sanctuaries housing a garbhagriha with a cult statue, similar to a cave temple except above-ground with external walls. From the first, their most conspicuous feature was their shikhara (Sans.> peak,) the tower above their cellae, which could be seen as isomorphic with a chaitya-griha’s elevated nave, suggesting they may have had prototypes other than cave temples. These superstructures were made with tiers of rows of unitary aedicules or compressed versions of themselves layered to form a stepped-pyramidal profile (now often decapitated.) Their garbhagrihas’ devakostas, the thick walls necessary to support these superstructures, contained portals, “blind” doors and windows and were decorated with accomplished terra cotta or brick reliefs of gods, as well as scenes from the two Hindu epics. Their portals’ arches or pediments were carried up through the talas or tiers of their shikharas, forming a projecting central strip which would become a bhadra.
India: Bhitargaon
The best-preserved and possibly earliest free-standing Hindu temple in India is located at Bhitargaon, a village near Kanpur in Uttar Pradesh state dating from late 5th Century Gupta rule. Its dedicatee is unknown but a nearby shrine of similar design and distinction from the early 6th Century, the Dashavatara temple at Deogarh (below right,) was dedicated to Vishnu’s ten avatars. The Bhitargaon temple is additionally noteworthy for its use of the “true” or Roman arch, that is, with keystone and voussoirs, not employed in subsequent Hindu or Khmer architecture.
The brick prasada (below left,) 31 feet on a side, with an unusual, double-story, inner sanctum half that width, has an adhisthana (base or pedestal) comprised of kumbha (pot,) gala (shrine recess or antarapatta) and kapota (eave) moldings which rise from a jagati (raised platform.) The prasada is triratha, with a wide, vertical central strip, a proto-bhadra, flanked by prati-bhadras, which are in this case also its karna (corner) rathas (pagas, or projections) and run from its jagati to the top of its shikhara. The superstructure is believed to have been divided horizontally into four talas (bhumis, pidas levels, tiers,) lined with arcuated bays, most containing square terra cotta bas relief.

These rows or colonnades of arched niches are capped by a kapota (eave) molding, itself containing smaller arches similar to nasis, (kudus, gavaksha dormers.) A second eave is placed above the first, a peculiarity found six centuries later in Hoysala temples, On the upper talas of the west face, (center, below) the five central arches are replaced by long shala aedicules, a reminder that at this early date the canonical sequence of miniature shrines and moldings had yet to gel.
The layering of kapotas may have evolved into the stacks of aedicules on the pagas of later Nagara Latina and Bhumija shikharas. The superstructure’s conjectural step-pyramidal profile, however, has closer affinities with Karnata Dravida vimanas than convex-curved Nagara mulaprasadas. It may derive from the stacked kapota slabs, alternating with galleries lined with arpita (false, implied, incised) shrines, characteristic of the early Kannadiga Kadamba Dynasty (345 – 543 CE.) Their feudatories and later conquerors, the Early or Badami Chalukyas (543 – 754) adapted the single-cell or cella shrines at Bhitargaon and Deogarh by adding an antarala (vestibule.) ardhamandapa (eastern porch) sandhara pradakshina patha (internal circumambulatory) and gudhamandapa (enclosed assembly hall).
Cambodia: Sambor Prei Kuk
The notably unchanging Khmer prasat had already assumed its normative form at the earliest major Cambodian site, the Saivite sanctuary at Sambor Prei Kuk, originally Isanapura, 176km east of Angkor and 200km north of Phnom Penh, dating from the early 7th Century Chenla/ Zenla kingdom, (sometimes referred to as the “First Khmer Empire.”) Its one hundred fifty shrines, symmetrically arranged around three principal enclosures each with two or three terraces prefigure later Khmer “temple mountains” but contrast with the more haphazard clustering of vimanas at the Early Chalukya sacred precinct at Pattadakal (c. 675 – 753) and the Hoysalas’ Vaishnava complex at Belur (1117.)
Most of the prasats at Sambor Prei Kuk are square with only the suggestion of triratha projections around their portals, similar to those at Bhitargaon; some are octagonal which also finds early Indian precedents, for example, the Mundesvari (Mandalesvari) Durga temple at Ramgarh, Bidar (625.) Even at this nascent stage, their doors are surmounted by lintels with garlands and a lobed torana pediment with a replica of the tri-partite shrine itself inside its tympanum.

One of the most distinctive features found on several of these Khmer prasats is the so-called “flying palace” depicted in brick relief on the exterior walls and representing the vahanas, “vehicles” or “mounts,” in which the gods were imagined to commute from Mt. Meru to the terrestrial plane in aircraft patterned after simple, single-celled alpa vimanas, In the photo on the left below, a panjara shrine, (a shala seen end-on,) is depicted like a gondola with four columns and a billowing horseshoe gable, (nasi or gavaksha dormer.) A god (presumably Shiva) and his retinue crowd this upper space, while below Brahma is recognizable from three of his four heads. Above him, two makaras or sea monsters perch on the middle pair of pillars forming an undulating torana or garland of foliage which also suggests a chaitya-griha or Buddhistcave temple’s arched nave and aisles (and, more surprisingly, a Serliana.) Like the portal, these “flying palaces” resemble the prasat itself which thus might be regarded as the palace having landed.
This tri-partite ground-story was then replicated in unitary “aedicules” of diminishing height and breadth, “telescoped” above the garbhagriha to form the talas of a shikhara. What is remarkable is that this simple prasat was the only type of shrine or “shrine aedicule” used with little variation in the immensely more immense “temple mountains” constructed in subsequent centuries during the “Second Khmer Empire.”