7  - Construction

In Cambodia" - Henri Parmentier tells us - "it would seem the construction was but a tedious necessity that was skimped in order to realise as quickly as possible the form which was more or less determined by tradition".

It is a fact that the Khmer, who had specialised for so long in timber architecture - in which they are considerably skilled - showed some delightfully incompetent technique in the art of building in durable materials, ignorant even of the rudiments of stereotomy. Too often the stone blocks were neither squared nor arranged in coursing by the natural lie of the stone. Vertical joints running from bottom to top without any horizontal overlapping, as in the towers of the Bayon, created veritable fault planes. The mass of the large walls was rarely homogenous, the main structure being surfaced in a simply adhered covering that was often quite thin and of a different material. The porticoes or galleries with wide opening bays see the whole weight of the fronton or vaults distributed on long monolithic architraves resting on pillars, which in turn almost invariably cracked under the excessive load.

Everywhere the mistakes and errors are flagrant, with nothing to correct them other than, in certain critical cases, double T form iron straps set into and linking adjacent stones. The excessive corbelling and the mixed use of materials in the construction of the vaults - such as in the 10th/11th century practice of strengthening the sandstone lintels by doubling them with timber beams - also caused extensive structural failure. The stone is constantly used in the manner of timber with the same means of assembly, taking no account of the fact that the material is not able to act effectively in tension.

And yet the ensemble remains despite the ravages of time and climate. All of these faults that trouble us or provoke our reproach are happily tolerated by the Khmer - as Orientals less concerned with shortcomings in detail - with neither eye nor spirit offended, and their general appreciation for the quality of the work certainly unchanged.


The temples of ancient Cambodia are constructed either in sandstone or in brick, often combined in a greater or lesser proportion with laterite.


The Cambodian "thma puok" - literally 'mud stone' - is variable in colour and is, with the exception of the particularly durable rose coloured sandstone used notably at Banteay Srei, a soft stone with little strength. The predominant grey sandstone decomposes and becomes friable under the action of the elements, cracking with the growth of roots and, often laid against the grain, defoliating - it rarely maintains the clear surface and keen line of its decoration and profiles. It weighs from two to two and a half tonnes per cubic metre.

Some large open quarries have been found on the hillside between the temple of Beng Mealea and the south-east extremity of the chain of the Phnom Kulen, at about forty kilometres from Angkor. Transport must have been in part by water, in part by carrying on shoulders or pulling on rollers. The regularly placed round holes of a few centimetres circumference and depth, apparent in most of the stone blocks of the monuments, probably took either wooden pegs tied by vines or metal lugs for a kind of hoist arranged to allow the stone to be manoeuvred during the course of construction - these holes, which legend has it are the impressions of the fingers of Indra, were then filled either with cut stone inserts or with mortar plugs.

Sandstone was initially used sparingly, and almost exclusively for the surrounding elements of openings and false doors, but gradually became used for all the elements of construction - though with the exception of thick internal block walls, utility buildings and certain areas of paving.


Brick was used in all the early structures and then in numerous temples of the first half of the classical period (9th-10th centuries). It was manufactured on site and well baked in order to enable sculpting and to be used in the forming of corbelled vaults. Their size could vary from 22 x 12 x 4 to 30 x 16 x 8.5 centimetres and more. Generally a pale pink in colour they were apparently rarely seen in elevation, having been preferably covered in a sculpted coating of decorative lime based mortar - the brick backing having been previously rough-hewn for the thicker layers.


Laterite or "baļ kriem" - literally 'grilled rice' - is a porous, reddish brown stone that has certain analogies with our mudstone. Abundant in the subsoil of the southern part of the Indo-Chinese peninsula, it is easily cut when it comes out of the ground but then hardens in the open air. Unfortunately some blocks undergo a decomposition, rendering them friable and leading to inevitable collapse.

As a material used for in-fill that can be cut and shaped, laterite was also used in the construction of retaining walls in the tiered temples, for utility buildings, the piers of bridges, enclosure walls and for the paving of courtyards.


From some of the most durable of species, timber served even in the monumental architecture of the classic period for the building of certain external light-weight elements in combination with the stone.

Internally, timber was used in beams either for supporting or reinforcing, for the roof carpentry, for the double-leafed pivoting doors - whose pivot socket-holes can often still be seen in the door cills - for the dais sheltering the idols, or for the richly decorated panelling of the walls and ceilings - some remnants of which, decorated with flowers and lotus blossoms (though deteriorated by the humidity and the termites) were still in place in Angkor Wat at the time of its clearing. Fragments of beams also remained, as they do in many of the other monuments.


The roof tiles of the annexe buildings are excellent in quality, and numerous specimens have been found during excavation. In baked clay and either plain or glazed and with a fixing nib, they are of two types; - the first with a flat edge to form channels, the second curved to form joints enabling a roof covering of so-called "Roman tiles". The ridge can be marked with a line of turned stone finials and, at the base of each slope, the stop tiles can also curl in the form of lotus petals or some other decorative motif.



The angkor monuments where generally constructed on a firm ground of clay-bound sand, so reducing t he foundations to their simplest expression - one or two courses of laterite laid sometimes on a bed of consolidated gravel. There was little settlement, except where there was banking.


The base platforms are a common feature. Often formed in combination, they can be crowned by a simple band when built as retaining walls for the tiers of a pyramid, or abundantly moulded and ornate as the base platform for terraces that may carry some other structure. They can be one of the most remarkable elements of the architecture.

The Khmer substructure has the peculiarity that it remains independent of the movement in vertical expansion of the building it supports. It is a base, a plateau emerging like the very mound of the celestial Meru - it is the horizontal component in the composition. This is reinforced by the moulding with its horizontal axis of symmetry, expressed as a central band between two opposing ogees. The symmetry is reflected in the smallest detail of the ornamentation, where only the lines of lotus petals invariably turn upwards.


Whether in sandstone, brick or laterite, the walls have dry joints without mortar. For the brickwork, only a kind of vegetal adhesive, of unknown composition, serves to reinforce the bonding.

In an architecture where the sequences of moulding and sculpture are carried out on an in-situ masonry structure, it is important to obtain as near as possible a monolith by the perfect adherence of the beds and vertical joints, rigorously dressed and made filiform. This was achieved through polishing each block by rubbing it against the surrounding stones in its immediate contact - a bas-relief at the Bayon (the internal gallery, west side, southern part) gives a precious indication of this operation.

The wall thickness is essentially variable but always far in excess of the limits imposed by the strength of the material - depths of one metre to a metre and a half are not uncommon, and it is not unusual for enclosure walls to be built with the length of the block laid perpendicular to the face of the wall

And while it is true that the same wall, plumb from top to bottom on its internal side, can correspond externally to steps corresponding to various false elements, this feature can also simply relate to a juxtaposition of cladding on course blockwork that has lost cohesion.

It is interesting to note that the door frames, set in main or internal walls and treated with an assembly of mitred or straight joints - as in timber - always have their bottom member set above the level of the paving. The existence of these high cills, which can make a visit to the temple quite tiring, must be intended to accentuate the cellular character of the space so enclosed in order to increase the number of sanctuaries by compartmenting the galleries to the extreme, rather than for any technical necessity.


The tiered temple is like a "stairway to heaven" - which is perhaps sufficient to justify the steep incline that can be set from an angle of 45° to 70° - unless otherwise the stone steps are simply a replica of the wooden stairs that give access to the timber houses, where the absence of a riser allows the foot to be placed despite the gradient.

Whatever the reason, the respective dimensions of the steps and the risers are the inverse of those to which we are accustomed, and the arrangement - where the stairway is presented from the front and generally set into the substructure without intermediate landings - transforms the ascension into a veritable climb, confirming that it was not intended for the advance of a crowd but rather only for use by certain officiants. From the monumental point of view the advantage is clear - the square of the base not having to spread in surface area, the entire building rises to its zenith with a particular thrust.


The problem of the vault conditions a characteristic of the Khmer temple - and indeed of all religious architecture of Hindu inspiration - which is the absence of any large internal space - a disadvantage since there is no place to shelter any assembly of the religious faithful.

Only a keyed arch will enable large openings. Known since antiquity in the West, it was used as far as China, so seeming strange that the Khmer of the 9th to the 13th centuries were ignorant of it - despite their use of radial joints in the lining of circular wells, as, for example, at the western Mebon. Perhaps it is necessary to see some ritual reasoning for this abstention, or a respect for the Hindu saying in which we are told by Henri Parmentier that - "the dressed arch has no rest, only the corbelled arch sleeps..."

The Khmer vault does not transfer its thrust to the points of its support, so that no reaction is provided by its elements - it is formed by a continuation of the walls which overhang until they meet on the axis of the covered space. The beds are therefore horizontal and the elements successively corbelled and finally topped with a bridging stone, so linking the two walls.

The sloping inner face, usually following the line of a slender cone, is left rough when masked by a timber ceiling at the height of its springing. However, when remaining in view, it is carefully finished and can be decorated, particularly on the half vaults of the gallery side-aisles. The outer face is much smoother and almost semicircular in profile, with its curve serving as a template for the mass of the fronton.

In the cruciform-planned buildings the intersection of the two barrels is normally formed as a groined vault, while for the square-planned "prasat" the principle of the cloistered or coved vault applies, though often interrupted by some vertical elements corresponding to the projections of the illusory external storeys.


We do not know what means the Khmer had at their disposal for the construction of their temples. The bas-reliefs give only some cursory indication of the methods used for polishing the stone blocks, but show no lifting apparatus. We are therefore reduced to hypotheses.

Present day Cambodian labourers still rely on the building methods of their ancestors. Highly skilled at erecting a sturdy scaffolding with a few simple timbers cut from the forest and tied with lengths of vine, they can lift the heaviest of loads. Using their climbing ability to good advantage, they carry them on their shoulders suspended from two sticks or bamboo poles, or haul them to great heights on rails of logs. One can therefore suppose that similar methods of lifting were previously used, with ladders or inclined planes perhaps aided by winches or capstans.

George Groslier undertook some interesting research into the time required for the construction of a large temple in the north-west of Cambodia - Banteay Chmar. His calculations, based on reasoning and logic rather than on actual fact, led him to conclude a construction period of 32 to 35 years. We would tend towards this latter figure, which noticeably corresponds to the duration of the reign of Suryavarman II, the builder of Angkor Wat. The uniform style of this monument enabled the assumption that it was built without interruption under a single direction.

On the other hand, Groslier's thesis provides a strong argument against the attribution to the single king Jayavarman VII, who reigned for some 20 years, for the totality of the temples in the so called "style of the Bayon", where there is abundant proof of alterations and which particularly lack unity.