A vastu purusha manduka (8 x 8-pada) or paramasayika  (9 x 9-pada) mandala can be generated from a point, as in figure 3, by projecting a line in one direction from the primal dot or bindu, creating the first dimension, then three more in the other cardinal directions, creating an x- and  y-axis, a plane and a second dimension. If each of these four lines then projects a line laterally on either side of it, (that is, at right angles to itself,) at equal distances from the axial crossing, they will form four squares or a single larger square whose center is the axial crossing; It appears that the earliest Hindu shrines were square wattle or mud brick huts with curved thatched or bent bamboo roofs. A model of this type has been preserved, the Draupadi Ratha, figure 3D above, carved from a solid granite outcrop at Mamallapuram (formerly Mahabalipuram) on the Coromandel Coast of Tamil Nadu near Chennai (formerly Madras,) one of five monolithic temple “prototypes” or rathas at this site. These early shrines are also thought to be recalled in the rounded roofs of the chaitya halls of early Buddhist cave temples, such as those at Ajanta in Maharashtra near Mumbai (formerly Bombay.)  A square shrine or cella with a roof gathered at the center like a “cap” placed on top of solid walls or pillars constitutes a Dravida or Southern Indian alpa vimana or kuta, one of the most common shrine “types” in Hindu architecture.

​

Continuing the dynamic of “aedicular expansion,” each of these four squares projects one additional square along each axis to form a Greek or equal-armed cross with eight squares. The four central padas constitute the garbagriha, “womb chamber’, sanctuary or cella, while the added eight form four, 2-pada porches. Since the square module or aedicule from which they emerge is 2 x 2 padas and since they are only 2 x 1 padas, they have emerged only partially, half-way or 50% of their width. The sanctuary and its four porches of Thommanon, figure 3E, a small temple built just before Angkor Wat, approximate a Greek cross. Projecting one more aedicule, not just axially but laterally results in a central square with four rows of four padas each with a superimposed, equal-armed cross, two padas wide and six padas long, producing a staggered or redented, triratha outline, with three rathas or angles in each quadrant. A very large example of a triratha shrine would be Candi Siwa, figure 3F, a Saivite complex at Prambanan in southern Java (c. 850 C.E.,) a date when it could have influenced the earliest temples at Angkor. Adding another square to those above, results in two rectangles, 4 x 6 padas at right angles to each other forming a broad cross on which a narrower, equal-armed cross, 8 x 2 padas is superimposed, as in figure 3G. This can then be expanded laterally to form a manduka mandala of 8 x 8 squares or padas, composed of four concentric squares of 4, 12, 20 and 28 padas around their edges, figure 3H.

​

A paramasayika mandala is employed almost as frequently as a manduka; it begins from a single square (red in the diagram below, cf.  figure 3D) rather than a point, which emanates four squares or porches (blue) forming an equal-armed Greek cross three padas wide. Following the process of aedicular expansion, it adds green, yellow and oranges squares resulting in concentric crosses 3, 5, 7, and 9 padas wide, containing 5, 13, 25, and 41 padas respectively. The last consists of one cross 9 x 1 padas long, superimposed on another cross, 3 x 7 padas long, both superimposed on a 5 x 5-pada square, resulting in a pancharatha shrine, that is, with five rathas (numbered on the diagram,) the most common Khmer shrine. This cross’ forty-one padas can then be expanded into a 9 x 9-pada square with 81 padas, a paramasayiika mandala, consisting of five concentric squares of 1, 8, 16, 24, and 32 padas. Indian architects discovered they could increase the number of facets by using overlapping aedicules, that is, aedicules only partially emerged, ⅛, ¼ and ½ their width (see figure 3E and figure 8.)

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

It should be noted here that the Khmer throughout their nearly 500 year floruit employed essentially the same “aedicule,” actually a L. > “aedes” or building, the ubiquitous pancharatha (with five rathas) prasat  Sans. > castle, palace, temple > Sans. prasada, dwelling, residence (of a god,) temple. This consisted of a garbagriiha, the shrine, sanctuary or cella, with a shikhara, tower or superstructure, of four compressed tiers, each repeating the outline of the shrine beneath it; these qualify as the only strictly defined aedicules used in Khmer architecture. This “aedicule, in the larger sense, was then cloned and projected at unprecedented dimensions. Viewed strictly from the perspective of “aedicular expansion,” the pre-eminent Khmer monument, Angkor Wat, might be described as three, immense, identical tiers or talas, (its terraces,) each with eight identical prasats or “aedicules,” (its towers,) linked by endless harantaras, (its galleries) with the garbagriha or cella,not underneath the talas, but at their tip like a giant stupi. This reductio ad absurdum results from reading one “architectural language” in terms of another, a danger given the lack of a Khmer building lexicon.