Pillar base architecture in ancient Kedah
B.A.V. Peacock’s 1974 article, published in the Journal of the Malaysian Branch of the Royal Asiatic Society, challenges the prevailing reconstruction of Kedah Site No. 8 (Chandi Bukit Batu Pahat) as a massive stone sanctuary tower, arguing instead that the building was a low-walled sanctuary supporting a multi-tiered wooden roof of the meru type. Working from a close study of the pillar bases preserved in situ at the site and drawing on comparative evidence from Sumatra, Bali, and East Javanese temple reliefs, Peacock demonstrates that the structural and foundation evidence is incompatible with the stone tower hypothesis proposed by H.A. Lamb in 1960.
Summary
The ancient Hindu-Buddhist settlement at Bujang Valley, Kedah, has been studied for over a century, yet the near-total absence of superstructural remains has left the architectural form of its temples largely unresolved. Peacock takes as his starting point the reconstruction of Site No. 8 by Dr. H.A. Lamb, who proposed a stone sanctuary tower of more than thirty feet rising from the square basement, of the prasada type. Peacock identifies several interlocking problems with this interpretation. The curvilinear stone block Lamb identified as a roof finial was not recovered by the excavators but was reported by a householder who claimed to have removed it from the site, leaving its provenance uncertain. More critically, when Lamb dismantled the sanctuary basement for reconstruction, he found it to be a masonry shell enclosing rubble and earth rather than solid piers capable of bearing the downward thrust of a massive stone superstructure. The reconstructors themselves felt compelled to add masonry piers not present in the original to carry even the weight of the lower wall courses.
Peacock’s alternative interpretation rests on the arrangement of the pillar bases on the sanctuary basement. An outer square of smaller pillars (mortises 4” × 4” × 3”) supported the lower roof timbers, while an inner square of substantially larger pillars (mortises 7” × 7” × 3”) was positioned close against the sanctuary walls. Eight specialized bases (Lamb’s type D) carried angled mortises indicating a system of cross-struts sloping from the base of each corner pillar to the top of its neighbours—a bracing system that would be structurally purposeless if a solid stone tower occupied the interior. The battering of the sanctuary walls, clearly visible in the stepping of the lower courses, would have separated these inner pillars from the masonry by a gap of two to three feet, further undermining the tower hypothesis.
Peacock supports his meru interpretation with comparative parallels. The sanctuary of Biaro Si Topajan at Padang Lawas, Sumatra, shows stone pillar bases standing directly on low sanctuary walls, conclusively precluding any masonry rise above that point. Kedah Site No. 15, a small laterite sanctuary, similarly places pillar bases on the inner edges of low walls. East Javanese temple reliefs, as studied by Galestin (1936), document widespread use of tiered roofs on small wooden pillared structures, while the surviving Balinese meru demonstrates the dramatic effect achievable with thatched tiered roofing. The rarity of ceramic tiles at Kedah sites points to the use of perishable roofing materials.
Key Findings
- The sanctuary basement at Site No. 8 was a masonry shell enclosing rubble and earth, not solid piers; Lamb’s reconstruction required the manufacture of at least 8,000 new granite blocks to replace originals too decayed for reuse (p. 72).
- Outer pillar mortises measured 4” × 4” × 3” (implying pillars of 6–7 inches square); inner pillar mortises measured 7” × 7” × 3” (implying pillars of 9–10 inches square) (p. 81).
- The inner pillar bases alone contained circular depressions 1½” in diameter at the bottom of the mortises, which held foundation deposits of precious stones in small silver capsules (p. 81).
- Eight type D pillar bases with angled mortises indicate a system of cross-struts sloping from the base of each corner pillar to the top of adjacent pillars, a bracing system designed to resist lateral wind stresses on a tall multi-storied roof (pp. 72, 81).
- The mandapa platform carried six rows of five pillars each, suggesting a separate double roof structurally distinct from the sanctuary roof (p. 85).
- Peacock proposes an eleven-tiered roof for the sanctuary, the number ritually prescribed for a temple dedicated to Siva, whose worship is strongly indicated by the ritual associations of the site (p. 81).
Conclusion
Peacock concludes that Site No. 8 was a meru-type structure whose monumental construction, associated terraces, enclosure wall, and elevated position on a spur of Gunong Jerai overlooking the Bujang valley and Merbok estuary invite comparison with Pura Besakih in Bali. The temple likely functioned as the temporary abode of ancestral spirits summoned from the mountain during seasonal festivals, serving as a “state temple” and place of pilgrimage for the inhabitants of the settlement below, and underscoring the pre-Indianized megalithic associations previously noted by Lamb.
Context
- Peacock’s argument is grounded in fieldwork conducted in Kedah in 1972, supported by the Ford Foundation, and in direct examination of the pillar bases in situ at Site No. 8 prior to and during Lamb’s reconstruction (p. 86).
- The article represents a significant historiographical intervention, being the first sustained critical challenge to Lamb’s tower hypothesis, which had dominated interpretation of the Kedah settlement since 1960 and had passed without critical comment into popular literature (pp. 70–71).