Next to Angkor Wat the Bayon is the most widely recognized Khmer site, if for no other reason than the mystery of its "face towers" and their enigmatic smiles. Unlike many of the preceding "temple mountains," the monument's present state provides among the least accurate impressions of its makers' intents and its original appearance. This photograph of the southern and western ranges of the 1st gallery and terrace, containing the temple's most famous bas relief, "The Battle on the Lake," lacks most of its external colonnade and roof, including its gopuras, cannibalized as a quarry for later projects. As a result, we are denied a satisfying view of the temple's all-important elevation, its rise from the terrestrial plain to its spiritual pinnacle of which today only shards of a shikhara remain. We know the central shrine and tower were cylindrical, more precisely conic, one of the Bayon's many unusual features, which in its decayed state appears jagged and irregular, a more naturalistic impression of a mountain than its architects probably desired for a representation of Mt. Meru, the Platonic form of mountain.  In addition, the omnipresent face towers, placed so deliberately and regularly in a dense array, as detailed in figures 23A, B and C of the introduction, are barely discernible today even from this close a range; many have been lost and the rest appear haphazard, peeping coyly from their masonry, quite the opposite of their original effect.

The Bayon is the only "temple mountain" at Angkor clearly to have been constructed in separate stages which contain anomalies suggesting that its plan may have changed or expanded over its three building phases. If so, an understanding of the temple's architectural history may provide insights into its unprecedented design and its significance in Khmer society at a time of critical change and renewal. The site plan at right, color-coded to reveal each stage in the Bayon’s construction, as determined by recent radiocarbon dating and spectroscopic analysis of the stone used in different parts of the temple, is discussed in detail under figure 22 in section XIV of the introduction.

SOUTHERN SECTION, EAST FACE, 1ST GALLERY, BAYON (1181-1220)

SOUTHERN SECTION, EAST FACE, 1ST GALLERY, BAYON (1181-1220)

Such stylistic changes seem consistent with the objectives of Jayavarman VIII's long reign (1243-1295) the restoration of legitimacy to Hinduism and Khmer traditions in the face of the unavoidable skepticism following the civil war, the Cham conquest, Jayavarman VII’s challenges to tradition and the failure of his innovations to arrest the Empire’s decline. Jayavarman VIII did not rule until his death but abdicated, probably under pressure from his court, following a string of humiliating defeats by the Mongols under Kublai Khan in 1283, the always restive Mon minority and the first Thai kingdom, centered at Sukhothai. Its successor Ayutthuya would occupy Angkor twice in the coming century, eventually leading to its abandonment in 1431, while Theravada Buddhism would replace Hinduism as Cambodia’s permanent creed, presumably one more suited to a period of declining expectations since its highest goal was non-existence.

Characteristic scenes depicted in the 2nd gallery include: a)  kings and soldiers departing on military expeditions; b) women of the court amusing themselves, waiting their uncertain return; c) Shiva Bhikshatana, living as a forest hermit among ascetics and wild animals;  

d) Vishnu venerated in a temple by prostrate royal suppliants; e) legends of the loss and miraculous restoration of the prerogatives of noble birth, common in every feudal society, but especially timely for a period of contested legitimacy. In the Bayon’s 2nd gallery, this theme is represented by Pradyumna, one of the four vyuhas, avatars or aspects of Vishnu and Krishna's son, who, abducted by a demon, is thrown into the sea, swallowed by a fish, miraculously rescued, raised by a servant and eventually kills his adversary, assuming his rightful royal destiny (a theme as old as Moses in the bullrushes and  Oedipus.) f) Finally, there is the unexplained legend of the "leper king," (mistakenly attributed to a statue on the nearby terrace of that name,) who slays a serpent which has been terrorizing his kingdom but contracts the disease from his bite. Since this incident is not mentioned prior to the 13th Century, it is thought it must refer to a contemporary sovereign, either Indravarman II (1220-1245) or Jayavaraman VIII (1243-1295.) Since there is no epigraphic evidence to confirm this, it might  be a metaphor for a monarch “sacrificing” himself for his subjects, only to become a pariah driven from his throne, in other words “autobiography”.

often appear together, as if to fuse (or confuse) their identities, a venerable trope in Khmer art stretching back to pre-Angkorian statues of Harihara, such as the outstanding example from Prasat Andet (c.675.) 

In the lower register, two ladies of rank, in identical boats with “double, staggered shala” roofs, their panjara  openings cleverly everted into blossoms, are rowed towards a lotus in the middle of a lake above which three apsaras, like the three graces, bounce back and forth in the peculiarly acrobatic dance step adopted at the Bayon; their string-like arms are tangled in similarly carved lotus stems. The lady on the right clutches an aryballos, usually associated with amrita/amrta, the elixir of immortality, the most important ratna or jewel from “The Churning of the Ocean of Milk” along with apsaras, perhaps providing another Vaishnavite allusion. The co-existence of asceticism with apsaras and immortality was not contradictory for the gods of the Vedic pantheon who were granted human desires; it was Buddha who "stood Hinduism on its head.”

The apsara is probably the most pervasive and characteristic motif in Khmer art, certainly one of its most ingratiating. With the passage of years and the growing confidence of their empire, these "celestial musicians" or "dancers" gradually replaced the apotropaic kalas, the ravenous kirtimukhas, nagas and makaras, lurking in the earth and water always lapping at the Khmer lands, lodged deep inside the collective reptilian brain. The hybrid narasimha, garuda, and bestial asuras constantly threatening rebellion, even the Vedic gods with their fickle limbic rages and affections, their cycles of destruction and creation, the ceaseless mayhem of the feudal Hindu epics, may have too closely reflected daily Khmer life. Apsaras were, above all, “above it all” – untethered from gravity, weightless, dancing on air, no longer conceived as mortal, court dances but self-delighting denizens of the Buddhist rupadhatus or "form worlds," free from the insatiable desire for external, cataleptic pleasure which drove the gods and humans of the kamadhatu.

 Apsaras are Khmer sisters of Greek "aurai," "wind nymphs" or “breezes” who could be heard and felt but not seen, their billowing robes indistinguishable from clouds; like apsaras, they too were nikes or "winged-victories," singing paeans over heroes at the moment of their short-lived triumphs. They might also be seen as the Khmer equivalent of angels except without a god to send them on errands announcing his admonitions. They may represent the aesthete’s dream of any refined court, a life sans souci, whether in Frederick’s Charlottenburg, Heian Kyoto or Renaissance Urbino, of a life where taste could be cultivated as its own end, relieved of the bitter struggle needed to sustain it. This idyll is also glimpsed in the hazy intimations of the heavens (perhaps never fully realized or realizable) on the eastern section of Angkor Wat’s southern gallery, above the hells which remain as vivid today, as the world we live in.  

The ideal of bliss necessarily changed between the reigns of Survavarman II, the "post- traumatic stress" from which Jayavarman VII's empire was recovering and the sense of irreversible decline under Jayavaraman VIII. In Angkor Wat's 2nd gallery apasras are portrayed, not so much dancing as swaying, poised gracefully, possessing the peculiar Khmer quality of sensuality without sexuality, a charm which may lie in their self-possession and indifference to the viewer. In contrast the thousands of apsaras which decorated Jayavarman VII's foundations can seem like gyrating cheerleaders, springing gymnastically, even manically, from foot to foot, like puppets jerked by strings or wind-up toys, knowing just one step, their legs and arms always flexed at the same acute angle like a chorus line. Perhaps their exaggerated peppiness can be excused as a response to that monarch’s snatching victory from the jaws of defeat, however brief that renewal may have been but they seem to try too hard. The characteristic S-curves of Angkor Wat's apsaras from the skirt across their bodies to the lotus stalks they hold, stiffen at the Bayon into angular zig-zags, arms and hands flailing in all directions a far cry from the subtle mudras of classical Bharatanatyam or, for that matter, the kbach of today's ersatz Apsara Dance. Not too much need be made of the noticeable decline in artisanship at the Bayon; as in most matters, the quality of sculpture seems to have been inversely related to its quantity, victims of Jayavarman's VII's "excessive (architectural) exuberance" and Jayavarman’s VIII’s strident Hindu fundamentalist propaganda, a Narendra Modi for his times.

AURA, AGORA MUSEUM, ATHENS 

(4TH CENTURY B.C.E.)

The Bayon’s conic, central tower was initially a standard Khmer prasat, a square sanctuary with four porches forming a Greek cross. While still in the 1st construction phrase, second thoughts converted this into a circular cella ringed by eight rectangular sub-shrines with eight, interstitial triangular niches, forming a 16-sided 1st story unprecedented in Angkorian architecture. A portico surrounded the base of this faceted tower, (reminiscent of Bramante's Tempietto, itself inspired by the Temple of Vesta and Tholos at Delphi) making it circular. It has largely vanished along with a large part of the upper tiers and finial of the superstructure which rose above it. Just as compressed copies or aedicular stories of a Khmer prasat were layered above it to form the tiers of its shikhara, at the Bayon the eight rectangular sub-shrines were each stacked with versions of themselves, in what Adam Hardy has characterized as “piling up” or clumping aedicules together around a core to make a superstructure. These aedicular shrines at the Bayon protruded more distinctly than the terrace sub-shrines because the triangular niches did not extend up the tower’s face, leaving acute angle recesses between the vertical strips and horizontal rows (actually segmented circles) of similarly jutting rectangular masonry. This is probably the closest Khmer architecture comes to the process described by Hardy as “aedicular expansion,” illustrated in section II of the introduction to this catalog. 

While the talas or tiers of Karnata Dravida temples were lined with ingenious variations on a traditional repertory of shrine types (kuta, shala, panjara,) the Bayon, as in so many ways, invented its own aedicules without clear precedents. For example, the shrines on the Bayon’s 2nd tala, have pilasters, lintels and pediments like those of the 1st but are turned into dormer windows with three baluster columns in antis screening some interior mystery, (evoking the enigmatic enfilade of chambers along the  processional path to the temple’s central shrine.) In India, gods and goddesses would poke their heads over these saedicule’s  vedika/vedi, (railing, hbalustrades or  sills,) stare from inside their gala recesses or give darshan from their gavaksha’s gadhas (windows.) On the Bayon’s 3rd tala, in contrast, heads crowned by a miter and amalaka burst through the shrine’s enclosed blocks, becoming “face towers,” inverting the relationship between a shrine and its contents. Two are visible in this photo on the tower’s cardinal axes; (a third looks down from an adjacent tower, in the upper right corner.) 

The baluster windows and “face towers” above the intercardinal shrines have collapsed but enough of their masonry still projects to suggest that originally there were eight stacks of aedicules around the tower. These suggest bhadra, (the raha or central rathas or paga,) vertical strips of similar aedicules (often panjaras, known as nasi, ) which unify the horizontal tiers (talas) of many Dravida (Southern Indian) vimanas; their nearest Khmer analog would be the cascade of pediments down the shikharas of temples such as Phimai or Phnom Rung. In the context of Karnataka and Tamil Nadu temple architecture, the Bayon might be classified as asta-bhadra, with eight bhadra and stellate, a square rotated once by 45° with principal projections, points or rays  at its cardinal and intercardinal points, a common configuration ofor Hoysala temples in the 12th and 13th Centuries. 

(A ) above, the  rings of shrines on the Bayon’s shikhara are so dilapidated it is impossible to tell how it might have been completed – which need not and has not stopped speculation. A brief review of other conic towers might suggest possible solutions. Rigorous trigonometry would demand something like (B,) Norman Foster's 30 St. Mary's Axe in London, known, not necessarily fondly, as "The Gherkin," or Jean Nouvel's Agbar Tower in Barcelona.  Khmer architects could have found models for geometric towers closer to hand – the tall spires of the two "temple mountains" immediately to the Bayon’s north, the Baphuon and Phimeanakas (now both lost.) Such a finial, however, would have shifted the towers distinctive convex curve to concave, although this was no impediment at the early conic pagodas at Bagan, Sri Ksetra and Anuradhapura; for example,  (C) the Bupaya Phya, Bagan (9th Century,) reversed course for its concave tip. Eventually, this evolved into the convex-concave, bell-shaped pagoda which became the norm for Buddhist chedi or zedi in Southeast Asia, such as the celebrated Shwedagon in Yangon (1362.)

The Bayon’s shikhara, however, seems to have been generated more by what Hardy calls “expanding repetition” where the aedicules of each tala increase in size as they move down the tower which would render a smooth parabolic 

curve impossible. A closer analogy might be the Bagan temple or gu/gyu, (cf Sans.> gudi/gudhi temple,) a ground floor shrine followed by aedicular tiers with miniature shrines, as with a Khmer prasat, crowned by a convex, “corncob” finial or shikhara, itself striated into multiple, schematic talas, for example, (D) the Sein-Nyet-Ama, (12th Century.) In Lanna, with its strong Dvaravati influence, temples were sometimes constructed out of circular or rectangular tiers of individual shrine aedicules, for example, (E) Wat Puac Hong, Chiang Mai (16th Century) and Wat Ku Kuk (Chamthewi,) Lamphum (1218.) Tibetan kumbums were multi-story, step pyramids composed of chapels or shrines, somewhat like pueblos or the jagamohana roofs but surmounted by a cylinder, a square harmika and cone; the most famous may be the pancharatha one (F) at Gyantse (1427.) 

The 19th Century model at the beginning of these notes on the Bayon probably came closest, informed, as it was, by archaeological data, however limited. It proposed four aedicular tiers of alternating shrines and “face towers” arranged in eight vertical stacks like the stambhas of a Nagara bhumija temple,topped ingeniously by a single tower with four faces and a four-tier miter – an aedicule of the shikhara it crowned The plan implicitly recognized that as a cone rises its sections 

BUDDHA, BAYON (1181-1220)

​

The identity of the Bayon's "face towers" continues to be a source of fascination and folly in direct proportion to how little can be said with certainty about them. Their size and number taunt viewers that the answer must be as obvious as they are – literally staring them in the face. The fact that each new study expands rather than narrows the candidates, rather than discourage speculation, has only made the dubious seem more plausible. Similarly, descriptions of the expression on the faces have become a Rorschach test of what people think they should be thinking or what they might be thinking about them. Like the Mona Lisa their "enigmatic smiles" convince us that they must know something we don’t know, simply because we don't know what it could be. Could the smile possibly be a "smirk" mocking the assumption that a stone block could harbor one correct meaning, that we could know it after eight centuries or ever agree on it? This photographic overview of the Bayon will not add to this speculative prolixity; (the introduction to this catalog has already done more than its fair share of that.) 

​

One thing seems undeniable; the identity of the "face towers" shares a characteristic Khmer vagueness not to say evasiveness on doctrinal issues, already noted in reference to the devaraja cult, the easy co-existence of opposing sects, its eclectic or syncretic iconography and the striking epigraphic silence about specific religious practices. Asked to decide between alternatives, the Khmer seem to have preferred to choose both, eliding differences and accepting dichotomies as different manifestations of the same divinity, venerated in whichever aspect appealed at the moment. Such post-modern relativism, of course, suffers from a "presentism" as “imbricating,” as blinkered, un-self-critical  and perhaps unavoidable, as every other hermeneutic. 

​

Although it seems self-evident that the answer to this conundrum must be implicit in the number and placement of the “face towers,” it has proved impossible even to determine how many there were and hence their pattern. The discussion of the positioning of these towers in this catalog’s introduction began with what appeared a modest goal: describing rather than explaining the arrangement of the conjectural original array of 52 towers. But even this seemingly straight-forward task could not avoid imposing some, hypothetical “idea of order,” indeed three different organizing principles, graphed on figures 23A, 23B and 23 C at right, displayed full-size and analyzed in section XIV of the introduction. Figure 23A attempts to generate the array through the process of “cruciform expansion” (axial and lateral projection) described in section XI of the introduction on the Baphuon, producing an orthogonal pattern which accounts for the majority but not all towers. Figure 23B attempts to catch the rest in a net of diagonals cast across that orthogonal grid. Finally, figure 23C looks for a countervailing, radial impulse rippling out from the central tower and its circle of sixteen shrines and niches — though the most it finds are some broken arcs and semi-circles, fractures more than rotation. The analysis manages to evade integrating these disparate arrays by suggesting they are intended to frustrate any single reading of the towers' arrangement, thus producing a sense of disorientation. Such a transparent feint says more about the author’s own bewilderment than the intentions of the Bayon's sthapakas, an appropriate ending for these stray notes on the Bayon. 

“FACE TOWERS,” 3RD TERRACE, BAYON (1181 - 1220)