International Journal of Tantric Studies

Sexual Imagery on the "Phantasmagorical Castles" at Khajuraho

by Michael Rabe

Setting the stage.

Plate 1 For one glorious century starting just before the year 1000 Vikrama (A.D. 943), the political center of North India's Imperial formation1 shifted from Kanauj in the Yamuna-Gangetic Doab (Mesopotamia), southwards to Khajuraho in the heavily-jungled uplands the Vindhyas. There, a succession of five Candella kings presided over the final, but also finest efflorescence of the NAgara-order temple, an architectural style whose origins are can be traced to Gupta-period prototypes in the same Bundelkand region,2 and whose extinction as a vibrant creative tradition was perpetrated by iconoclastic Islam. Yet as Fate would have it (personified in the Zri-LakSmI consort of the Candella dynasty's brief Imperium),3 geographic remoteness insured the relatively intact survival of a dozen masterpieces of NAgara temple architecture at Khajuraho.4 Apparently, they escaped notice of even SultAn MahmUd himself, who lay futile siege to the impregnable Candella fortress at Kalanjar, not 50 miles away in AD 1022, on his 14th predation out of Ghazni, Afghanistan.5 It is true that in 1342 Ibn Battuta saw evidence of muslim mutilations, on the consecrated images at Khajuraho,6 and the largest Ziva-linga in North India, still in worship in the MAtangezvara, suffers the indignity of two Persian inscriptions.7 But otherwise, especially in their overwhelmingly figural outer fabrics [e.g., Plate 1], these temples survived centuries of neglect (following the final collapse of Candella power in 1308) with remarkable integrity, and soon after initial rediscovery by an Englishman in 1838,8 concerted efforts to restore them were undertaken, first by the local Chatarpur rajas, and since by the Archaeological Survey of India.

Returning to consideration of the site in the year 1000 Vikrama, it must have seemed a most propitious time for a hitherto insignificant tributary of the Imperial PratihAras to concretize dynastic aspirations for genuine sovereignty by the erection of a magnificent temple. As Ron Inden has noted in his brilliant reconstruction of the Deccan-based imperial formation of the Rashtrakutas,9 the grand finale of a tributary king's metamorphosis to overlordship in this period, equivalent to the cakravartin's horse sacrifice of earlier times, was often the construction of major temple. Unfortunately, the Candella's earliest royal temple dedication is lost--all but a roughly 1'4" square fragment of what must have been an approximately 2' x 5' slab.10 Because the fragment was found near the VAmana, that temple may have been the one dedicated therein by the Candella king HarSadeva,11 presumably after wresting the fabled Kalanjar fortress12 back from the Rashtrakuta invader, Indra III and restoring his feudal liege MahIpAla to the PratihAra throne in Kanauj. However, since, atypically, its only overtly sexual imagery appears in unpublished subsidiary niches of the roof-pediments,13 it can receive no further comment here.

Plate 2 HarSadeva's son, Yazovarman, expanded the boundaries of Candella authority mostly at the expense of their Kalachuri rivals to the south and the imperial PratihAras, even forcing MahIpAla's successor, DevapAla, to relinquish a priceless palladium of VaiSNava kingship, a VaikuNTha-ViSNu icon, and built the suitably magnificent LakSman temple to house it. The icon, almost certainly gold,14 and originating in Kashmir,15 has long since disappeared, but an Candella-style facsimile in sandstone survives in its sanctum.16 The impressively long dedicatory inscription (in 49 verses and a prose colophon), detailing Yazovarman's ancestry and conquests is actually a posthumous record, dating from the year Vikrama 1011 (A.D. 954/5) in his son Dhanga's reign but any patron-specific references in the temple proper (or benefits accruing from its construction) clearly belong to the father. Rarely noted evidence in support of this contention is that the father, Yazovarman is also titled LakSavarman (vv. 37, 39), whence derives the temple's traditional name in a slightly garbled form.17 Thus, precisely because it is the earliest dated temple at the site, any attempt to make sense of the sexual imagery at Khajuraho must concentrate first and foremost upon the LakSman complex [e.g., Plate 2, detail of 1], completed in 954 with 4 subsidiary chapels and a shared masonry terrace around which the most notorious obscenities at the site appear.

Plate 3 Before cutting to the nitty gritty, three other temples need introduction by virtue of association with inscriptions (two of them dated), that name individual patrons. First the PArzvanAth, largest of the Jain temples on the east edge of Khajuraho [Plate 3], must be nearly contemporary with the LakSman--in proximate sculpture style, together they represent the finest phase of the Candella idiom [Plate 4]. The PArzvanAth also bears a brief inscription dated to the same year, Vikrama 1011. Though recopied in 13th c. characters, it enumerates a gift of seven gardens to the temple on a day equivalent to Monday April 2, 955 A.D., given by PAhilla, presumably a minister of state, "held in honor by king Dhanga."18 Much is often made about the relative discreetness of sexual imagery on Jain temples, but the word relative needs emphasis, where even goddess LakSmi is not immune to ViSNu's fondling, and maithuna couples actually coupling appear in the ambulatory within!

King Dhangadeva's own panegyric (prazasti), dated equivalent to 1002, was found in rubble at the base of a Ziva temple now known as the VizvanAth, and is presumed to be its dedication.19 This temple too, like the LakSman, was built on an imperial scale, with subsidiary chapels at the four corners of its high masonry terrace (just one of which survives)20. Its primary icon, rivaling in pedigree his father's golden Vaikuntha, was an emerald Ziva-linga, called the Maraketzvara, descended it was said from Indra's heaven.21 Coincidentally as well, its even longer dedication (in 64 verses) is also posthumous: though explicitly crediting Dhanga with construction of the temple, it proceeds in v. 55 to commemorate his voluntary suicide by drowning in the sangam at PrayAg: Plate 4

Having lived for over a hundred years and protected this earth enclosed by the girdle of seas as its undisputed sovereign, the celebrated king Dhanga obtained liberation by abandoning this mortal frame at the confluence of the GangA and the YamunA, with his eyes closed and mind concentrated on and reciting the name of Rudra Ziva.22

One wonders if any wives accompanied him in aquatic satI as did the 100 of king GAngeyadeva, of the rival Kalachuri dynasty, at the same holy confluence some 40 years later.23 Whether or not in partial reference to his wives, female figures certainly predominate in the iconography of Dhanga's VizvanAth [Plate 5], as indeed they do in all the major temples at Khajuraho. The speculation is hardly idle, for dramatic allusion to captive women in Dhanga's train is explicit grist for his panegyrist's mill, engraved as fair notice on this very temple:

v. 46.

Who are you?
Wife of the king of KAncI.
Who are you?
Spouse of the chief of RADhA.
Who are you?
Wife of the king of Anga.

Such conversations took place amongst the imprisoned wives, whose lotus-eyes were wet with tears, of his enemies over whom he had gained victory.

v. 47.

Who and whose are you? What for have you, resplendent as the moon, come here? I am the accomplished fame [KIrttir-ahaM] of the monarch Dhanga, the sole friend of the learned. After wandering over the whole of (the) universe, I have come, with whetted curiosity, to behold the beauty of the peaks of the high mountain ranges of LokAloka.24

Plate 5 Ironically, though a translation of this inscription was included in the earliest publication on Khajuraho,25 these verses have never previously been cited as relevant to the site's iconography. Yet undeniably they are--cut literally from identical sandstone and bonded as integral units. Thus, with the admitted proviso that this illustrative figure [Plate 6, another detail from Plate 1] belongs to the earlier king Yazovarman's LakZman temple, is it not fitting to imagine one such as her being the visual analogue to the poet YazapAla's KIrtti--singing yes, but also embodying, a Candella king's Fame? Reunited at long last, in cyber.

Plate 6 The last temple to be introducted here is also the largest26 and most perfectly proportioned [Plate 7], the myriad-peaked sine qua non of NAgara architecture, called the KandAriyA MahAdeo. Absent a more definitive etymology of the famous name I parse it as poetic reference to Ziva as Great Lord of the Cavernous Peaks.27 Such a name is certainly appropriate, given its wonderful clustering of uruzRnga replications, like foothills and lesser peaks, surrounding its tautly soaring primary zikhara. While, unfortunately, no dedicatory prazasti survives for the KandAriyA, the mountain metaphor, so recurrent in Indian architecture and perfectly captured here, is proclaimed in the Yazovarman's earlier inscription for the LakSman:

v. 42.

He erected this charming splendid home...(of Visnu) which rivals the peaks of the mountain of snow; the golden pinnacles of which illumine the the sight of which the inhabitants of heaven, met together on festivals, filled with increasing delight, are struck with wonder.28

There is, however, a tantalizing single-line epigraph on a pillar of the interior which, according to Krishna Deva, corroborates the attribution on stylistic grounds of the KandAriyA MahAdeo to Dhanga's grandson, VidyAdhara (c.1018-1022):

rAjaVirindasamaye navasura-samAgame varastrInAM

Of the VarastrIs on (the occasion of the) Navasura-samAgama in the time of king Virinda29

Plate 7 While granting that Virinda may be an affectionate pet name for VidyAdhara, it is less certain who exactly these varastrIs or choice women might have been,30 --royal wives as noted earlier, or commoner-accessible dancers called devadAsIs? The great Arab historian al-Biruni, who accompanied MahmUd of Ghazni to the siege of Kalanjar, reported matter of factly that Hindu kings defrayed military budgets with revenues skimmed from temple prostitution,31 but Khajuraho, never a large metropolis and lying considerably off beaten pilgrimage routes through religious centers like Mathura, PrayAg and Benares, was more likely the private preserve of Candella kings and their courtiers.32 As for the ceremonial context named in the KandAriyA epigraph, one possible scenario for imagining what navasura-samAgame, "welcoming the new god" (?), might have entailed, is preserved in KalhaNa's history of Kashmir, wherein a slightly later king HarSa (1089-1101) was plied with slave girl goddesses by his courtiers:

Others brought slave girls before him and said they were goddesses. He worshipped them, and abandoning his exalted position and wealth was laughed at by the people.

Plate 8 These [slaves girls], instructed by the parasites, who taught them [to give] counsels etc. [pretended to have obtained] from conversations with the gods, confused his mind.

Some among these [slave girls] showed themselves eager for amorous intercourse at those occasions, and the king forsook his good fortune by touching them with his own body.

As he was anxious to live for a very long time, they [the goddesses] granted him, when in his foolishness he asked for a long life, hundreds of years to live.

When he desired to give magic perfection to his body (piNDasiddhi) some Domba [low-caste laundress] made him swallow a drink which he pretended was an elixir having that power.


What respectable man could relate the other even more shameful practices of his which he followed to obtain strength and beauty?33

Plate 9 Somewhere in that spectrum, between consorting with slave goddesses and siddhi-acquiring practices too shameful for KalhaNa to mention, must figure the celebrated group-clenches of the KandAriyA MahAdeo [e.g., Plate 8] and their precursors on the LakSman and ViZvanAth temples. Great significance has been inferred by Michael Meister and Devangana Desai from their distinctive location, exclusively on antarAla juncture walls between mahAmaNDapa worship halls and the vimAna sanctuaries proper.34 Thus it appears, (as a partial explanation), that three varieties of intimate engagement--sexual, mystical and architectural--overlay one another as mutually illuminating glosses. Far fetched to be sure, but not without sanction in texts, and I quote but two. First, from the BRhadAraNyaka UpaniSad these famous lines:

In the embrace of his beloved a man forgets the whole world,--everything both within and without; in the same way, he who embraces the Self knows neither within nor without.35

Second, from an Orissan architectural treatise exactly contemporary with Khajuraho's imperial monuments,36 these very components of the temple are correlated in gendered terms:

The vimAna is the best bridegroom and the mukhazAlikA main hall is the bride. The place in front, where the bridegroom and the bride meet becomes the place of junction (sandhikSetra).37

Plate 10 From these lofty heights of metaphoric fancy (building nuptials, God-consciousness) I descend to earthiness for a final introductory sampling of the erotic provocations on view at Khajuraho. Perhaps the site's most idiosyncratic motif is the varastrI/choice woman provocatively shown in the act of dislodging a scorpion from her thigh by removing her skirt. Plate 9 is but one of several figures on the KandAriyA MahAdeo that expose themselves on this pretext and there are other earlier examples at Khajuraho, but none elsewhere to my knowledge until later,38 at 12 c. Belur for example [Plate 10]. Unlike the later, more straight-forward depiction of skirt and scorpion removed, at Khajuraho the motif has an uncanny ambiguity. Why, for example, does the lusciously hip-shot poseur gingerly pinch at her skirt's right upper hem, when the ominous scorpion is firmly planted on her left leg? Pardon the anachronism, but talk about strip-tease! Not wishing to prolong suspense unnecessarily, I hasten to credit Devangana Desai with discovery of a site-specific connotation of the scorpion at Khajuraho:

KharjUra-vAhaka, the ancient name of Khajuraho, mentioned in King Dhanga's inscription of V.S. 1059 (A.D. 1002), can be interpreted to have two meanings. The well-known meaning of the word kharjUra is date-palm tree, and vAhaka means a carrier or bearer. So KharjUra-vAhaka can mean date-palm bearer, and a later legend associates the town with two golden kharjUra trees at its gate. But the word kharjUra also means a scorpion. So kharjUra-vAhaka [also] means scorpion bearer.39

But here our interpretations diverge. While she cites a sAdhana verse for Aghora Ziva as wearing a necklace of scorpions (however, vRScika)40 I prefer, more directly to the point, to identify these skirt-shedding women themselves as the eponymous Scorpion-bearers. Or, if not eponyms whence the site was originally named, at least they constitute namesake allusions to it--they too are kharjUra-bearers. Here it is relevant to note that the oldest architectural structure at Khajuraho is a yogini precinct,41 its sixty-four niches devoid of sculpture, I suggest, because originally live women may have figured there in rites intended to augur military victories.42 Thus the kharjUra/scorpion bearing attribute may have belonged to a local manifestation of the preeminent goddess of victory, DurgA.43

That said, there is yet another vital facet to disclose in the purport of these unnerving surasundarIs, or divine beauties with scorpions on their thighs. Looking again at the kharjUra-vAhakI in Plate 9, one would not be mistaken to see hints of inebriation, for kharjUra is also a word for date wine.44 Seen in light of this punning zleSa (a literary ornament much loved by Sanskrit poets), her pose struck becomes a perfect visual analogue to the literary figure known to alaMkArikAs as sandeha, intriguing doubt.45 Why indeed, does she delicately lower her garment? Due to kharjUra fear or arousal? These may be idle witticisms to some, but not, I insist to the sculptor who carved this masterful image of coquettish seduction. Notice, lest there be lingering doubt, that allusion to drinking is overtly mimed in the abhinaya gesture of the cupped thumb and fingers of her left hand, pressed, moreover, against a breast as if it were the cup.46 Welcome to KharjUravAhaka, the varastrI suggestively pantomimes, with both hands!

The Artha of Temple KAma.

Plate 11 So there it is: lascivious iconography in abundance and integral to the conception and design of the NAgara order temples at Khajuraho. But still the perennial bafflement remains: whatever for? By what criteria and evidence might genuine intentions of the architects, patrons and clergy of the Candella court be isolated from the unwarranted inferences of later apologists, present company not excluded? All one can really do is amass collateral documents of the period and test their relevance against careful reading of the monuments themselves according to one's own best lights at any given moment. Fortunately, the challenge to present day scholars is more one of feast than famine. A great wealth of potentially relevant documents have been identified, by T.P. Bhattacharya and Devangana Desai among others, and their findings are more than sufficient to start (or, rather, to continue) the winnowing process.

Thus, before examining four text-certifiable rationales for the plethora of sexual imagery at Khajuraho, I propose to clear away five other, less-credible alternatives. First, and notwithstanding the erudition of Alain Danielou, the variety of coital bandhas (clenches) or were not rendered in stone for sexual education of the general populace, newlyweds and kuNDalinI physiologists included.47 True, matches may be found, as between the straddling pose that tops the south antarAla of the VizvanAth [Plate 11] and prescriptions in sex manuals like the 12th century Ratirahasya,48 but divergences in matter, even between these relatively contemporary corpora, are far greater than their commonalties.49

Plate 12 Nor, for a second rationale applicable elsewhere but not at Khajuraho, were deliberately enticing images erected as a moral-filtration against weak-kneed aspirants still prone to lust. Obscene Idol Houses for such a purpose have been reported in VajrayAna Tibet,50 but parallel intentions cannot be presumed where sexual motifs are concentrated on exterior walls, to be glimpsed by one and all during pradakSina circumambulation, rather than by senior adepts undergoing final ordeals in emulation of prince Siddhartha's encounter with Mara's daughters just before achieving Buddhahood. Touchstones of renunciation51 these are not.

Plate 13 Third, there is no credible basis to believing that the Khajuraho temples were intended to replicate a morally stratified universe, like the Buddhist mandala architecture exemplified by Borobudur in which diminishing degrees of carnality correspond with spiritual attainment. The presumption that they might arises from the appearance of bestiality and group orgies on an outermost terrace frieze of the LakSman temple [Plates 12, 13 to be discussed below]--there and nowhere else at higher levels of the temple proper. But unlike Borobudur, where patently sinful acts are juxtaposed with resultant punishments,52 the LakSman terrace lacks any indication of cautionary admonition. Nor can a dichotomy be supposed between grosser exterior versus more ethereal interior formulations of religious experience53 when couples coupling (maithuna-mithuna ) appear on inner sanctuary walls, even in the ParzvanAth of the relatively circumspect Jains. [Plate 14]

Plate 14 Fourth, and not withstanding the inherent causalities between sex and fertility, both in biology and agrarian rituals since the neolithic,54 the Khajuraho temple builders were not primarily motivated in their choice of iconography by desires for offspring or harvest. Again, by contrast with earlier Buddhist monuments where fructifying yakSis were so prevalent, zAlabhanjikAs beneath trees are conspicuous by their rarity at Khajuraho, and maternal images even more so [exceptionally, Plate 15] among literally hundreds of variations on the theme of feminine grace and the preoccupations of women. To the proverbial list of the gods' four distinguishing features, a fifth could be added for these surasundarIs--they rarely get pregnant.55

A fifth untenable explanation for the sexual imagery at Khajuraho is that they were designed or dedicated for the use by any known Tantric sect like the PAzupathas, Kaula-KApAlikAs,56 KAlamukhas, or even the MattamayUra Saiva SiddhAntas known to have been extremely influential in the neighboring kingdom of the Kalachuris of DAhAla. The discounting of a any significant role of the latter monastic order comes as an admitted disappointment, given the absorbing historical reconstruction of events by Goetz, to the effect that MattamayUra ascetics infiltrated the Candella court and thus facilitated (with debauchery) their territorial loses to Kalachuri GAngeyadeva, c. 1025-1040.57 This otherwise plausible scenario is at odds both with the strictly orthodox tenor of the Candella's surviving temple dedications,58 and the comparatively chaste iconography (and epigraphic content) of proven MattamayUra temples and monasteries.59 Perhaps goose and gander were receiving different alchemical sauces (as from abstinent dealers of illicit drugs anywhere), and tantric admonitions to secrecy could explain the relative dearth of overt sexual imagery in the otherwise impressive stone-work at proven MattamayUra sites like Candrehi, Bilhari and Gurgi. But by the same token, their very prevalence at Khajuraho sets them apart as not responsive to those same guiding lights. Not that tantric beliefs and practices were unknown in Candella circles--far from it--simply they were not foremost among the motivations of their builders, as I reconstruct them.

1. Protection Gems

Plate 15 The architectural treatise closest to the royal temples at Khajuraho, both in date and region of composition, is the SamarangaNa-sUtradhAra, written by ParamAra king Bhoja who suffered a humiliating defeat at the hands of a vassal of Candella VidyAdhara, patron of the KandAriyA MahAdeo.60 By happier coincidence, this is also the text that best corresponds with iconography on the ground in its abandon of any lingering Gupta-period reticence about loving couples. Comparability in spirit, or rather in flesh, may be appraised, for example, by recalling Plate 2 (the only mithuna in Indian art with pubic hair notation) when reading the following exhortation:

[Temples] should be decorated with beautifully bejeweled youths, their attractive limbs entwined making love.

Heroes and women gratifying their desire for one another in sex-play, their pale bodies adorned with a few choice ornaments, with their limbs slightly enervated from sexual indulgence.61

(Notice how they hang in each other arms.) These graphic images of sensual abandon contrast markedly with the simpler prescriptions and carvings of earlier centuries, when temple door frames (primarily) were adorned with a variety of auspicious motifs, mithunas not predominant among them, and rarely engaged in more than hand holding. The 6th century BRhat-saMhitA is typical of the earlier period when it enjoins simply that:

The remaining part (of the doorjambs) should be decorated with auspicious birds, swastika designs, vessels, mithunas, leaves, creepers etc.62

At the same time one can readily accept Thomas Donaldson's thorough-going argument with the Orissan analogues in mind, that no categorical differences in kind or probable intent distinguish casually disposed mithunas from the more heated maithuna couplings.63 Presumably the differences constitute mere degrees in kind or shades from a single spectrum whose predominance at any given moment can best be explained by the standard dynamics of stylistic development and desire for variation on perennial themes. Thus a complex, but single rationale applies to all such architectural embellishments: in them inheres a dual complementary symbolism of propitious and apotropaic qualities.64 Whether their efficacy, alternatively, is perceived in terms of attraction or revulsion, the underlying purpose, very much from the architect's perspective, is protection. For the promotion of structural integrity, the aversion of untoward disasters, the building is dressed in auspicious ornaments, like a person with magical amulets. Since a vast literature on this and allied subjects is readily available,65 I merely defend one of the more risible claims to this effect with a proof text. Much scoffing has been elicited by the claim that erotic display may protect a temple from lightning. And for decades a sole textual citation in support of this conviction, ascribed still to zilpins in the 20th century, has bounced through the literature (though never actually quoted) as from the UtkalakhANDa (the Orissa section), severed from its resident text name, the SkandapurANa:

In order to ward off strokes of lightning, cracks in the structure and other calamities, gems etc. were suitably fitted in the manner prescribed in the treatises on architecture.66

Plate 16 Rather than question the rationality of such a belief, or even its dubious antiquity (in a text pertaining to the still-active JagannAth temple, at Puri), I prefer to speculate about how exactly apotropiac gems etc. (i.e., alaMkAras,including sexual motifs) were imagined to work. Given the Vedic god Indra's perennial and pan-Indian identity as vajra-wielder par excellence, coupled with his equally notorious affiliation with heavenly dancers--the apsarases whom he both enjoys and dispatches to debauch alarmingly powerful ascetics, one wonders whether images, acts and utterances of a frankly sexual nature are intended to curry his interest or shame him away? Arguments can be made for both alternatives, and not wishing to leave any stone unturned, here is one for each, applicable, perhaps, to different genres of temple iconography.

On the one hand, the architect responsible for giving prominence to such alluring females as this dancer from the PArzvanAth [Plate 16] is easily charged with intent-to-entice (and thus to attract divine favor) on the basis of this exactly contemporary passage in the Zilpa PrakAza (I.392-395):

...the NArIbandha [frieze of women] is indispensable in architecture. As a house without a wife, as frolic without a woman, so without (the figure of) woman the monument will be of inferior quality and bear no fruit.

Gandharvas, YakSas, RAkSasas, Pannaga (NAgas), Kinnaras become enchanted on seeing the graceful postures of women.

Woman is the most beautiful, when adorned with all ornaments.

Contemplated in various postures, she is known as AlasA-(Indolent) and is decorating...the walls and other parts of the mukhazAla (main hall).67

[Conversely, moreover II: 502b, 503]

A place without love-images (kAmakalA) is known as a place to be shunned.

In the opinion of the KaulAcAras (Tantric authorities) it is always a base, forsaken place, resembling a dark abyss, which is shunned like the den of Death.68

No surprises here: images of beautiful women and artful lovers are solicitous inducements for the gods to be present.

Plate 17 But on the other hand, what of the repellent capability of images? Might not deliberately obscene images be intended to keep lightning strikers, the evil eye, and any other untoward spirits and their attendant calamities at bay? In a word, yes, though here, in the absence of known proof texts the evidence is somewhat hypothetical: hypothetical, but not dependent upon standards of decency arbitrarily imposed from the outside. Compared to the initially shocking range of transgressive perversions (bhraSTa-kriya) found elsewhere in India (much of it comically absurd, like over-the-shoulder self-fellatio69), the famous terrace reliefs of Khajuraho's LakSmaN may seem pretty tame [Plates 13, 17]. But still the fact remains that oral sex acts (aupariSTaka) are disparaged in VAtsyAyana's KAmasUtra,70 and in the sex manual closest in date to the Khajuraho temples, dismissed entirely as unworthy of comment.71 Yet there is an abundance of good humored variety, at the outermost pale of LakSavarman's precinct--ideally situated, I infer, for repelling the evil eye.72 Interestingly, humor seems to be a consistent devise to assuage the more fastidious viewer. Consider, for example the dire consequences to a second soldier trying to coax his buddy's put-upon horse to open its mouth [Plate 12]--while a child absconds with his mount, riding bareback. And laity the world over enjoy laughter at the expense of lecherous clergy exposed. These on the LakSman terrace almost certainly include tantrics, including one who presides from a double-occupancy bed over the preparation of some performance enhancing elixir [Plate 18]--reminiscent of the pills Muslims also sought to acquire from sAdhus still resident at KajjurA in the 14th century.73

Plate 18 One of the great unsolved mysteries at Khajuraho concerns the label inscriptions that appear on several of the male participants in the group orgies [e.g., Plates 19, 13, 16]. No systematic collection and decipherment has yet been attempted, to my knowledge, but from causal observance of their orthography and distribution patterns in photographs, they appear contemporary with the temple's construction. Yet unlike the numerous other mason marks and sculptor names that Cunningham reported from throughout the site (often upside down, thus indicative of engraving before assembly into the temple walls),74 these seem to identify individuals or generic types known to the court and citizenry of ancient KharjUravAhaka. Corroboration of this hypothesis comes from the single reported decipherment of a label of this type: beneath a damaged maithuna couple at the west end of LakSman's south antarAla wall (he taking her from behind) is written ZrI SAdhu Nandi Khapanaka: holding a flagged-staff, perhaps a rajoharaNa duster used by Jains to avoid injury to insects in their path, the male must represent a particular reverend kSapanaka or Jain mendicant, named Nandi.75 Since corresponding labels do not accompany the more dignified central groupings, mostly royal figures to judge from their rich accouterments [e.g., Plates 8, 11], the labeling must have been intended to poke fun--again in the context of apotropaic sexual display. Thus, in summary of this first of four legitimizing purposes for lascivious iconography, I say, the artha of adharmic kAma is [not mokSa, but] AvarNa--i.e., shielding from evil spirits, while simultaneously attracting (with more dharmic, dignified enticements) an encircling, protective host of good gods.76

2. Power Plays

Plate 19 From the perspective of yajamAna interests, that of the royal donors, by contrast to that of the zilpa-builders, a primary function of sexual imagery on temples was declaration of personal charisma and capability to rule. Both for iconography and written edict, sexual prowess was a favorite metaphor, by which knowing visitors to a royal temple were appraised of the ruler's well-merited splendor and fame (tejas and kIrti). As noted earlier, for example, sculptures of beautiful women (varaStrI) on the VizvanAth correspond implicitly with its dedication's reference to the imprisoned wives king Dhanga had wrested from other kings, defeated in battle.77 To any unwilling to accept such overt insinuation of royal prerogatives into the function_ and fabric of medieval Indian temples, I reiterate the overwhelming preponderance of verses in dedicatory inscriptions of the period that eulogize the king, by contrast to the miniscule few that cover the obligatory nods to celestial deities for whose residence the temple is being prepared. Deity is the subject of only the first three out of 49 verses belonging to the LakSman dedication; the final two identify poet and engraver--the remaining 44 pertain primarily to the king. Royal temples concretized their patron's claim to centrality in a rAjamaNDala or imperial formation as king of other kings, the one most favored by Fortune, Earth and Victory, among other goddesses and their human equivalents. To be sure, they also served as places for worship of consecrated icons of divinity, but that was a secondary given, I would argue. Admittedly, this may be terrible theology--not a fair portrait of medieval Hinduism in general, but like it or not, the inscriptions and temple iconography speak mainly of and for kings, their builders, and not only at Khajuraho, of course.78

Plate 20 Here is another propitious pairing of Candella sculpture and poetry earnestly devoted to the portrayal of royal power in sexual terms. Plate 20 represents an architectural fragment of untraced provenance, now in the Cleveland Museum of Art. In style and subject matter--a king disrobing some girl--it is clearly Candella, eloquently embodying these sentiments of a late Candella inscription at Kalanjar:

He [ParamArdi, c.1166-1202], the greatest of kings, having drunk, like draughts of honey and curds, the shining fame of [other]kings, his enemies..., like the wind of the Malaya mountain [king ParamArdi] kisses sportively the lips of the maidens, red like the pomegranate, seizes them by their beautiful tresses, removes the garments that shine brightly on the high bosoms of the maidens, and easily dries the perspirations occasioned by sport from the brows of the fair.79

Yet another pairing of Candella text and sculpture, I present by way of insistence that something far more important than frivolous hedonism is here at work. Plate 21 is a gripping detail from one of the complex center-panels of group sex, on the south antarAla exterior of the LakSman, d. 954. Perfectly congruent with its imagery are these sentiments from a land grant inscription forty-four years later:

His son [i.e., son of the LakSman's patron LakSavarman] was the illustrious Dhangadeva, a fit dwelling for the goddess of victory, renowned in countless battles...

Strange it is, that the fire of separation is ever increasing in the hearts of the wives of his enemies, although it is incessantly sprinkled with the water of their tears. So long as he is the sole lord of the earth, (only) the curls of the damsels of the female apartments are loose,80 there is seizure by the hair (only) in amorous dalliance, hard are (only) the two breasts, crooked (only) the brows [etc.]...81

Plate 21 This metaphoric relationship between king and kingdom as that between lover and beloved(s) is certainly not unique to the Candella court, and perhaps the seemingly greater fixation on the theme at Khajuraho is unduly exaggerated by the comparatively greater devastation suffered by many of their contemporaries' monuments. In any event the non-frivolous character of this eroticized display of royal authority needs to be stressed. The harems of rAjAdhirAjas (paramount sovereigns) were not just repositories of sexual booty, nor only pleasure grounds where heirs essential to the regime's perpetuation might be conceived. Given the feudal array of tributary states beholden to one man, harems also functioned as virtual departments of state, destinations to which vassals, not always under coercion, might wish to depute marriageable daughters as sureties of allegiance to a suzerain. KanyopAyanadAna was the term used for this means of cementing political alliances, often negotiated in the terms of peace treaties following military engagements! In such a system, with the polity so thoroughly dependent upon the whim and vigor of single individuals, it is not improbable that pleasures of the royal bed chamber seemed at times more like work, bhoga more like yoga.

3. The Yoga of Bhoga

A canard of alarming prevalence in the apologist literature on Indian erotica is that yoga (disciplined action) and bhoga (pleasure) are one. If not absurd, to my lights, such assertions are indicative of gross negligence of nuance, for if there is anything tantric about the famed maithuna couplings at Khajuraho then they stand in eloquent testimony to the daunting arduousness of the left hand path (vAmamArga). Broad and easy it is not, though by no means straight and narrow either.

Plate 22 It is true, yoga and bhoga are frequently juxtaposed in Sanskritic sources, but the seductive appeal of their assonance would never arise were it not for an underlying antonymy. Yoga, cognate with yoke, means to harness (the vital breaths), yes, to discipline by denial of bodily impulse. Bhoga, pleasure, by contrast, arises from sensual gratification--from indulgence versus denial. Thus, the trick to be turned in tantra is to hotwire the system, to reverse motives and paradoxically to practice disciplined indulgence of the flesh-resident vital forces. Practices normally pleasurable (eating meats, drinking wine, and sexual intercourse), become sAdhana means of empowerment, precisely if and only if gratification as a motive can be denied, if consequences like orgasm can be suppressed (the hormones hopefully rechanneled up the spine) or at least deferred as long as possible.

With a quick disclaimer that experts can cite any number of exceptions to the above generalities, the great differences between Buddhist and Hindu formulations of tantric theory, to name only one,82 I propose to consider just two of the 14 major maithuna panels of group sex at Khajuraho83 as exemplifying the yoga of bhoga. The first, from the earliest of the three temples upon which they are found, is semi-light hearted [Plate 22]; the second, from among the latest set, is not. [Plate 23]

Plate 23 A detail of the first was already introduced [Plate 21] as an instance of seizure by the hair during good sex (surata-krIDAsu). If the reader can redirect her attention to pedantry for a moment, I find it amazing how frequently in the literature the whole point of this foursome has been missed. To me, at least, and at present, Plate 22 proposes an answer to this question: how great is their love? Put another way, how great a lover is king Lakzavarman (aka Yazovarman), patron and beneficiary of this temple? Permit me to count the ways this masterpiece would have us believe in his inordinate capacities. So arousing are his presence and amorous attentions that She84 seizes his forelock, throws an arm around his shoulder, the better to climb for a kiss. In effect this sculpture conflates (or intensifies by reduplication?), two postures from the KAmasUtra called latA-sAdhana, clinging like a vine to a tree, and vRkSAdhirUDaka, the tree-climbing pose.85 In VAtsyAyana's definitive words:

When a woman, clinging to a man as a creeper twines round a tree, bends his head down to hers with the desire of kissing him and slightly makes the sound of sut sut, embraces him, and looks lovingly towards him, it is called an embrace like the twining of a creeper.

When a woman, having placed one of her feet on the foot of her lover, and the other on one of his thighs, passes one of her arms round his back, and the other on his shoulders, makes slightly the sounds of singing and cooing, and wishes, as it were, to climb up him in order to have a kiss, it is called an embrace like the climbing of a tree.86

Meanwhile, behind her back, the bearded king affirms consummation of the embrace by displaying an ascending trikoNamudrA, the term being a necessary neologism for the male triangle symbol in tantric yantras (by contrast and in intersecting mesh with the inverted female triangle). This is a unique instance of such a gesture, so far as I know, though a six-pointed star signifying male/female union is one of the best known of Indian symbols, especially as reduplicated five times in (locus classicus) the Zri Yantra.87

But wait! There's much more being said in this prazasti to royal love, in the guise of attendant figures. Despite what others may have hoped, the naked female with her back against the king is not shyly covering herself like an ostensibly modest Venus Pudica.88 Rather, I fear it cannot be denied, given the masturbatory activity of the poor Jain monk on the opposite side,89 that she is too. So that's how strong His love is--enough to excite other women in self-absorbed revery and to turn regretful heads of celibate monks. Is it humor or profound insight into gender difference that while the female prefers to look away, the male is aroused while craning his neck to watch?

Pressing on, one may be permitted, I trust, a note of cynicism. How difficult can it have been for a king to shine as the supreme paramour when surrounded by hyper-stimulated, under-gratified varastrIs, each pampered with every luxury save exclusive access to her man, and never too sure of her position in the pecking order? Some yoga.

About the yogic and yes, even tantric, significance of the later example, however, there can be no doubt [Plate 23]. No male could possibly contemplate such acrobatics with three women for bhoga alone. At the very outset I confess to lacking the temerity to attempt any definitive exposition of this group clench, -- not one that occurs among the sanghATaka bandhas of either the KAmasUtra or RatIrahasya, I might add, chapters I.4 and 10 respectively. Instead, as an essential preliminary to some future article, I offer two more relevant texts, one from fiction of the period and the other from the tantric architectural treatise already utilized, RAmacandra KaulAcAra's Zilpa PrakAza. A great story first.

Once, as told in the 11th c. Ocean of Story,90 a young Brahmin left town on some errand. In his absence a kApAlika ascetic, skull-mounted khaTvAnga scepter in hand, spied his beautiful wife when approaching their house for alms and cast a spell on her. Immediately she was struck down by fever and died before evening. Before her husband returned home, grief stricken-relatives had placed her body on a funeral pyre. The young man, named CandrasvAmin, arrived just as the flames started shooting into the night sky. Then from out of the crowd stepped that kApAlika who resurrected her unscathed with a handful of sacred ash (Ziva's vibhUti). Still under a spell, however, she accompanied the kApAlika straight out of town, her husband following in hot pursuit with his bow and arrows. On the banks of the Ganges they entered a cave where the kApAlika had already imprisoned two other women, daughters of the king of Benares and a merchant. Presenting CandrasvAmin's wife, the kApAlika exalted: She, without whom I could not marry you, though I had obtained you [by identical means], has come into my possession; and so my vow has been successfully accomplished. Just then CandrasvAmin jumped forward to grab the kApAlika's khaTvAnga staff and throw out into the river. Thus bereft of his magic powers the scoundrel tried to run; but I drew my bow and killed him with a poisoned arrow, CandrasvAmin later told another king (contemplating multiple-partner sex also, but that's another story)!

Before turning to the equally relevant Zilpa PrakAza, let us consider for a moment the amazing parallels. Not only do Somadeva's story and Candella king VidyAdhara's temple both feature sexual designs between a primary couple and two attendant women, but they both date from the 11th century and are located in the Gangetic heartland of North India; even Benares was under Candella control at the time.91 Too much of a coincidence, even for me, is the fact that the name Candella denotes affiliation with Candra, the moon, as does the Brahmin hero's name, CandrasvAmin. At a minimum, it would be foolish not to infer some common origin for both, and more likely, the story reflects common knowledge of the imagery at Khajuraho and/or lost equivalent images or practices from elsewhere in the Gangetic Doab.

In the Zilpa PrakAza, a leitmotif is the essential role of yantras, schematic geometries into which deities are distributed, for the protection and longevity of temples. One of the most important, incorporating far more participant beings than the panel in question on the KandAriyA MahAdeo, but anyway, is the KAmakalA Yantra. In wry deference to its author's insistence on total secrecy, I forgo the task of summarizing its constituent parts and significance as detailed in verses 508-541 of the text's second prakAza. In fact, for our purposes it suffices to quote but two, astonishingly revealing verses:

This yantra is utterly secret, it should not be shown to everyone (to others). For this reason a love-scene (mithuna mUtri) has to be carved on the lines of the yantra.

In the opinion of KaulAcAras it should be made on the lovely jAngha in the upper part of the wall. The kAmabandha is placed there to give delight to people.92

Returning for one last look at the inverted king in intercourse with one women while fondling two others [Plate 23], I ask the reader's indulgence to make this poetic leap: paraphrasing Mark Twain on the disputed authorship of the Homeric epics, if this panel wasn't intended to conceal by delight the KamakalA Yantra, then it conceals another worthy of the same name, Love-art. And if I may be permitted to defer further exposition till such time as all 14th kAmakalA bandhas at Khajuraho can be studied as a corpus, I take it as proven that some sort of Yoga of Bhoga is being demonstrated, or rather, is being used to camouflage with titillating flesh something still more esoteric and inscrutable to the uninitiated.

4. Phantasmagorical Castles

Plate 24 Of all the metaphoric formulations of the Hindu temple--mountain, palace, altar, divine embodiment, chariot--for me it is the last that provides the surest key to unlocking the mystique surrounding its sexual imagery. To expand upon my favorite phrase to the Pali Text Society Dictionary's definition of vimAna (undeniably the most common architectural term for the sanctuary structure proper), they are "immeasurably" palatial residences of the meritorious celestials (devatAs), capable in myth of appearing suddenly or darting off again at their occupants' will, UFO-like. Exactly like the Candellas planned for their vimAnas at Khajuraho:

...these towering mansions (are surrounded) by lovely, well-planned gardens...Lotus ponds with cool waters invite to refreshing baths; a host of birds mix their songs with the strains of cymbals and lutes, played by heavenly musicians. Angelic maidens perform their dances, filling the atmosphere with a radiant light which shines from their bodies. Peace and happiness reign everywhere, the joys of such a vimAna cannot be expressed in words. This elysiam lasts for aeons...93

Shifting facets on this metaphoric jewel of inquiry, I wish to add that implicitly at least, royal temples on the order of those at Khajuraho were semi-funerary in function (their dedications twice posthumous, remember), standing like their Mt. Meru-styled cousins of Cambodian kings, or like the pyramids of Egypt for that matter, as memorial aids for apotheosis of identifiable individual rulers. By way of textual authority for this admittedly bold assertion, I cite Krishna Deva's passing mention of an early 12th c description of svargArohaNa-prAsAda, lit. temples for flying to heaven.94 Differences or similarities in structural components between the prototypic heavenly-flyers of that text and those on the ground in Khajuraho (and there are important instances of each) are less material here than commonalties of intention, as registered by two epigraphic parallels. First, in the year 1000/01 a Candella feudatory named Kokkala, was:

desirous of crossing the deep ocean [i.e., the abyss of mortality]..., he caused to be erected this (temple, high like?) the spotless great peaks of the mountain of snow, the lofty golden dome of which, because it is in contact with the fierce splendour of the sun, became a spotless canopy for the glorious lord VaidyanAtha.95

Second, and at about the same time, in neighboring Kalachuri country:

By him [PrazAntaziva, a MattamayUra ascetic] was established a temple of Ziva (built) to the north of the palace which was built by the illustrious (Kalachuri king) YuvarAjadeva and was like the shining peak of the KailAsa mountain, which (temple) aspired to be as high as the peak of the Sumeru mountain, was famous on the earth, caused wonder in the three worlds and acted like a stair-case to his fame marching towards heaven.96

All that remains to flesh out here is the peculiar fixation upon sexual gratification that ancient Indian texts promised the heaven-bound. Consistently from the Vedic period the dancing nymphs of heaven called apsaras are said to be eagerly awaiting new arrivals.97 A poignant favorite I offer in closing, together with a matching sculpture type of great frequency at Khajuraho [Plate 24]. According to the MahAbhArata:

Thousands of handsome Apsarases run up in haste to the hero who has been slain in battle (exclaiming) be my husband.98

Accordingly, can we not see special significance in the frequent image of a temple apsaras applying kunkum to the center part of her hair? Since vermillion powder there signifies a married woman, this must be the anticipatory gesture of these varApsarah, the choice nymphs of heaven, preparing for nuptial union with the temple donor upon his decease, and by extension offering an alluring foretaste of paradise to every subsequent visitor to Khajuraho. In the mysterious words of the KandAriyA MahAdeo's brief inscription,99 VarastrIs are welcoming a new god, in the time of king VidyAdhara.


  1. As defined by Ronald Inden, Imagining India (Oxford: Basil Blackwell, 1990) esp. ch 6.2 Reconstructions: The Imperial Formation of the Rashtrakutas.

  2. E.g., 5th century ruins at Nachna, Bhumara and especially Deogarh (where an 11th c. Chandella inscription also appears) J. C. Harle, Gupta Sculpture (Oxford: Clarendon Press, 1974.); Joanna Gottfried.Williams, The Art of Gupta India: Empire and Province (Princeton: Princeton University Press, 1982). The previously cited Encyclopedia of Indian Temple Architecture, ed. Michael W. Meister, is not yet complete and the volume with coverage of the major monuments at Khajuraho is still awaited: their antecedents, however, are well documented in Vol II, North India; Part 1, Foundations of North Indian Style, c. 25- B.C.-A.D. 1100.; Part 2, Period of Early Maturity, c. A.D. 700-900 New Delhi: AIIS, 1991.

  3. In their inscriptions Indian kings were frequently beloved of Earth and Fortune, wives of the most regal god, ViSNu, of whom they in turn were considered aMSas, fractional avatars.

  4. Plus half again as many minor shrines of the same period. There is also evidence in rubble piles for a handful of additional temples, but legends as recorded in the MahobA Khand (c. 17th c) about an original 85 are specious. For critical appraisal of the late bardic traditions of Bundelkand, Sisir Kumar Mitra, The Early Rulers of Khajuraho (Calcutta: Firma KLM, 1958), pp. 6, 12-26.

  5. On coming to power in A.D. 999, Mahmud resolved to make annual forays into India and did in fact undertake at least 17 devastating campaigns in the next 28 years. Muhammad Nazim, The Life and Times of SultAn MahmUd of Ghazna (Lahore: Khalil, & Co., 1973).

  6. ...then we came to...KajarrA where there is a great pond about a mile in length near which are temples containing idols which the Muslims have mutilated. Ibn Battuta (1304-1377), The ReHla of Ibn BaTTUTa (India, Maldive Islands and Ceylon): translation and commentary by Mahdi Husain. 2nd ed. (reprint): Baroda: Oriental Institute, 1976. Gaekwad's Oriental Series, no.122, p. 166.

  7. And to be fair, a four-line pilgrim's record in devanAgari. The MAtangezvara linga is a polished cream-sandstone cylinder, looking for all the world like a recycled Maurya pillar fragment, 2.5m high by.93m in diameter. Infrequently published, it may be seen in figs. 32 and 34 of Shobita Punja, Divine Ecstasy: the Story of Khajuraho, photographs by Toby Sinclair. (New Delhi; New York: Viking by Penguin Books India, 1992).

  8. Captain T.S. Burt, Bengal Engineers, Extract from the Journal of his travel in the hands of Thaker & Co, for the Press, Bengal Asiatic Society vol VIII Part I (1839) pp. 159-184. Of course, they had never completely disappeared from regional consciousness, as confirmed by their mention in late medieval ballad cycles. A major pilgrimage fair coinciding with Ziva RAtri was already (or still) flourishing there when Alexander Cunningham made his archaeological tours through the region in the mid-19th century. Cunningham's first visit, as reported in the Archaeological Survey of India Report, vol II, was in January 1852; see also his subsequent reports in vols. VII, X and XXI.

  9. Inden, esp. p. 248.

  10. To extrapolate from parallels in orthography and content with two subsequent inscriptions, Yazovarman's of Vikrama 1011 (A.D. 954), and Dhangadeva's of V. 1059 (A.D.1002), to be introduced below. This is the implicit inference of their editor, F. Kielhorn, Inscriptions from Khajuraho, Epigraphia Indica, vol. I (1890-91): 121-153.

  11. This attribution is hypothetical, pending my own stylistic analysis of the monuments.

  12. Celebrated in the MahAbhArata and purANas, with Gupta-period remains and a natural cave identified with Ziva as NIlkaNTa, the neck-stained poison-drinker.

  13. Kanwar Lal, Immortal Khajuraho (Delhi: Asia Press, 1965), p. 220; and Krishna Deva, Temples of Khajuraho (New Delhi: Archaeological Survey of India, 1990), vol 1, p. 193, both following Krishna Deva, The Temples of Khajuraho in Central India, Ancient India no. 15, Bulletin of the ASI (1958), p. 57.

  14. In the opinion of Hermann Goetz,The Historical Background of the Great Temples of KhajurAho, Arts Asiatiques, t. V fasc.1 (1958): 35-47; Given the following record of loot MahmUd of Ghazni seized from much less politically powerful Mathura, I concur: The booty captured included five idols of gold, one of which was set with two rubies of the value of 50,000 dInArs, 200 idols of silver, and a sapphire of unusually large size. Nazim, , p. 108. The gold images were estimated to be 15 feet in height and so heavy as to require breaking apart before weighing.

  15. Yazovarman's inscription speaks of its acquisition from Ziva's mythic mount KailAsa by a Tibetan king who presented it to the SAhi kings of the Northwest who passed it on to the PratIhAra king HerambapAla in exchange for a contingent of elephants and cavalry. Kielhorn, p. 134, v. 43.

  16. Multi-headed with VarAha-boar and Narasimha-lion masks protruding on either side of the central human head: Lal, Pl. 27; Krishna Deva, vol 2, Pl. 41. For examples of the presumed prototype, always with a 4th demonic mask at the rear, see Pratapaditya Pal, Some Kashmiri-style bronzes and problems of authenticity, John Guy ed., Indian Art and Connoisseurship: Essays in Honour of Douglas Barrett, (Ahmedabad: Mapin with Indira Gandhi National Centre for the Arts, New Delhi; Middletown NJ: Grantha, 1995), pp. 86-98.

  17. LakSanAth > LakSman [the epic Rama's brother], says Goetz (p. 37), with interposition (unnecessary in my opinion) of nAth, protector, a frequent component of deity and temple names.

  18. Kielhorn, no. III, pp. 135, 136.

  19. Kielhorn, no. IV, pp. 137-147; for its translation Krishna Deva, 1990, vol 1, pp. 363-370.

  20. The two are also surprisingly close in basic dimensions: the LakSman being 85 ft long by 44 ft in breadth and 89 ft in total height (including a 9 ft terrace), and the VizvanAth just slightly larger at 87.5 by 46 ft and 91 in height.

  21. v. 48 "Glorious in the world is the divine lofty linga, made of emerald (marakata) which was worshipped by Indra, and which having been obtained from him as a favour by Arjuna was brought to the earth by him for worship by YudhiSThira and (lastly), was installed by the illustrious Dhanga with due obeisance." Krishna Deva, vol. 1, p. 368. In the next verse a second image of Siva, presumably the sandstone linga still in situ, is mentioned as installed (pratiSThitaH) by the same king.

  22. Krishna Deva, 1990, vol. 1, p. 369. Though composed in V. 1059 (A.D.1002), the inscription survives as recopied (with two additional verses) 114 years later, during the reign of Jayavarman.

  23. In c.1041 "GAngeyadeva...a thunderbolt falling on the heads of enemies...fond of residing at the foot of the holy fig at PrayAga...striving after final beatitude...he had found salvation there together with his hundred wives." F. Kielhorn, Jabalpur Copper-Plate Inscription of Yasahkarnadeva (A.D. 1122), Epigraphia Indica, 2 (1894), p. 6.

  24. Krishna Deva, 1990, vol .1, p. 368, with this note appended: "LokAloka is the legendary mountain outside the earth's sphere where the Sun never reaches. The poet [named KAyastha YazapAla, v. 59] means that his fame went where even the Sun could not reach." A minor quibble: I take it to mean the temple here at Khajuraho is a metaphoric LokAloka: ie., worthy not only of the more commonplace comparison with the Himalayan axis mundi, but with mountains out of this world.

  25. J.C.C. Southerland, (appended to Burt, no. 17 above) Journal of the Bengal Asiatic Society vol VIII Part I (1839) pp. 159-184.

  26. 109 x 59.5 ft and reaching a height of 116.5 ft, according to Cunningham, Archaeological Survey of India Report vol II (1871), p. 419. Krishna Deva's scaled elevation drawing (1990, vol. 2, Fig. 18) projects a height of 112.5 ft, including its 10 ft terrace.

  27. after kandaravat, containing caves, or valleys, and kandara-kAra for mountain in Monier-Williams, A Sanskrit-English Dictionary (1883. reprint ed. Delhi: Motilal Banarsidass, 1970). Goetz (p. 36) infers derivation from Maraketzvara, the Emerald-Lord of Dhanga's inscription, despite the latter having been found at the VizvanAth, 250m to its East. Pramod Chandra, whom he cites on another point does not (as implied) subscribe to that attribution, in The Kaula-Kapalika Cults of Khajuraho, Lalit Kala 1-2 (1955-56): 100, but rather the one detailed below. For the transliteration, KandAriyA MahAdeo, I follow Krishna Deva.

  28. Krishna Deva, 1990, vol. 1, p. 346

  29. Ibid, p. 371.

  30. Krishna Deva has found one or two not-terribly helpful textual citations from Bhoja's 11th c. samarangaNa-sUtradhara and the much earlier arthazAstra of Kautilya, ibid, pp. 370, 371.

  31. Al-Biruni, Kitab-ul-Hind, trans. E.C. Sachau under Alberuni's India, 2 vol. (London: 1910) II, p. 157. J.C. Harle derides A.L. Basham for positing linkage between devadAsi prostitution and temple iconography: respectively The Art and Architecture of the Indian Subcontinent. (The Pelican History of Art). Harmondsworth: Penguin Books; New York: Viking Penguin, 1986), p. 238, n. 56, and The Wonder that was India (London: Sidgwick & Jackson, 1954), p. 362 (citing the then-unpublished opinion of P. Rawson). Vociferous defense of the devadAsI hypothesis is made by Nirad C. Chaudhuri, The Continent of Circe: being an essay on the peoples of India (London: Chatto & Windus, 1965) pp. 217-220; approvingly cited by David Lorenzen, The KApAlikas and KAlAmukha: Two Lost Zaivite Sects (Berkeley: University of California Press, 1972) p. 139. For concurring views and much documentation see also Y. Krishan, The Erotic Sculptures of India, Artibus Asiae 34 (1972), pp. 331-343; Devangana Desai, Erotic Sculpture of India: a Socio-Cultural Study (2nd ed. New Delhi: Munshiram Manoharlal, 1985); Thomas Donaldson, Kamadeva's Pleasure Garden: Orissa (Delhi: B.R. Publishing Corp, 1987) and David Kopf, Dancing Virgin, Sexual Slave, Divine Courtesan or Celestial Dancer: In Search of the Historic DevadAsI, George Kliger, ed., Bharata Natyam in Cultural Perspective (Delhi: American Institute of Indian Studies and Manohar, 1993), pp. 144-180.

  32. A rare non-royal temple donation was jointly made by a palace guard (pratihAra) and prima dancer (mahAnAcanI), PadmAvatI, at Kalanjar in 1129; Cunningham, Archaeological Survey of India Report vol XXI (1885), p 34.

  33. KahlaNa, RAjatarangiNI VII, 1129-35, trans. M.A. Stein (1900. reprint Delhi: Motilal Banarsidass, 1979) vol 1, pp. 355, 356.

  34. Michael W. Meister, Juncture and Conjunction: Punning and Temple Architecture, Artibus Asiae XLI (1979): 226-228; Devananga Desai, Puns and Intentional Language at Khajuraho, Kusumanjali: New Interpretation of Indian Art & Culture: Sh. C. Sivaramamurti commemoration volume, ed, M.S. Nagaraja Rao (Delhi: Agam Kala Prakashan, 1987) vol. 2, pp. 99-108.

  35. BRhadAraNyaka UpaniSad IV.3.21. First art historian to cite this passages, Ananda Coomaraswamy, La Doctrine de la "Binunit" Etude Traditionelles, 1937, pp. 289-301; then in turn by Stella Kramrisch, The Hindu Temple (Calcutta: University of Calcutta, 1946), vol. 2, p. 346; Alain Danielou, An Approach to Hindu Erotic Sculpture, Marg, 2:1,2 (1947), p. 88; Allan Watts, photographs Eliot Elisofon, Erotic Spirituality: The Vision of Konarak (New York: Macmillan, 1971), p. 90; and in the final paragraph of the most handsomely illustrated of all books on the subject, Krishna Deva, photographs by [Ms.] Darshan Lall, Khajuraho (New Delhi: Bhrijbasi Printers Ltd 1987), p. 205.

  36. RAmacandra KaulAcAra, Zilpa PrakAza, Alice Boner and Sadasiva Rath Sarma, trans. (Leiden: E.J. Brill, 1966): as the temple components it prescribes are best exemplified by the VarAhI at Caurasi (p. xix), I accept Thomas Donaldson's 10th century date for the latter as most likely for the text (rather than the translators' broader 9-12th c. time frame): Donaldson, pp. 422, 423.

  37. Zilpa Prakaza II.595b, 596a; trans., p. 111.

  38. The so-called scorpion beneath a 5th c. Mathura figure fragment, is more likely a lizard [godhikA], with all due respect to Pramod Chandra, The Sculpture of India: 3000 B.C.-1300 A.D. (Cambridge: MA, Harvard University Press and National Gallery of Art, Washington, D.C., 1985), fig. 26, p. 79.

  39. Desai, 1987, p. 384.

  40. bhuangakeyUaradharam sarpahAropavItinam /
    gonasam kaTisUtram ca gale vRzcikamAlikAm //
    ity-aghore tRtIyo rudraH; AparAjitapRcchA 212,15 and almost verbatim in RUpamaNDana IV.6.

  41. Indeed, according to Krishna Deva, (1990, vol. 1, p. 26)the earliest among fourteen other yoginI temples known in North India, mostly clustered in regions surrounding Khajuraho.

  42. This was its remembered purpose among local Brahmins in the 19th century: Cunningham (1871), p. 417. I infer recourse to living yoginIs, not only from the absence of stone images (not withstanding 3 that were found--not matching as parts of a set), both also from the peculiar passageway, just 20" wide, cut through the south end of the rectangular precinct, beside the principal shrine. Obviously, its location is ideal for emanations from a primary deity to make their entrance, unbeknownst, perhaps, to some knowing only of the temple's visible north entrance. Of course, textual sources leave no doubt that actual women (at times coercively) were pressed into service for tantric rites of all kinds. E.g., Sir John Woodroffe, Zakti and Zakta 6th ed (Madras: Ganesh & Co, 1965), pp. 609-610. 617 ff on three classes of women for pancatattva secret ritual: wives and among non-wives, a) those for bhogya enjoyment; b) those for pUjya veneration (only).

  43. Groupings of 64 and 81 yoginIs are inextricably linked with permutations of nava-(9) DurgAs, and other groupings of one central deity and 8 AvarNa attendants. For details, Vidya Dehejia, YoginI, Cult and Temples, A Tantric Tradition (New Delhi: National Museum, 1986).

  44. Third among twelve varieties of liquor said to be enumerated by Pulastya: 1 panasa (bread-fruit, jack-liquor); 2 drAkshA (grapes); 3. kharjUrI (from the date-palm); 4 common palm (tAlI), 5 coconut (nArikela); 6 ikSu (sugar cane); 7 the MAdhvIka plant; 8 saira (long pepper); 9 arishTa (soap-berry); 10 mAdhUka (Bassia Latifolia); 11 GauDi or Maireya (rum from molasses); and 12 surA (arrack from rice or other grain). Monier-Williams Brahmanism and Hinduism, p.194.

  45. E.g., DaNDin, KAvyAdarza .2.26, cited by Edwin Mahaffey Gerow, A Glossary of Indian Figures of Speech (The Hague: Mouton, 1971), p.166, 312. Lest think I'm over reaching, a verse from DaNDin's KAvyAdarza is quoted in a Candella inscription: in a still-unpublished monograh I discuss numerous visual analogues to alaMkArika figures of speech at MAmallapuram, the Pallava port that DaNDin is known to have visited in c. 700.

  46. In effect the figure is performing a charade, an antarAlApa prahelikA in the terminology of Sanskrit alaMkArikAs. Ludwik Sternbach, Indian Riddles: a Forgotten Chapter in the List of Sanskrit Literature (Hosiarpur, Vishveshvaranand Vedic Research Institute, 1975), pp. 67-84

  47. His 8th of 10 pronouncements: ...Sexual education through images, of every possible form of sexual enjoyment, is a useful means of mental clarification... Alain Danielou, An Approach to Hindu Erotic Sculpture Marg, 2:1,2 (1947), p. 89 in 79-91l; see also Mulk Raj Anand, Of KAma KalA, p. 60.

  48. "If [while he stands] she sits in his hands with her arms round his neck and her legs round his waist, moving herself by putting the toes of one foot against the wall, throwing herself about, crying out and gasping continually, this is the suspended position (avalambitaka)." Kokkoka, Ratirahasya, Alex Comfort trans. as The Koka Shastra (New York: Stein & Day, 1965), p. 140. Without corresponding citations, several other sculpture groups are assigned Sanskrit nomenclature by R. Nath Art of Khajuraho (New Delhi: Abhinav, 1980), plates CLX-CLXXI.

  49. According to Comfort, in a lengthy note, p. 140.

  50. Urmila Agarwal, The Mithunas: Why 'Obscene' Sculptures?, Oriental Art XIV (1968), p.260, citing Harrison Foreman, Through Forbidden Tibet (London, 1936), pp. 107-109.

  51. Danielou's poetic phrase for his rationale no. 6, p. 88.

  52. Thus illustrating causal laws from a text called the MahAkarma-vibhanga: Heinrich Zimmer, Art of Indian Asia, (Princeton: Princeton University Press, 1955), 2 vols., Plate 479 b.

  53. Coomaraswamy, Arts and Crafts of India and Ceylon (London, 1913), pp. 63-65.

  54. These are thoroughly surveyed by Desai in her chapter 6, Sex in Religion: Magico-Religious Beliefs and Practices.

  55. According to the MahAbhArata Nala-Damayanti story, the gods appear disguised as men they still don't blink, sweat, cast shadows or quite touch the ground. Elsewhere, of course, apsaras do give birth, after Indra sends them to seduce overly powerful ascetics.

  56. Contra. Pramod Chandra, The Kaula-Kapalika Cults of Khajuraho Lalit Kala 1-2 (1955-56) 98-107. Attributes Chandra believed to mark Saiva ascetics (chiefly tonsure, and a scepter like implement with a flaring -head) have subsequently been proven to denote Jain monks, holding the picchikA, peacock-feather whisk for clearing sentient beings from their path. This correct identification was first made by L. K. Tripathi, The Erotic Scenes of Khajuraho and Their Probable Explanation, Bharati vol. 3 (1959-60) 104 f [still on order], after discovery of a label inscription beneath one such, identifying him as ZrI sAdhu Nandi Khapanaka, i.e., a kSapaNakaH, according to Monier-Williams, ... a religious mendicant, especially a Jaina mendicant who wears no garments (MBh i.789).

  57. (The same king who took 100 woman with him, via drowning , above, n. 33) Hermann Goetz, The Historical Background of the Great Temples of KhajurAho, Arts Asiatiques, t. V fasc.1 (1958): 35-47.

  58. E.g., in Dhanga's prazasti of A.D. 1002: v. 53: "Benevolent Brahmins of pure lineage, busy with the six functions, spotlessly clean though their bodies were smoky due to the smoke from the sacrifices undertaken, were settled by him in mansions, high, like the peaks, of KailAsa, after being honoured by gifts of wealth, grain, cattle and land." Krishna Deva, 1990, vol. I, p. 369

  59. Banerji, R.D. The Haihayas of Tripuri and their Monuments Memoirs of the Archaeological Survey of India, no. 23; and V.V., Mirashi, The Zaiva AcAryas of the MattamayUra Clan, Indian History Quarterly vol. 26 (1950) pp. 1 -16.

  60. After having the temerity to invade spheres of Candella hegemony; to be fleshed out from S.K. Mitra, p. 75; EI I pp. 219, 222, v. 22.

  61. SamarangaNa-sUtradhAra 34: 32-34a.
    "kumArakaizca krIDadbhiryuktA lalitabAhubhiH /
    vAsadhAmni nivezyante vicitrAbharaNAmbharAH // 32
    ratikrIDAparA nAryo nAyakastu yadRcchAyA /
    ApANDudehacchavayaH svalpacAruvibhUSaNAH // 33
    kiJcitpratanubhirgAtraiH kAryAH suratalAlasAH / 34a
    Translated here for the first time, though first cited in part by Tarapada Bhattacharya, Some Notes on the Mithuna in Indian Art, Rupam, (Jan, 1926) pp. 22-24; and again in his Canons of Indian Art, (Calcutta, 1926), p. 230.

  62. ZeSaM maGdalyavihagaiH zrIvRkSaiH svastikairghaTaiH /
    mithunaiH patravallIbhiH pramathaizcopazobhayet //

    VarAhmihira, BRhatsaMhitA, 56.15 This and similar citations from the Agni PurAna , HayasirSa-pancarAtra, Mayamatam and Zilparatnam are cited by Bhattacharya.

  63. Thomas Donaldson, Propitious-Apotropaic Eroticism in the Art of Orissa, Artibus Asiae 35 (1975): 75-100.

  64. Ibid., p. 88.

  65. See also, Donaldson's book-length study, Kamadeva's Pleasure Garden: Orissa (Delhi: B.R. Publishing, 1987), and Devangana Desai's chapter VI, Sex in Religion: magico-religious beliefs and Practices, Erotic Sculpture of India: A Socio-cultural Study 2nd ed. (New Delhi: Munishram Manoharlal, 1985).

  66. Skanda PurANa II.ii.21,45
    vajrapAtAdibhaGgAdivArNArthaMyathocitam /
    zilpaSAstreSumaNyAdivinyasyapauruZAt //

    trans. G.V. Tagare (Delhi: Motilal Banarsidass, 1994), V, p. 125

  67. Loukike kathito nArIbandhaH zilpasamudbhavaH /
    vinA nArIM yathA vAsaH krIDA nArIM vinA yathA //
    II. 392
    vinA ca lalanAM loke kIrtirjAyeta niSphalA /
    gandharvayakSarakSAMsi pannagAH kinnarAsstathA //
    darzanAt tatra muhyanti nAgarIbhaGgimuttamAm /
    UttamAM ramaNIM jurhAt sarvAlaMkArabhUSitAm //
    NAnAbhaGjIsamAhAri alsA sA vidhIyate
    GavAkSe zikhare vApi mukhazAlaGgamaNDane //

  68. SthAnaM kAmakalAhInaM tyaktamaNDalamucyate // II. 502
    KaulAcAramate hInaM sarvadA tyakhamaNDalam /
    KAlakakSasamaM tyaktaM tatsthAnaM gahanopamam //

  69. at Bagali and elsewhere in Karnataka especially; Desai, Erotic Sculpture of India, plate 124; and S. Settar, The Hoysala Temples: Study on the art and architecture of temples constructed during the reign of Hoysalas, 1000-1336 (Bangalore : Kala Yatra Publications, c1991-1992). 2 vol., plates; 262, 270, 271, 310.

  70. "The Acharyas are of the opinion that this Auparishtaka is the work of a dog and not of a man, because it is a low practice, and opposed to the orders of the Holy Writ... [though later conceding, ] all these things connected with love, everybody should act according to the custom of his country, and his own inclination."
    KAmasUtra II.9.22 ff. trans. Sir Richard Burton and F.F. Arbuthnot, The Kama Sutra of Vatsyanana (New York: G.P. Putnam's Sons, 1967) p. 118.

  71. "Why should we concern ourselves about oral congress when VAtsyAyana has declared it as utterly detestable?" Kokkoka, Ratirahasya VIII.66 quoted by Desai, p. 216 but excised from Comfort's translation, The Koka Shastra.

  72. The evil eye per se is more widely reported from South India, where, not coincidentally, bhraSTa sex acts are a ubiquitous accouterment carved on the wooden temple rathas (chariots). E.g., W.T. Elmore, Dravidian Gods in Modern Hinduism (1915, reprint New Delhi: Asian Educational Services, 1984),pp. 142-144.
    "The fear of the evil eye among the Dravidians is most easily explained by this fear of evil spirits...The placing of obscene figures and carvings on idol cars and temples is often explained in the same way."

  73. As noted earlier (n. 16), Ibn BaTTUTa visited KajarrA in 1342: "At the four corners of the pond [a mile long] are cupolas in which live a body of the jogis who have clotted their hair and let them grow so that they became as long as their bodies and on account of their practicing asceticism their colour had become extremely yellow. Many Musalmans follow them in order to take lessons from then"
    He goes on to report in this context that "one of [these yogis, in far South India] made pills for SultAn GhiyAs ud-dIn ad DAnghAnI, king of Ma'bar--pills which the latter was to take for strengthening his pleasure of love. Among the ingredients of the pills were iron filings. This effect pleased the sultAn, who took them in more than necessary quantity and died." The ReHla of Ibn BaTTUTa, p. 166.

  74. Cunningham, Khajuraho in Archaeological Survey of India Report vol. II, pp. 420-437.

  75. L.K. Tripathi, The Erotic Scenes of Khajuraho and Their Probable Explanation, Bharat vol. 3 (1959-60) p. 93; and cited by others including, Hiram W. Woodward, The Laksamana Temple, Khajuraho, and Its Meaning, Ars Orientalis vol. 19 (1989), p. 28.

  76. The same term in a compound, AvarNa-devatA, is standard for the named protective deities that are commonly assigned niches, aligned with the cardinal directions. E.g.

  77. Above, n. 34.

  78. Consider South Indian parallels, for example, where temple names typically subsume the patron's, like Cola RAjaRaja I's RAjarAjezvara at TanjAvUr, India's largest temple ever, built exactly at the same time as the KandAriyA MahAdeo. For earlier instances of this pattern, see my chapter, Royal Temple Dedications of the Pallava Dynasty, Donald S. Lopez, ed., Religions of India in Practice, Princeton Readings in Religion series, (Princeton University Press, 1995): 235-243.

  79. vv. 25 and 27 of a large back stone inside the NIlakaNTha temple, d. Vikrama Samvat 1258 (A.D. 1201), emphasis supplied, trans. Lieutenant Maisey, Bengal Asiatic Society Journal, vol. XVII (1838), p. 313; reprinted by Cunningham Archaeological Survey of India Report vol. XXI (1885), p. 37.

  80. alternatively twisted, crimped, by contrast to negative connotations of *bhanga* in terms of social order or polity; see also the translators note 16.

  81. tasya ZrI DhaGgdevo=bhUt=putraH pAtraM jaya-zriyaH /
    asaMkhya-saMkhya-vikhyAtaH khaDga-dhArA-parAkramah//
    CitraM yad-ari-nArINAM hRdaye virah-AnalaH /
    ajasram=azru-pAnIya-sicyamAno=pi varddhate//
    BhaGgo=ntaHpurik-AlakeSu surata-krIDAsu keza-grahaH kAThinyaM kucayor=bhruvoH...

    emphasis supplied: NAnyaura Plate A of Dhanga, dated equivalent to November 6, 998; trans. F. Kielhorn, Three Chandella Copper-Plate Grants Indian Antiquary vol. XVI (1887), pp. 203, 204.

  82. For an excellent introduction to tantra in all its diversity, see Mircea Eliade, Yoga: Immortality and Freedom 2nd ed., (New York: Bollingen Foundation, 1969), chapters, 6-8.

  83. Namely the vertical sets of two each on the anatArala walls of the LakSman, and three each at the same locations on the VizvanAth and KandAriyA MahAdeo

  84. queen, goddess, is all the same--Everywoman.

  85. For citations on the former, Stella Kramrisch, The Hindu Temple, vol. 2, p. 347.

  86. Burton trans., p. 94.

  87. For a readily accessible commentary, Heinrich Zimmer, Myths and Symbols in Indian Art and Civilization ed.Joseph Campbell (New York: Harper Torchbooks, 1946).

  88. Kenneth Clark, The Nude: A Study in Ideal Form (New York: Doubleday, 1956), p. 130.

  89. Tell-tale picchikA, or rajoharaNa, whisk topped with peacock feathers, in hand.

  90. Somadeva Bhatta (11th cent.), Kathasaritsagara. i.e., The Ocean of Story, C. H. Tawney's trans., ed. N. M. Penzer (Delhi : Motilal Banarsidass, [1968]), vol. IX, pp. 68-70.

  91. Whence Dhanga's NAnyaura copper plate A was issued in 998; F. Kielhorn, Three Chandella Copper-Plate Grants Indian Antiquary vol. XVI (1887), pp. 201-204. Bengal-based PAlas were also active in the area as were the Kalachuris, especially under GangeyAdeva and Karnadeva in the mid-11th century (as noted earlier, no. 33).

  92. Zilpa PrakAza, II. 538, 539; trans. p. 106.

  93. The Pali Text Society's Pali-English Dictionary ed. T. W. Rhys Davids and William Stede (Chipstead, Surrey: Pali Text Society, 1925), 4 v (in 1), p. 89.

  94. Such is the title of a sole recovered chapter from the SiddhArtha pRicchA, as reported in the Gujarati journal Svadhyaya vol. V: 2 (1967-68): 191-198, by M.A. Dhaky and P.O. Sompura; cited in passing by Krishna Deva in Temples of Khajuraho vol. 1, p. 33.

  95. With emphasis supplied, to the phrase gahanottArArthinA.
    Which of the Khajuraho temples he dedicated to Siva as [healer] VaidyanAth is unknown: the inscription is now preserved in the VizvaNath, attributed by most to king Dhanga; F. Kielhorn, Inscriptions from Khajuraho, no. V, Epigraphia Indica, vol. I (1890-91): 152.

  96. Emphasis supplied to the phrase, yat=svarggaM vrajatas tadIya-yasa(za)saH sopAna-mArggAyate; R.D. Banerji, The Gurgi Inscription of Prabodhasiva, Epigraphia Indica vol. 22 (1933-34), v. 11, p.133.

  97. For full citations on the retention of sexual capacity in heaven, from the Rg and Atharva Vedas, Zatapatha BrAhmaNa and the UpaniSads, etc., John Muir, Further quotations from the hymns on the subject of paradise and future punishment, Original Sanskrit Texts (London: 1868-1873; reprint 3rd ed., Amsterdam: Oriental Press, 1967) vol. 5, pp. 305-313.

  98. VarApsaraH-sahasrANI zUram Ayodhane hatam / tvaramANA'bhidhAvanti mama bhartA bhaved iti //
    MahAbhArata XII. 3657; Muir trans, p. 308.

  99. Above, n.39.