Pandukeshwar, Architectural Knowledge, and an Idea of India

Abstract

This essay probes a temple complex at Pandukeshwar—a sacred center near a glacial source of the Ganga River in the Central Himalayas—where lithic shrines in local, regional, and transregional architectural typologies stand side by side. I first descriptively analyze the forms and layouts of these edifices, then connect the archaeological evidence to the historical record to date the shrines. Thereafter, I show how their design, construction, and use appears to be connected to the emergence, refinement, and dispersion of an idea of India as a geo-cultural landscape extending from the Indian Ocean up to the high Himalayas. This conception, I conclude, was transmitted along knowledge corridors, routes that connected institutions of learning.

Introduction

Pandukeshwar Village is an extraordinary place. It is situated in a Central Himalayan gorge near the glacial source of the mighty Alakananda-Ganga River (map 1).[1] It is also the last hamlet in Uttarakhand State in India to host a year-round population on a perilous ancient route that leads from the lush northern plains to Mana Pass, an entry into the frosty Tibetan plateau. Four early medieval lithic temples stand at Pandukeshwar (fig. 1). One temple, honoring Viṣṇu as Yogabadrī, is built in a typology of the Drāviḍa or southern Indian mode. The other three temples, honoring Gaṇeśa and Viṣṇu in his manifestations as Vāsudeva-Badrī and Lakṣmī-Nārāyaṇa, are built in architectural typologies that reference the Nāgara or northern Indian mode in dissimilar ways. Enshrined in these temples are gold and copper images of the Vaiṣṇava pantheon and four copperplate inscriptions. The later documents are written in Sanskrit and are datable to the ninth and tenth centuries.

Map 1. Places of historical and cultural significance in the Central Himalayas.

1. (left to right) Lakṣmī Nārāyaṇa temple, Yogabadrī temple, Gaṇeśa shrine (this diminutive monument abuts the Yogabadrī), and Nārāyaṇa temple, Pandukeshwar, Uttarakhand

Unless otherwise noted, all maps, plans, and photographs are by the author

To find such a variety of material remains at Pandukeshwar is astounding for at least three reasons. First, even today it is dangerous and difficult to reach this far-flung settlement.[2] Snowfall frequently obstructs access during winter months. Flashfloods and landslides triggered by cloudbursts periodically prevent visitors from reaching the village at other times of the year.[3] Second, despite the abundance of lithic monuments in India, very rarely does one find medieval temples built in the Drāviḍa and Nāgara modes standing side by side. In fact, Mahakuta and Pattadakal—the closest settlements where medieval temples in these modes can be seen next to one another—are located more than 2,800 kilometers away in Karnataka State in peninsular India (map 2). Third, despite the fact that the four temples at Pandukeshwar are visited by thousands of Hindu pilgrims annually and have been protected by the Archaeological Survey of India (ASI) since 1942, until now they have not been recognized as belonging to different modes of Hindu temple architecture. On the contrary, they have been regarded as specimens of Buddhist architecture and have been compared to edifices in such places as Nalanda in northeastern India.[4]

Map 2. Places where medieval temples in Nāgara and Drāviḍa modes stand side by side (italicized and in gray). Key sites traditionally associated with Ādī Śaṅkara’s life are in black

The quality and diversity of architectural modes and typologies at Pandukeshwar raise a series of questions. When and under what circumstances did the more broadly applicable knowledge of lithic temple construction and the more specialized knowledge of the Drāviḍa and Nāgara modes reach the Central Himalayas at large and this village in particular? How was this knowledge received and deployed? How did the accretion of monuments affect religious practices? Who were the original and subsequent patrons, builders, and users, and how were they connected to each other? How, if at all, was Pandukeshwar related to other fords (tīrthas) and sacred centers?[5] And finally, how do temple construction and the expansion of religious networks in the Central Himalayas intersect with social, political, and cultural changes elsewhere in South Asia?

I adopt an integrated method to investigate these questions. Given the extreme paucity of written records at Pandukeshwar and elsewhere in the Central Himalayas, it is the material record—built, landscaped, and sculpted—that constitutes the bulk of the historical evidence. Therefore, I carefully record and analyze this material evidence by (a) making maps that visualize spatial relationships among and between settlements; (b) drawing plans showing the orientation and arrangement of parts of individual temples; (c) using a Sanskritic terminology to understand temple forms; (d) employing precise language to record relationships between plans and elevations, external structures and internal systems, and original temple fabrics and later additions; and (e) forging a relative chronology. In addition, I look at epigraphs, oral narratives, toponyms, and contemporary practices. I use analytical tools devised by historians and anthropologists to decode the aforesaid items. I record geographical features that have had an impact on the locations of tīrthas and movements of people. I consider bridle paths that connect hill settlements to one another as well as cart roads that lead in and out of the mountain range. I also survey artistic accomplishments and political processes of earlier epochs and other parts of South Asia that influenced cultural production in the early medieval Central Himalayas.[6] Finally, I look briefly at the cultural expressions and political movements of subsequent centuries to understand how they may complicate my interpretations.

Using this methodology, I recover important moments in the history of Pandukeshwar and its region. I also cast new light on the transmission of architectural knowledge in early medieval India. This phenomenon, I posit, shaped Central Himalayan communities and reflected their variable aspirations. It also solidified a long-lived idea about India that continues to resonate with its residents today—that it is a cohesive geo-cultural entity that extends from the sandy shores of the Indian Ocean to the snow-clad Himalayas.

Temples and Communities in Early India: Plotting Trends and Transformations

To understand the forms, layouts, and functions of Pandukeshwar’s temples, we must review the genesis and development of stone temple architecture in early India. Stone temples, a new technology that connected Hindus to divine forces, began to be constructed in India in the early fifth century CE. Located in the fertile plains in the middle of the subcontinent, the earliest temples consisted of small, cave-like sanctums (garbhagṛhas) preceded by shallow porticos. The garbhagṛhas housed either aniconic or anthropomorphized images of deities. The porticos accommodated devotees. In the fifth century, these devotees were likely members of esoteric cults rather than the lay populace. Goaded by ascendant dynasties eager to accrue merit and preceptors seeking to strengthen authority, the interest in erecting temples spread from middle India to other regions. By the seventh century, temples were being constructed in distant corners of the subcontinent. In addition, two distinct architectural modes, the Drāviḍa and Nāgara, had emerged.

In Tamil Nadu in southern India, the early Drāviḍa formula consisted of a shrine (vimāna) composed of a molded platform beneath the floor level (adhisṭhāna), masonry walls punctuated by pilasters and occasionally niches, a prominent eave (kapoṭa) adorned with semicircular window motifs (kuḍus), and one of two kinds of superstructure. The superstructure might be composed of a balustrade (hāra) and a diminutive cottage with a domical roof, as rendered in a famous seventh-century relief at the port town of Mamallapuram (fig. 2).[7] Or it might consist of several stories (talās)—with balconies enclosed by railings, aedicular cottages (kutas), and vaulted halls (śālas)—surmounted by a small cottage with a domical roof, as exemplified by a nearly contemporaneous temple by the seashore in the same town (fig. 3). In contrast, the early Nāgara formula consisted of a square shrine (mūlaprāsāda) with central offsets (bhadras) surmounted by a curvilinear tower (latina śikhara). At the well-preserved Parasurameśvara temple, built in the seventh century in Bhubaneshwar, Orissa, the latina śikhara consists of compacted cornices, neckings, projecting central bands (madhyalatās) accentuating the plan’s bhadras, cogged discs at the corners (karṇāmalakas) and a crowning cogged disc (āmalaka) (fig. 4). In subsequent decades, the Drāviḍa became the preferred mode of builders in southern India who saw it as an apt symbol for a “new” temple-centered Hinduism and territorial sovereignty. Meanwhile, patrons and masons in northern India enthusiastically embraced the Nāgara as a sign of their autonomy. At Mahakuta and Pattadakal, Drāviḍa and Nāgara monuments were erected besides each other in the eighth century; the pairings possibly signaled the strategic marital alliances and extravagant political ambitions of their patrons, the Cālukyas (figs. 5, 6).[8]

2. Great Narrative Relief, Mamallapuram, Tamil Nadu

3. Seashore temple, Mamallapuram

4. Parasurameśvara temple, Bhubhaneshwar, Orissa. Photo courtesy of the American Council of Southern Asian Art (ACSAA) Collection, University of Michigan, History of Art Department, Visual Resource Collections

5. Nāgara and Drāviḍa temples, Mahakuta, Karnataka. Photo courtesy of the American Council of Southern Asian Art (ACSAA) Collection, University of Michigan, History of Art Department, Visual Resource Collections

6. Nāgara and Drāviḍa temples, Pattadakal, Karantaka. Photo courtesy of the American Council of Southern Asian Art (ACSAA) Collection, University of Michigan, History of Art Department, Visual Resource Collections

To appreciate the forms of the Drāviḍa and Nāgara temples at Pandukeshwar, glean when they were built, and elucidate what they signaled, it is important to summarize formative political processes and cultural trends in this mountainous frontier. In the third century BCE, the installation of a monumental rock edict brought Kalsi Village—where the Yamuna River enters the plains—into the ambit of the Mauryan Empire. Shortly thereafter, events such as the importation of large red sandstone sculptures of Śiva and Pārvatī to Rishikesh, where the Ganga leaves the mountains and enters the plains, deepened contact between the foothills and the “Hinduizing” civilizations of upper India.[9] In time, the riverways that connected Kalsi to Mathura and Rishikesh to Pāṭaliputra drew Śaiva and Śākta pilgrims and traders toward the high Himalayas and the Tibetan plateau. Over time, these visitors contributed to the gradual transformation of Kalsi, Rishikesh, and nearby Haridwar into trading posts, gateway towns, and tīrthas in their own right.

Conquests also played their part in establishing enduring settlements, spreading Brahmanical ideologies, and fostering a sense of regional identity. As the hold of the Kūṣāna rulers over the fertile Doab weakened, minor princes expanded their territories and performed legitimizing Aśvamedha yajñas (Vedic sacrificial rituals), as Sīlavarmaṇa did at Jagatgram near Kalsi.[10] The imperial Guptas and their armies—who ruled over much of northern India between the fourth and sixth centuries—did not reach the Himalayas. However, some architectural styles and aesthetic conventions that developed in the Gangetic heartland during their reign did make inroads along the upper courses of the Ganga and Yamuna. One such monument is a seventh-century brick temple at Koteshwar, sited just upstream from the celebrated confluence of the Alakananda-Ganga and Mandakini. Perhaps this temple’s large front hall was built to accommodate pilgrims.

Other seventh-century temples in the Central Himalayas seem to have been the preserve of specialized cults and their royal patrons. At Lakhamandal, on the Yamuna’s upper reaches, a preceptor honored the memory of widowed hill queen’s husband by asking her to build a lofty brick-and-stone terraced podium with a columnar shaft connoting Śiva’s presence (liṅga) at its summit.[11] A panegyric (praśasti) celebrating the queen’s virtues also records that this ensemble was designed by Īśvaranāga, an architect (sūtradhāra) from Rhotak in Haryana, about two hundred kilometers downstream from Lakhamandal. The panegyric also identifies Īśvaranāga as its composer. Exotic architectural projects and literary compositions—occasionally imported to the Central Himalayas by the same individual—appear not only to have enjoyed respect in the region but also to have spurred local artisans to erect lithic monuments. In the decades after the terrace temple’s construction, local artisans constructed small shrines nearby. These edifices have simple base moldings, small square sanctums, and plain walls. Their tiered superstructures are articulated like cornices and neckings and crowned by āmalakas. This architectural formula quickly came to be associated with the Central Himalayas.

In the eighth and the ninth centuries, builders at incipient tīrthas in the Central Himalayas constructed subsidiary shrines and commemorative monuments using the recently devised architectural formula. A few latina Nāgara temples also were built in the region. At Lakhamandal and Jageshwar, the latina Nāgara mode was selected for Śiva temples and the regional typology reserved for funerary monuments. And at Paithani, the latina Nāgara mode was used for the principal temple at the center of a quincunx complex (pañcāyatana) and the regional typology adopted for the four corner shrines (fig. 7). The previously mentioned and other latina Nāgara temples in the Central Himalayas appear to have been designed by émigré architect-priests with greater knowledge of how to use the Vāstupuruṣamaṇḍala to control the sanctum’s proportions and generate its plan and elevation.[12] The names of the patrons who commissioned these lithic edifices have not come down to us. However, architectural sculpture such as frontons (śukanāsas) bearing relief carvings of Lākuliśa, the deified founder of Pāsupata Śaivism, and statuary such as columnar shafts bearing Śiva’s four faces (caturmukhaliṅgas) at many of these sites suggests the influence of Pāśupata Śaiva ascetics.[13]

7. Pañcāyatana temple complex, Paithani, Uttarakhand

Vāstu Vidyā at Pandukeshwar

In the ninth and tenth centuries, communities in the Central Himalayas were smaller and socially less striated than many others in the subcontinent.[14] Extra hands were needed for building and repairing everything from irrigation ducts to the retaining walls of terrace fields and from cow pens to private residences. Therefore, most householders in the region were, in part, seasonal builders.

A scrutiny of the four temples erected at Pandukeshwar during this period indicates that the possession of Vāstu Vidyā, an integrative body of architectural theory, distinguished their builders from those who periodically provided hands for construction projects. Vāstu Vidyā included everything from site selection to the consecration of the ground and from plan development to the erection of robust three-dimensional spaces. Although comprehensive in its scope, Vāstu Vidyā was capable of being enlarged and adapted by its holders.

The location of the temples within Pandukeshwar indicates what their builders knew about site potentials. It is a short walk from the riverbank, the destination of many pilgrims visiting the tīrtha. At the same time, it is significantly removed from the torrential river’s pathway. Furthermore, like the gently swelling shell of the cosmic turtle, described in classical Sanskrit literature as bearing the world’s weight, the site’s center faintly rises above its edge.[15] Therefore, neither snowmelt nor rainfall accumulates in this sanctified space. Moreover, the surface of the earth where the temples stand is firm—strong enough to support the load of several buildings, yet porous and fertile enough to allow seedlings to break through the ground and grow into healthy plants.

The temples’ carefully calibrated square plans divulge other details about the builders’ capacities. These skills likely included observing the sun’s movement across the sky and using the resulting knowledge to generate cardinally orientated square plans. A general understanding and acceptance of Hindu associations between and among cardinal directions and planets as well as the square form and the characteristics of the ordered cosmos must have accompanied such knowledge. It is difficult to establish how much builders at Pandukeshwar knew about the closely allied concept of the Vāstupuruṣamaṇḍala—with its intricate subdivisions, imagery of the body, and the significance of its lines and their points to intersection. What is clear is that, unlike many of their contemporaries, they did not use it to precisely control plans and elevations.

As suggested by the millennium-old stands of Himalayan cedar (Cedrus deodara) at Jageshwar and elsewhere in the Central Himalayas, timber has long been an abundant resource in the region.[16] Therefore, the consistent use of a grayish stone for the fabric of the temples and black schist for architectural sculpture, rather than wood, is noteworthy. Both stones are smooth, and neither is fragmented by deep veins nor flaking at the edges, which shows that builders knew how to obtain and work with particular stones. Such understanding may have been accompanied by the knowledge of chants and rituals associated with quarrying, hauling, dressing, and laying stone courses and with the purported merits of building with that material.[17] Finally, the decision to construct using as such an enduring material shows the builders’ engagement with the emphasis on the eternal nature of Hindu deities that is found in many genres of Sanskrit literature.[18]

Finally, the Pandukeshwar temple morphologies—and measurements taken at the khura course of their base molding sequences and at the point where their wall friezes begin—lead to the conclusion that their designers and executors were knowledgeable about alignment, classification, juxtaposition, and proportion. The first shrine encountered by devotees entering the sacred enclosure is aptly dedicated to Gaṇeśa, the remover of obstacles (figs. 1, 8). Like subsidiary shrines at Paithani and votive monuments at Jageshwar, the east-facing Gaṇeśa shrine at Pandukeshwar has a truncated vedībandha (base molding unit), walls composed of unornamented single slabs of stone, and a superstructure made up of progressively diminishing slabs whose edges are articulated as cornices and neckings. A large āmalaka and a pot (kalaśa) crowns the formation. Furthermore, like the subsidiary shrines and votive monuments that orbit and shield the largest temples at the previously mentioned tīrthas, Pandukeshwar’s Gaṇeśa shrine is set just left of the entrance of the Yogabadrī temple, the biggest and most important monument in the village.[19]

8. Gaṇeśa shrine, Pandukeshwar

The Yogabadrī temple is so different from the Gaṇeśa shrine and all other monuments in the Central Himalayas that it baffled Kanti Prasad Nautiyal, the sole scholar to write about it (figs. 1, 9, 10). He wrote that it was “either influenced by the Buddhist stupa type because of the proximity of the place to the Tibetan region or it may have been just a reproduction of the Pala temple.”[20] Close examination of the Yogabadrī temple’s structural components, however, shows its affiliation with temples in southern India, suggesting that it is a Drāviḍa alpa vimāna.[21] For reasons that I shall explain shortly, I date it to circa 850–1000 CE.

9. Detail, Yogabadrī temple, Pandukeshwar

10. Schematic plan, drawn at floor and wall frieze levels, Yogabadrī temple, Pandukeshwar

The foundational moldings of Yogabadrī’s square-plan vimāna appear in the following sequence: minor projecting molding (kṣudropāna),[22] lotus-formed molding (padmopāna), recessed band (antarita), tri-faced torus molding (tripaṭakumuda), recessed molding (golapāda) with block-projections (galas)[23] and with crocodilian creatures (makaras) only at the corners, upper-fillet molding (ūrdhvakampa), minor molding (upāna), and oversailing fillet (mahāvājan). In contemporaneous Drāviḍa alpa vimānas in Tamil Nadu, this sequence constitutes the molded base (adhisthāna), and it invariably is set beneath the garbhagṛha’s floor level. Furthermore, in Tamil Nadu, a water chute (praṇāla) typically cuts through the mahāvājan or equivalent crowning course. The Yogabadrī temple’s floor level, however, is sited in the middle of this set of molding courses, a feature externally indicated by the water chute that cuts through the tripaṭakumuda.[24]

Such variances apart, the Yogabadrī temple compares well with better-known alpa vimānas erected at villages such as Viralur, Nartamallai, and Kalayadipatti in Tamil Nadu, in the ninth and tenth centuries (figs. 11–13).[25] A wall frieze punctuated by rather plain pilasters rises above the temple’s foundational mahāvājana course. The uppermost portion of the wall frieze features a row of galas, which are sheltered by a heavy kapoṭa decorated with gavākṣas (dormer-window motifs) with trefoil finials. Sculptures of sejant lions are set on the four corners of the plank molding (pratis). Calling to mind a sculptural typology and an architectural convention first seen at the seashore temple at Mamallapuram, they symbolically protect the temple, weigh down a course carved with a diamond-and-pearl pattern, and safeguard the vulnerable junctures of walls (figs. 14–17).

11. Śiva temple, Visalur, Tamil Nadu

12. Ancillary shrine, Nartamallai, Tamil Nadu

13. Shrine, Kalayadipatti, Tamil Nadu

14. Sejant lion, Yogabadrī temple, Pandukeshwar

15. Sejant lion, seashore temple, Mamallapuram

16. Sejant lion, Chettiapatti, Tamil Nadu. Photo courtesy of the American Institute of Indian Studies, Gurgaon

17. Horned lion (vyāla), Tamil Nadu. Photo courtesy of the Metropolitan Museum of Art, New York

A tall dome rises behind the lions. More cylindrical in shape than the circular and octagonal domes of alpa vimānas in Tamil Nadu, it features a necking (grīva), wide flange, and tall, arched window-motifs (kuḍus) crested with heads of composite beings (kirtīmukhas) from whose mouths fluids and flames cascade down. At each kuḍu’s base is a pedestal. It is not possible to determine whether these pedestals ever bore sculpted images of gods (grīva devatas), as was customary in alpa vimānas in Tamil Nadu.[26] Lastly, such alpa vimānas were crowned with jar finials (stūpīs). As a large umbrella-like wooden canopy (chattra) currently surmounts the Yogabadrī temple’s dome, it is not possible to determine whether it too originally had a stūpī.

A rectangular, enclosed front hall (guḍhamaṇḍapa) fronts the vimāna. It shows the same base molding courses as the vimāna, but its wall frieze is plainer. Beaten copper sheets clad an older roof composed of sloping stone slabs.[27] A short flight of stairs contained by a banister—reminiscent of the curvaceous balustrades of so many Drāviḍa temples in Tamil Nadu—leads devotees toward an unornamented portal. Stepping through this portal, one enters a dark hall, with a door at its far end leading into the garbhagṛha. A cluster of hitherto unpublished metal images of Viṣṇu and his celestial retainers are enshrined in this exalted space (fig. 18). The largest of these images, a golden sculpture depicting Viṣṇu seated in a cross-legged meditative posture (padmasana), appears to date from the time of the temple’s construction in the ninth or tenth century.

18. Garbhagṛha deities, Yogabadrī temple, Pandukeshwar

The mulaprāsāda and guḍhamaṇḍapa of a latina Nāgara temple with two planes of offset (dvi-aṅga) stand to the Yogabadrī temple’s left.[28] Today, this Nāgara temple is dedicated to Nārāyaṇa (figs. 1, 19–21). I tentatively date the oldest portions of the Nārāyaṇa temple’s mulaprāsāda to either the tenth or the eleventh century, as it appears to have been self-consciously paired with the Yogabadrī’s vīmāna.

19. Nārāyaṇa temple, Pandukeshwar

20. Yogabadrī temple (foreground), Nārāyaṇa temple (background), Pandukeshwar

21. Schematic plan, drawn at floor and wall frieze levels, Nārāyaṇa temple, Pandukeshwar

The Nārāyaṇa temple’s vedībandha comprises the customary sequence of courses seen on Nāgara temples. However, the distinctive characteristics of the individual courses are attenuated. The minimally projecting kapota course barely resembles a sheltering awning. And the blocked-out dormer-window motifs on it bear greater resemblance to the galas on the Yogabadrī temple’s adhisṭhāna than to gavākṣa motifs that customarily occupied this position. The Nārāyaṇa temple’s wall frieze (jaṅghā) also shows such modifications. It is unadorned by stacks of richly carved pilasters and stringcourses. Wall niches (rathikās) crowned by tall pediments (uḍgamas), flanked by pillarets and relief carvings of celestial maidens (apsarās), seers (munis), composite leonine creatures (vyālas)—all frequently seen on latina Nāgara temples—are also absent. On the contrary, its surface features barely projecting pilasters that echo those on the walls of the Yogabadrī. Rising above the jaṅghā is a cyma recta transition zone(varaṇḍikā) bearing a latina śikhara. The śikhara’s mismatched stones, abnormal taper, and oversized chattra all point to a reconstruction.[29] The temple’s śukanāsa in contrast, has survived well. It features a fierce masculine face carved of black schist, the same stone as the Yogabadrī’s ferocious lions. Finally, the Nārāyaṇa temple is also fronted by a plain guḍhamaṇḍapa with a pitched roof.

Standing near the Nārāyaṇa temple is the north-facing Lakṣmī-Nārāyaṇa temple (figs. 1, 22). This unevenly preserved monument consists of a mūlaprāsāda preceded by a kapilī (porch). Having two planes in plan and elevation, the mūlaprāsāda rises above a socle (pitha). Above it is the vedībandha composed of the hoof (khura), pitcher (kumbha), pot (kalaśa), and kapoṭa courses. The jaṅghā is punctuated by a deep rathikā framed by pillarets supporting a ribbed awning (khurachādya) and bifurcated by a stringcourse. Filled with a row of diamond and orb-shaped medallions, this course girdles the temple and helps date it to no earlier than the late ninth century.[30] A curvilinear latina śikhara, springs off the jaṅghā. However, its damaged state and heavy-handed reconstruction make it impossible to describe it too carefully.[31]

22. Lakṣmī Nārāyaṇa temple, Pandukeshwar

The lacunae of the recent reconstruction not withstanding, it is possible to glean something of the artisans’ abilities and objectives by examining those portions of the original fabric that have survived in fairly good repair. The pīṭha, half-diamond (ardha-ratna) motifs emblazoned on the khura, composite pillarets, and khurachādya framing the rathikā all establish the builders’ awareness of ornament associated with Nāgara temples in central and western India. Meanwhile, the diamond and orb-shaped motifs and some other elements confirm the builders concurrent interest in visually linking this temple to those already erected at Pandukeshwar.

An inspection of a hitherto unpublished copper image of Nārāyaṇa enshrined in the Nārāyaṇa temple’s sanctum illuminates our understanding of the capacities of metalworkers active around Pandukeshwar during the decades when the temples were constructed (fig. 23). About forty-two inches high, this image depicts the deity with a cone-shaped crown standing upright. His raised rear hands hold a conch and discus; his lowered front hands bear a lotus and a mace. Matched makara earrings emerge and disappear into wavy locks. The bare upper body, adorned with a jeweled necklace, contrasts with the lower body, which is wrapped in a dhoti with cascading, crescent-shaped folds. A parikara (frame) with makara-topped columns is set around the figure. A fiery arch springs from the wavelike tips of makara tails. Formally, the Nārāyaṇa icon and its parikara compare nicely with two tenth- or eleventh-century metal sculptures of the same deity said to be from the western Himalayas.[32] This said, certain details of this statue recall traits of metal mūrtīs cast elsewhere. Reminiscent of Vaiṣṇava images from Bengal are the sensuous fingers of the raised rear arms and the treatment of the parikara. Faintly recalling visions of portable icons (calamūrtis) from medieval Tamil Nadu are the necklaces, a diaphragm band intersecting the sacred thread (yajnopavīta), and the mace’s position.

23. Nārāyaṇa, garbhagṛha, Nārāyaṇa temple, Pandukeshwar

The metalworkers remarkable capabilities and their interest in citing, juxtaposing, and habilitating local, regional, and transregional forms and techniques parallels those of the master-masons and stone carvers. Where did all these artisans acquire their abilities and proclivities? And why did they, and their patrons, choose to display their accomplishments in this tiny Himalayan village over all other places? In the next two sections, I probe these questions by melding the geographical, literary, inscriptional, and archaeological record.

The Ganga’s Descent and the Drāviḍa’s Ascent

Pandukeshwar’s geographical location and cultural position are just as striking as the range and fineness of its temples and sculptures. The village is situated near a ravine’s end. From there, a road winds its way up a mountain and then down into an elevated alpine valley.[33] Bisecting the valley is the Alakananda-Ganga, a torrent born from the mouths of glaciers and rivulets trickling down from the surrounding snow-capped peaks. The seasonally inhabited hamlet of Badrinath is sited at the alpine valley’s center near four steaming sulfur pools. Meadows—sometimes carpeted with snow and at other times with wild flowers, medicinal herbs, aromatic shrubs, and herds of domesticated yak—encircle Badrinath. As a watercolor painted on-site makes clear, the sole trail leading northward and out of Badrinath initially veers past corrals of yak herders before meandering through ice-laced slopes and mounds of moraine. It eventually reaches Mana Pass (fig. 24),[34] through which the watercolor’s painter crossed into the windswept Tibetan plateau. Mana Pass has long been a corridor for the movement of men and goods.[35]

24. Hyder Young Hearsey (1782–1840), A View of the Badrinath Valley, 1808. Photo courtesy of the British Library, London

Badrinath (ancient Badrī, Badrikā, or Badrikāśrama) has occupied an immensely important place in the Indic imaginary for millennia. In the Mahābhārata, a Sanskrit epic redacted between 400 BCE and 400 CE, Badrinath is celebrated as the site where Nara, the primordial man, and Nārāyaṇa performed a protracted penance side by side.[36] According to the epic, many ages later when Nara was reborn as the Pāṇḍava Arjuna, he returned to the Ganga’s upper reaches. There, after an intense combat with Śiva—in his guise as a Kirāṭa, a wild man of the mountains—Arjuna received a formidable weapon. While Arjuna was away, his brothers and co-wife gradually made their way to nearby Badrinath, where they received the blessings of assembled hermits. They then dispelled the fatigue that they had accumulated during their travels to countless tīrthas by wandering around Badrinath’s sublime environs. These landscapes, in the epic poet’s understanding, were adorned with trees bearing lush fruits, expansive snowfields, mountains streaked with luminescent minerals, and cascading streams.[37]

From the post-Gupta period onward, the events narrated in the Mahābhārata were elaborated and recounted in the Purāṇas, courtly poems (mahākāvyas), and sculpted reliefs. Furthermore, Badrinath came be celebrated as the stage for more wondrous events. For instance, the redactors of the Skanda Purāṇa identified this tīrtha as the place where the legendary sages Nārada and Mārkaṇḍeya gained an auspicious sighting (darśana) of their cherished deity and where Viṣṇu as Varāha and Narasiṁha retreated after performing heroic deeds.[38] The Ganga’s upper reaches and Badrinath also feature in the Kirāṭārjunīya, Bhāravi’s (sixth-century) acclaimed poetic transcreation of Arjuna’s combat with Śiva.[39] It is also plausible that a large relief carved on the eastern wall of the sixth-century temple at Deogarh in the Gangetic plains depicts Nārāyaṇa instructing Nara at Badrinath.[40] Finally, it is likely that a scene included in the enormous seventh-century relief at Mamallapuram—featuring a preceptor and his students studying on the banks of a cascading river in the foreground, a Drāviḍa alpa vimāna honoring Viṣṇu in the middle ground, and ascetics performing penances in the background—is, in fact, a rendition of Badrinath and its environs (fig. 2).[41]

Neither archaeological remains nor inscriptional evidence assignable to the very centuries in which Indian sculptors and versifiers were valorizing Badrinath have survived in the alpine valley. However, documents preserved elsewhere establish that the village has attracted travelers for centuries.

Captain Raper, a British explorer who reached Badrinath in 1808 along with the previously mentioned watercolor painter, calculated that at the time some 45,000 to 50,000 pilgrims were visiting the tīrtha annually. Raper added that “the greater part of these were fakirs who came from the most remote quarters of India.”[42] That number had risen to 175,000 by 1975. A decade before this count, the ethnographer Surinder Bhardhwaj found that the average pilgrim to Badrinath traveled 789 kilometers from his or her hometown.[43] According to records kept by the temple committee, 981,000 Hindu pilgrims visited the tīrtha in 2011.[44]

Finally, it should be noted that at least since the late nineteenth century—when colonial authorities began documenting activities at Badrinath—local residents have recounted legends of the great South Indian Advaita philosopher Ādī Śaṅkara’s (788–820 CE?) visit to Badrinath.[45] One priestly account, first recorded by Edwin T. Atkinson, author of a pioneering study on Himalayan religion, runs as follows:

Another priestly account first recorded by Atkinson—and still communicated to pilgrims visiting Badrinath today—also credits the philosopher with the restoration of tīrtha’s ancient Hindu temples.[47] According to yet another local legend, the hill ruler Kanak Pāla helped Śaṅkara to expel Buddhism from the region and erect the Badrinath temple.[48]

Badrinath is not the sole Himalayan settlement associated with the great philosopher. Like their forefathers, the current residents of Joshimath (ancient Jyotirmaṭha)—a large village perched on a slope just south of Pandukeshwar—speak about how Śaṅkara established a maṭha (monastic center of learning) there to propagate his teachings. Furthermore, like many other modern Hindus, Joshimath’s residents believe that Saṅkara established at least three other maṭhas: at Puri in coastal Orissa, at Dwaraka in marine Gujarat, and at Kanchipuram in the Kaveri delta in Tamil Nadu.[49] Today, these monastic institutions are directed by Saṅkarācaryas, pontiffs who have taken the early medieval philosopher’s name, and are visited by thousands of Hindu pilgrims during their circumambulation of the land. [50]

In considering the previously mentioned scenarios, it is important to bear in mind that the great philosopher wrote nothing about his travels. However, it is equally important to recall that one of the few points of agreement in all seven premodern Sanskrit hagiographies of Saṅkara is his pilgrimage to Badrinath. All seven hagiographers report that upon reaching at Badrinath, Śaṅkara heroically defeated resident sages in debate, vanquished Buddhist heretics, restored the tīrtha’s purity, and wrote commentaries on theological texts.[51] The veracity of episodes recounted in these hagiographies cannot be independently confirmed, as they were composed many centuries after the philosopher’s life. Yet the possibility of Śaṅkara’s epic journey to the four quarters of India remains quite likely. As Siddha Kuśaladeva’s biography attests, learned individuals from South India did undertake long-distance journeys to the Himalayas despite uncertainties and privations.[52]

The earliest plausible record of an association between Pandukeshwar and Badrinath is found in an inscribed copperplate discovered in the mid-nineteenth century at the Nārāyaṇa temple. Bearing a date corresponding to 853–54 CE, this inscription directs priests based at Garuḍagrāma to aid celibates (brahmacārīs) at Badrinath. Given this copperplate’s findspot and Pandukeshwar’s position in a ravine at the threshold of the elevated valley, it is likely that the village historically occupied a subordinate position to Badrinath and was named for Garuda, Viṣṇu’s loyal attendant. Pandukeshwar’s stature has grown in the millennium since this order was issued. Today, Namboodri Brahmins from Kerala, Śaṅkara’s home state in southern India, officiate at its temples. In the summer, the Namboodris attempt to relax the pace of pilgrims hurrying on to Badrinath by recounting a meandering narrative about how Arjuna’s father, Pāṇḍu, performed a penance there in the guise of a deer. Furthermore, the Namboodris effectively transform Pandukeshwar into Badrinath each winter. This conversion begins at winter’s onset when the principal image of Viṣṇu in the Badrinath temple is laid in a supine position and a calamūrti is brought out of the temple in a palanquin. The temple doors are then locked, and the calamūrti is transported in fanfare to Pandukeshwar. Accommodated in the Yogabadrī temple’s crowded sanctum, it is regularly worshipped until its return to Badrinath at the start of summer, when the doors are unlocked, and the supine image is restored to an upright position and ritually reawakened.

It is undeniable that this custom parallels the seasonal migrations of other deities as well as administrators, pastoralists, priests, and traders in other Himalayan districts.[53] It is also evident that Pandukeshwar has enjoyed an enviable location on a transregional knowledge corridor—a passageway that connected far-flung places of learning—at least since the construction of its temples.[54] Still, questions remain. Who took the awareness of diverse architectural forms to Pandukeshwar? And who paid for the construction and upkeep of these monuments?

One possibility is that the Pandukeshwar temples were built by a hill dynasty who concurrently wished to celebrate and draw people to a spot sited at Badrinath’s threshold into their little kingdom, welcome the many and varied pilgrims passing by, and accommodate their equally diverse ritual practices. Keeping this scenario in mind, it is worth asking if Tamil architects (sthāpatis) and skilled builders, who occasionally traveled to distant lands to design temples for special patrons,[55] could have built the Yogabadrī temple in its entirety. Several conceptual, formal, and aesthetic discrepancies between “proper” Drāviḍa vimānas in Tamil Nadu and the vimāna at Pandukeshwar, however, make this possibility unlikely.

Another possibility is that an energetic, learned, and determined North Indian sūtradhāra conceived of pairing a Drāviḍa vimāna and a matched Nāgara mūlaprāsāda and enlisted the services of Central Himalayan artisans to construct them. Even if this scenario seems feasible, it is important to ask just how this sūtradhāra and the master masons, stone carvers, and laborers working under his supervision learned about the Drāviḍa typology.

It is tempting to imagine that the sūtradhāra was familiar with the general appearance of a South Indian temple from compendia that predated the encyclopedic Samrāṅgaṇasūtradhāra and like it had chapters on the construction of Drāviḍa and Nāgara temples.[56] However, this proposition has problematic aspects. No such ninth- to tenth-century śilpa śāstra compendium has come down to us. And even if one assumes that such works did exist, it is hard to know whether its prescriptions—with all their idiosyncrasies and contradictions—would have been intelligible to a sūtradhāra more familiar with Nāgara conventions. Furthermore, as śilpa śāstra texts tend to say little about anatomical articulation, construction techniques, and other important aspects, one wonders whether a sūtradhāra sitting in windswept Pandukeshwar systematically read them.[57]

Given Pandukeshwar’s position on a knowledge corridor, it is possible the sūtradhāra learned about the general appearance of a South Indian temple and its characteristic proportions from discourses given by Joshimath’s erudite monks; conversations with Badrinath’s priests from the South who were connected to Śaṅkarā; or even chance encounters with well-traveled pilgrims undertaking epic journeys to tīrthas in India’s four corners.[58] From encounters with such individuals, the sūtradhāra may have also learned something about listing, ranking, and scaling different types of temples, kings, and subjects as well as about linguistic, numerical, and categorical homologies connecting them to subjects as diverse as astronomy, medicine, and prosody.

The previously mentioned encounters also may have brought the sūtradhāra and artisans—who already possessed a wide repertoire of skills—into contact with models of different types of temples and icons (fig. 25). These models most likely would have taken the forms of miniature shrines and diminutive images of deities with which pilgrims sometimes traveled rather than purpose-built, scaled replicas of existing or potential buildings.[59] Indeed, the more private practice of carrying miniature shrines and icons of one’s favorite deities on a pilgrimage and the more public ritual of processing a mobile image in a palanquin around a sacred landscape continue to the present day.[60]

25. Medieval “model” of a Drāviḍa vimāna. Photo courtesy of the Ashmolean Museum, University of Oxford

In all, such activities allow me to reconsider the precision of a conclusion once reached by architectural historian M. A. Dhaky:

Great Temples, Little Kingdoms

Shifting from the production to consumption, we know neither the names of individuals who maintained and visited temples in and around Pandukeshwar nor what rituals were performed there in the ninth and tenth centuries. However, the construction of front halls (maṇḍapas) and an increase in the number of metal statues of various deities indicate greater pilgrim traffic, a diversification of rituals, and more wealth in the hands of these institutions.

Archaeologists have found only five documents from this moment in the history of the Central Himalayas.[62] Four of these records are written in Sanskrit on rectangular copperplates and are preserved at Pandukeshwar. The fifth one, written in Sanskrit and engraved on a stone slab, is kept at Bageshwar. Taken together, the documents provide important, if partial, insights into new patterns of patronage and ritual that complement findings recoverable from the foregoing study of the archaeological evidence. These records establish that the cult of Viṣṇu—especially one centered close to the glacial source of the Alakananda-Ganga—received support from three short-lived hill dynasties rather than from the Pāśupatas who had been especially influential in the Central Himalayas for several centuries. These three dynasties ruled from the ceremonial capital of Kārttikeyapura, possibly in the Katyur Valley. Tribhuvanapāla, the scion of one royal house, installed an image of Vaikunṭha Viṣṇu in a temple in the valley in 1002 CE.[63] Another hill dynasty included King Lalitasuradeva and his consort Sāmādevī. When Sāmādevi built a Nārāyaṇa temple at Gorrunasāri, her husband gifted the revenues of various villages to it and to the Nārāyaṇa temple at Garuḍagrāma (Pandukeshwar?).[64] A third lineage, which initially reigned from Kārttikeyapura and later from Subhikṣapura, included Padmaṭadeva and his son and successor, Subhikṣarājadeva. Father and son lent their support to the Badrinath temple.[65]

Earlier rulers, such as Princess Īśvarā of Lakhamandal who had enjoyed some authority over the Central Himalayas, had more interest in erecting Śiva temples than in making provisions for their long-term ritual and physical maintenance. In contrast, the copperplate inscriptions of later Himalayan leaders, including Lalitasuradeva, Padmaṭadeva, and Subhikṣarājadeva, indicate a personal interest in such activities as “providing perfumes, flowers, incense, lights, ointments, offerings of eatables, sacrifices, oblations of rice, &c., dancing, singing, music, charities, &c., for the repair of what may be damaged or broken, as well as for the execution of new work, and for the maintenance of servants and attendants.”[66] In other words, these rulers were interested in properly establishing Viṣṇu as a glorious sentient being with various requirements and expectations in this region.

The rulers of the little hill kingdoms of the medieval period also tried to selectively foster new networks between tīrthas. In one edict, Lalitasuradeva specifically directed the brahmacāris of Garuḍagrāma, the beneficiaries of his largesse, to aid brahmacarīs living at Badrinath. Such a royal intervention stands in contrast with the direction and nature of earlier networks such as those that existed between Jageshwar and other places in its vicinity and likely were maintained by the relatively unceremonious movements of ascetics and masons.[67]

The four copperplates preserved at Pandukeshwar are also all indicators of the transformative political developments that had occurred by the time of their composition. For although these metal documents record only the names of three or four prior generations of the donor’s family, they demonstrate a secondary phase of state formation. In this phase, previously measured and owned lands in well-established settlements were transferred to authorities either in the same or a different district. Judging from the functionaries listed on the copperplate inscriptions, this was also a period by which several courts—complete with hierarchies of aides, ministers, military commanders, poets, scribes, and guild foremen—had been established.

Finally, the foregoing discussion makes it possible to compare the rise of “little kingdoms” and pilgrimage circuits in the Central Himalayas with those in the Western Himalayas from the tenth until the thirteenth centuries. In this regard, the history of the erstwhile princely state of Chamba in the Western Himalayas makes for an especially illuminating comparison. In the mid-tenth century, a minor chieftain named Sāhilla Varman moved the ceremonial capital of the old and periodically conquered principality of Brahmapura to Caṃpakapuri (Chamba). From this new capital, Sāhilla Varman and his descendants enacted measures to maintain their authority and their territory’s sovereignty. Gathering revenues from traders using mountain passes and from agriculturalists working in fertile river valleys, they supported new Nāgara temples dedicated to Lakṣmī and Nārāyaṇa more enthusiastically than old shrines honoring Śiva and Śaktī. They also created roads and waterworks that connected them with their loyalists. Furthermore, Chamba’s fledging rulers increased their stature by initiating matrimonial alliances with the kings of Kashmir, their more powerful neighbors, and by undertaking journeys to renowned tīrthas such as Kurukshetra in the northern Indian plains. They recruited Brahmins from the plains to serve as their educators, genealogists, and ritual specialists, and they liberally awarded parcels of land to them. They also formed small courts, with officers whose grandiloquent titles conformed to the titles of the courtiers of the great empires of medieval India.[68] After their hold over the “little kingdom” was secure, Chamba’s kings periodically traveled to Haridwar to immerse the ashes of their deceased kin in the Ganga’s waters and make presents to needy Brahmins. From Haridwar, a few kings proceeded to Badrinath and Kedarnath with the objectives of washing away old sins, presenting jewels to the enshrined deities, and receiving divine grace.[69]

In subsequent centuries, residents of the area on the route to Badrinath and Kedarnath appear to have followed Pandukeshwar’s lead in redefining their settlements: molding landscape features, localizing events from the Mahābhārata, and habilitating distinctive architectural forms.[70] For example, modern pilgrims passing through Guptakashi, a hamlet on the way to Kedarnath, can bathe in Maṇikarnika, a stepped tank that shares its name with a celebrated set of stairs leading to the Ganga River (ghāṭ) at Kashi, the paragon of Śaiva tīrthas. At Guptakashi, they also can hear legends of the Pāṇḍavas’ escapades from Namboodri priests. Entry into a shrine resembling a Drāvida alpa vīmāna, which is set besides a latina Nāgara mūlaprāsāda, is also open to them (fig. 26). To give another example, in recent decades, two small Drāviḍa alpa vimānas have been erected at Rishikesh, besides the frequently repaired early medieval Bharata shrine (fig. 27). This cluster is sited just across the river from an imposing temple dedicated to Bhārata Mātā.

26. Nāgara and Drāviḍa temples, Guptakashi, Uttarakhand

27. Nāgara and Drāviḍa temples, Rishikesh, Uttarakhand

An Idea of India

Housing a theomorphic icon of the nation state, Rishikesh’s Bhārata Mātā temple is a modern building and a reflection of a modern belief.[71] Whether India ever existed as a unified entity prior to its independence from British rule in 1947 remains an open question. Like many Victorian scholars, John Strachey was convinced that “This is the first and most essential thing to learn about India—that there is not, and never was an India, or even any country of India, possessing according to European ideas, any sort of unity, physical, political, social, and religious; no Indian nation, no ‘people of India’ of which we hear so much.”[72] Echoes of such assertions continue to be heard. Only recently a postcolonial historian observed, “Ironically, therefore, the territory that we use to describe the landscape of Indian civilization was defined politically by the British Empire. India was never what it is today in a geographical, demographic, or cultural sense before 1947.”[73]

The temples and sculptures at Pandukeshwar suggest that by the tenth or eleventh century, a concerted effort had begun to reshape this far-flung hamlet in a harsh landscape into an anchor of an evolving idea of India: as a cohesive geo-cultural entity extending from the shores of the Indian Ocean up to the snow-capped Himalayan peaks. Mendicants, priests, philosophers, and pilgrims had introduced architectural theory to the village and encouraged resourceful and ambitious hill builders to erect an array of monuments. As well-engineered lithic structures, these monuments could concretely mark a site of growing importance almost for perpetuity. As architectural spaces, they could allow builders not just to posses the small plots on which they stood, but to extend their sphere of influence over land coveted also by Bhotiyas and others living on the fringes of the trans-Himalayan Tibetan plateau. As edifices honoring Viṣṇu—a deity who was increasingly a focus of widespread devotion—these monuments could spur individuals to forge and maintain alliances with neighbors, kinsmen, and people and congregations from distant lands. Moreover, as fully formed Brahmanical temples, rather than cenotaphs honoring deceased rulers or antipodean ascetics, the monuments could attract, store, and redistribute wealth in a comparatively impoverished realm.

At Pandukeshwar, each monument was built according to a mode or typology associated with one or another part of South Asia. Each monument, therefore, had the potential to reassure a weary visitor and perhaps even kindle a desire to settle in this region. As an ensemble of modes and typologies erected in a polyglot frontier, the monuments had the capability to visually and publically communicate a conception of India to travelers in a way could transcend the confusion of tongues. This said, as all the monuments were rooted in a comprehensive yet adaptable and extendable corpus of architectural theory and a thriving building tradition, they could allow builders and users to engage imaginatively with the previously mentioned idea.

The strategies enacted at Pandukeshwar to increase this status of this village, sanctify this region, and popularize a notion of India—among groups as varied as pilgrims undertaking grand journeys of the subcontinent’s four quarters; mendicants and traders journeying to and from India; and rulers other than those who patronized Pandukeshwar’s temples—appear to have been successful. Possibly, in acknowledgement of the region’s increasingly diverse resident and mobile population, hill rulers began to address not only Khaśas and Kirāṭas who had lived in these mountains for centuries but also travelers and émigrés from distant lands.[74] For instance, the copperplate issued by Lalitasuradeva is directed to

Today, more than a millennium later, members of many of these ethnicities and communities continue to undertake journeys to fords in this mountainous frontier.[76]

Nachiket Chanchani, PhD (University of Pennsylvania), 2012, is assistant professor of South Asian art and visual culture in the Departments of the History of Art and Asian Languages and Cultures at the University of Michigan, Ann Arbor. He has published articles on the linguistically hybrid and profusely painted hand-scrolls and manuscripts of pre-Mughal Western India, on the history of photography, on art writing and translation, and on the formulation of cultural policy. Chanchani is currently completing a monograph on the construction of sacrality in the Central Himalayas. E-mail: [email protected]

Author’s note: For their strategic guidance, I thank M. A. Dhaky, John Stratton Hawley, Darielle Mason, Michael W. Meister, and Romila Thapar. For facilitating my field research, I thank Kapila Vatsyayan, successive directors general of the Archaeological Survey of India, and the ASI’s officers. I am indebted to my family for their enthusiastic interest in this research project and for offering counsel at critical junctures. I am grateful to my brother, Pranav, for his assistance with the preparation of the maps.

In this essay, Sanskrit terms and ancient proper names—including names of historical characters, temples, and cultural regions—have been transliterated using diacritical marks established by the IAST scheme. However, to facilitate readability, the names of modern personages and toponyms have neither been rendered with diacritical marks (e.g., Sharma and Rishikesh rather than Śarmā and Hṛṣīkeśa) nor “Sankritized” (e.g. Mamallapuram rather than Mahābalipuram). Technical Sanskrit terms are briefly translated, e.g., guḍhamaṇḍapa (enclosed front hall).

Notes

1. The Ganga River has no one source. Two rivers—the Alakananda, which begins at Satopanth Glacier above Pandukeshwar, and the Bhagirathi, which commences at Gaumukh Glacier about one hundred kilometers northwest of Pandukeshwar—converge at Devaprayag to form the Ganga proper.

2. Pandukeshwar may be reached from Haridwar, Almora, or Tibet. The Haridwar route follows the Alakananda-Ganga upstream past Devaprayag, Koteshwar, and Joshimath. Travelers from Almora can reach Pandukeshwar via Baijnath, Gwaldam, and Joshimath. Travelers from the Tibetan plateau can get there by crossing the Mana Pass.

3. In June 2013 alone, for instance, more than 5,500 people perished in flashfloods and landslides in the Central Himalayas. Although no causalities from Pandukeshwar were reported, floodwaters swept away numerous cottages and destroyed roadways, leaving scores of villagers homeless and at least two thousand pilgrims stranded. For media reports on hardships encountered by local residents and the dramatic rescue of pilgrims, see Zee News, “Army evacuates 2000 People from Badrinath,” http://zeenews.india.com/news/uttarakhand/army-evacuates-2000-stranded-people-from-badrinath_858051.html, accessed November 11, 2013; “Uttarakhand flood: Rain, another cloudburst hits rescue ops, just 1,000 evacuated,” The Indian Express, June 24, 2013; and G. P. Semwal, “Flood Tragedy: Pilgrims Recall Nightmare, Rue State Inaction,” The Pioneer, June 25, 2013.

4. For such an understanding see Kanti Prasad Nautiyal, The Archaeology of Kumaon (Varanasi: Chowkhamba Sanskrit Library, 1969), 91.

5. For an overview of the development of this concept, see Diana L. Eck, “India’s Tirthas: ‘Crossings’ in Sacred Geography,” History of Religions 20, no. 4 (1981), 323–44.

6. The term “early medieval” is derived from European historiography where it once had pejorative connotations. However, it has been cautiously used and developed by scholars of South Asia to designate the period extending from the decline of the Gupta dynasty in the sixth century to the rise of the Delhi Sultanate in the thirteenth century. This period was characterized by the assertion of regional polities and by the increasing importance of temples in religious rituals. For a broad overview of these processes, see Brajdulal Chattopadhyaya, The Making of Early Medieval India (New Delhi: Oxford, 1997).

7. Such shrines are termed as alpa vimānas.

8. Over the long seventh century, successive Calukya rulers considerably expanded their empire’s extent by waging war against their eastern and western neighbors and obstructing the southward movement of northern India’s most powerful rulers. They also attempted to bring portions of South Asia under their sway. For analyses of morphologies and patronage structures of the temples at Mahakuta and Pattadakal, see Susan Buchanan, “Calukya Temples: History and Iconography” (PhD diss., Ohio State University, 1985), and George Michell, Pattadakal (New Delhi: Oxford University Press, 2002).

9. For details on these sculptures, see W. H. Siddiqi, “Two Newly Discovered Pre-Kuṣāṇa Sculptures from Ṛṣikeśa (Uttar Pradesh),” Journal of the Indian Society of Oriental Art, Dr. Moti Chandra Commemoration Volume n.s. (1978), 76–80.

10. For the excavator’s report, see T. N. Ramachandran, “Aśvamedha Site near Kalsi,” Journal of Oriental Research, 21 (1953), 1–31.

11. For a full translation of the inscription, which is still lying at Lakhamandal, see G. Bühler, “The Praśasti of the Temple of Lakkhā Maṇḍal at Maḍhā in Jaunsār Bāwar,” Epigraphia Indica 1 (1892), 10–15.

12. For a survey of the use of Vāstupuruṣamaṇdala in developing temple plans and elevations, see Michael W. Meister, “Vāstupuruṣamaṇdalas: Planning in the Image of Man,” in Maṇḍalas and Yantras in the Hindu Traditions, ed. Gudrun Buhnemann (Leiden: Brill, 2003), 251–90.

13. For a discussion on the development of architectural form at Jageshwar and an analysis on the activities of Pāśupata Śaivas at this sacred center, see Nachiket Chanchani, “The Jageshwar Valley: Where Death is Conquered,” Archives of Asian Art 63, no. 2 (2013), 133–54.

14. For details, see Marie Lecomte-Tilouine, ed., Bards and Mediums: History, Culture, and Politics in the Central Himalayan Kingdoms (Almora: Almora Book Depot, 2009), and Ramachandran Guha, The Unquiet Woods: Ecological Change and Peasant Resistance in the Himalaya (Delhi: Oxford University Press, 1989).

15. For details on the turtle’s symbolism, see Devangana Desai, “Kūrma Imagery in Indian Art and Culture,” Artibus Asiae 69, no. 2 (2009).

16. For a discussion on these forests, see Chanchani, “Jageshwar Valley.”

17. The great prestige associated with the construction of a stone temple is evinced not only by noting the large number of medieval lithic temples standing across India even today, but also by studying texts. For example, the Mahānirvāṇa Tantra notes that it is “100 times more meritorious to give a brick temple than a thatched temple; 10,000 times more meritorious to give a stone than a brick temple.” Quoted in Stella Kramrisch, The Hindu Temple, vol. 1 (Calcutta: University of Calcutta Press, 1946), 113.

18. For example, Avyayaḥ (undecaying), Akṣaraḥ (indestructible), Sthānuh (immovable), Dhruvaḥ (indestructible), are among the one thousand names of Viṣṇu anthologized in the Viṣṇusahasranāma portion of the Mahābhārata. For a full listing of names, see Śrī Viṣṇusahasranāma Stotram with a commentary by Āḍi Śaṅkara (Chennai: Ramakrishna Mission, 1986).

19. To the right of the Yogabadrī temple’s entrance is a tiny makeshift structure composed of a few stone slabs. It is impossible to say when this structure was erected. However, it is apparent that it was intended to endow the complex with a greater semblance of symmetry and possibly transform a four-shrine cluster into a pañcāyatana one.

20. Nautiyal, Archaeology of Kumaon, 91.

21. Because of this, I use appropriate Drāvida terminology to document and analyze its form.

22. A recent repaving of this courtyard has almost entirely obliterated the Yogabadrī temple’s kṣudropāna course. It is visible in older photographs of the temple such as AIIS photographic archive negative number 499.61, accessible at http://dsal.uchicago.edu/images/aiis/aiis_search.html?depth=Get+Details&id=57921.

23. The block projections of the golāpada course of ninth- and tenth-century alpa vimānas in Tamil Nadu are generally carved with animal motifs. Plain galas, likely a more inexpensive option, are occasionally seen on Tamil monuments, including the Manonamaṇīśvara temple at Vijaynarayanam, assignable to the late ninth century on the basis of a donative inscription. For a discussion of this temple, see K. V. Soundara Rajan, “Middle Pāṇḍīnāḍu style, ca. A.D. 850–950: Pāṇḍyas of Madurai, Phase II,” in Encyclopedia of Indian Temple Architecture: Southern India: Lower Drāvḍadeśa, 200B.C.–A.D. 1324, ed. Michael W. Meister (Philadelphia: University of Pennsylvania Press, 1983), 113–15.

24. For the important conceptual and formal distinction between a vedībandha and an adhiṣṭhāna, see Michael W. Meister, “Reading Monuments and Seeing Texts,” in Śāstric Traditions in Indian Arts, ed. A. Dallapiccola. (Stuttgart: Franz Steiner Verlag, 1989), 167–73, pls. 94–108 and pls. 81–88.

25. For discussions on these Drāviḍa monuments and their plausible patrons, see K. V. Soundara Rajan, “Early Pāṇḍya, Muttarayar and Irukkuvel Architecture,” in Studies in Indian Temple Architecture, ed. Pramod Chandra (New Delhi: American Institute of Indian Studies, 1975), 240–300, and K. V. Soundara Rajan, “Early Cola Nāḍu Style, c. A.D. 800–1000: Muttarayars of Neman and Sendai,” in Encyclopedia of Indian Temple Architecture: South India, 124–44.

26. Even as the Yogabadrī’s builders gave it the look of an alpa vimāna, they reduced the possibility of its collapse in an earthquake-prone area. Instead of punctuating the sanctum’s walls with niches, or puncturing them with windows, they fortified them with tightly bonded courses. By setting progressively receding courses above the sanctum ceiling’s slabs, placing sturdy sculptures of lions on their corners, and giving the dome a wide flange, they dispersed its weight outward rather than downward.

27. The original form of the guḍhamaṇḍapa’s roof is observable in older photographs, such as AIIS photographic archive, accession number 57917.

28. Because of this, I use appropriate Nāgara terminology to document and analyze its form.

29. In a 1930 photograph, the Nārāyaṇa temple’s śikhara looks almost as it does today. This establishes that the śikhara was reconstructed prior to ASI protection.

30. The convention of using a stringcourse to divide the jaṅghā horizontally appears to have made its first appearance at the Lālguan Mahādeva temple built in mid-ninth century at Khajuraho, Madhya Pradesh, in central India. This date helps assign a lower limit for that temple’s date. The convention grew in popularity in middle India during the tenth century, when it was used to create two tiered walls for stacking rows of sculpture.

31. Not visible in photographs of Pandukeshwar taken before 2011, the Lakṣmī Nārāyaṇa temple’s recent history is shrouded in mystery. Some villagers and ASI officers say that it lay buried under an ancient rubble mound at the temple cluster. Others hold that was concealed within a modern residence that has since been demolished.

32. One of these sculptures has been published in Pratapaditya Pal, ed., Himalayas: An Aesthetic Exploration (Chicago: Art Institute of Chicago, 2003), pl. 12; the other in the art dealer John Siudmak’s 2012 auction catalogue Indian and Himalayan Sculpture, pl. 26; http://www.johnsiudmak.com/indian_himalayan_2012.php.

33. The Yogabadrī and Nārāyaṇa temples face the mountain road leading to alpine Badrinath.

34. Hyder Young Hearsey (1782–1840), an artist in the entourage of the explorers Captain F. V. Raper and Lieutenant W. S. Webb, painted this hitherto-unpublished watercolor on-site on May 29, 1808. Today, the watercolor is in the collections of the British Library, London.

35. For a discussion of Indo-Tibetan trade through Mana Pass, see Maheshwar P. Joshi and C. W. Brown, “Some Dynamics of Indo-Tibetan Trade through Uttarākhanḍa (Kumaon-Garhwal), India,” Journal of the Economic and Social History of the Orient 30, no. 3 (1987), 303–17.

36. J. A. B. van Buitenen, “The Mahābhārata: An Introduction,” The Mahābhārata 1: The Book of the Beginning, ed. and trans. J. A. B. van Buitenen (Chicago: University of Chicago Press, 1973), xxv. For an alternate date, see Alf Hiltebeitel, Rethinking the Mahabharata: A Reader’s Guide to the Education of the Dharma King (Chicago: University of Chicago Press, 2001), 18ff.

37. For details of Nara and Nārāyaṇa austerities, Arjuna’s and Kirāṭa’s combat, and the Pandavas’ exhausting journey to Badrinath and their blissful stay in its environs, see The Mahābhārata 2: Book 2: The Book of the Assembly Hall; Book 3: The Book of the Forest, ed. and trans. J. A. B. van Buitenen, (Chicago: University of Chicago Press, 1975).

38. G. P. Bhatt, ed., and G. V. Tagare, trans., The Skanda Purāṇa, vol. 2, no. 3 (Delhi: Motilal Banarasidass, 1992), 3–4.

39. Romesh Chandra Dutt, Selections of Indian Poetry Rendered into English Verse (London: Kegan Paul, Trench and Trubner Ltd., 1894), 201. For an analysis of this mahākavya, see Indira Viswanathan Peterson, Design and Rhetoric in a Sanskrit Court Epic: The Kirāṭārjunīya of Bhāravi (Albany: SUNY Press, 2003).

40. For a photograph of this sculpture and a discussion of its architectural context, see Joanna Gottfried Williams, The Art of Gupta India: Empire and Province (Princeton: Princeton University Press, 1982), 130–37, and pl. 206.

41. This is the thesis of T. N. Ramachandran’s article, “The Kirāṭārjunīyam or ‘Arjuna’s Penance’ in Indian Art,” Journal of the Indian Society for Oriental Art 18 (1950–51), i–iii. For alternative readings, see Ananda K. Coomaraswamy, History of Indian and Indonesian Art (1927, repr. New York: Dover, 1985), 103, and Padma Kaimal, “Playful Ambiguity and Political Authority in the Large Relief at Mamallapuram,” Ars Orientalis 24 (1994), 1–27.

42. For Raper’s account, see James Ballie Fraser’s Journal of a tour through part of the snowy range of the Himalaya mountains, and to the sources of the Yamuna and Ganga (Delhi: Neeraj Publishing House, 1820 [1982]), 379.

43. See Surinder Bhardhwaj, Hindu Places of Pilgrimage in India (Berkeley: University of California Press, 1973), 118–27.

44. Badri Kedar Temple Committee, The Official Website of the Shri Badrinath Shri Kedarnath Temples Committee,” http://www.badarikedar.org, accessed November 10, 2013.

45. The most recent discussion of the probable dates of Ādī Śaṅkara’s life are found in Matthew Clark, The Dasnami Samnayasis: The Integration of Ascetic Lineages into an Order (Leiden: Brill, 2006).

46. Edwin T. Atkinson, Notes on the History of Religion in the Himalaya of the N.W.P. India ([Allahabad?], 1893), 85.

47. Atkinson, Religion in the Himalaya, 70.

48. Guha, Unquiet Woods, 64.

49. Some modern Hindus also believe that Saṅkarā founded a maṭha at Shringeri in Karnataka and consecrated Śiva temples at Jageshwar and Kedarnath in the Central Himalayas.

50. Clark—who doubts that Saṅkarā established the monastic institutions that today claim to have been founded by him—explains how in the modern period successive Saṇkarācāryas based at Joshimath have competed for authority over and acceptance by the spiritual leaders of peer institutions. For details, see Clark, Dasnami Samnayasis, 77–78, 118–19, and 143–44.

51. For a scholarly consideration of seven influential precolonial Sanskrit hagiographies and other sources for Ādī Śaṅkara’s life, see Jonathan Bader, Conquest of the Four Quarters: Traditional Accounts of the Life of Śaṅkarā (New Delhi: Aditya Prakashan, 2000), especially chap. 1, 2, and 5. Also pertinent is W. R. Antakar’s “Śaṅkara-Vijaya of Anantānandagiri” Journal of the University of Bombay 30 (1961), 73–80.

52. Kuśaladeva, known in Tibetan annals as Mahāsiddha Padampa Sangye, was one of these learned individuals. He was born into a Brahmin family in Andhra Pradesh in southern India. From there he traveled to the Himalayas before making several forays into the Tibetan plateau to teach. His lessons were so enthusiastically received that Kuśaladeva eventually agreed to live close to his Tibetan hosts until his death in 1117. For a biography of his life and a fourteenth-century portrait of him from the western Himalayas, see Dan Martin, “Padampa Sangye: A History of Representation of a South Indian Siddha in Tibet,” in Holy Madness: Portraits of Tantric Siddhas, ed. Rob Linrothe (New York: Rubin Museum of Art, 2006), 108–23.

53. For analyses of such migrations, see Shiv Darshan Pant, Social Economy of the Himalayans (London: Allen and Unwin, 1935), and William Sax, Mountain Goddess: Gender and Politics in a Himalayan Pilgrimage (New York: Oxford University Press, 1991).

54. This passageway may well have connected monastic institutions at Khojarnath and Tholing in the Tibetan plateau, at Joshimath in the Central Himalayas, and at Haridwar, Kashi, and Prayag downstream in the Gangetic plains.

55. Between the ninth and thirteenth centuries, teams of itinerant Tamil builders appear to have worked alongside local masons to erect several small Drāvidan alpa vimānas at places as distant as Quanzhou in China. For a discussion of a Śiva temple at this Chinese port city, see Risha Lee, “Rethinking Community: The Indic Carvings of Quanzhou,” in Nagapattinam to Suvarnadwipa: Reflections of the Chola Naval Expeditions to Southeast Asia, ed. Hermann Kulke et al. (Singapore: Institute of Southeast Asian Studies, 2009), 240–70. For a recent review of the transmission of Indian culture to Southeast Asia, see John Guy, ed., Lost Kingdoms: Hindu-Buddhist Sculpture of Early Southeast Asia (New Haven: Yale University Press, 2014).

56. Chapters 61 and 62 of the Samrāṅgaṇasūtradhāra, an eleventh-century compendium attributed to Rājā Bhoja of Dhar, list the characteristics of Drāviḍa temples, suggesting that some theoretical knowledge of this mode was known in courtly circles in middle India. For a discussion of these chapters, see Adam Hardy, “Drāviḍa Temples in the Samarāṅgasūtradhāra,” South Asian Studies 25 (2010), 41–62.

57. In the few known instances where particular texts did play a significant role in determining a Drāviḍa temple’s morphology, they were redacted close to the place were the monument was erected. The Sundarāvaradāperumalla temple, constructed in the early ninth century at Uttarameru in Tamil Nadu, is a case in point. A foundational inscription explains how Parameśvara Peruthachchan, its architect, cautiously adapted prescriptions of local āgamas and vāstu śāstras to generate the plan, elevation, and sculptural program. For details see V. Ganapathi Sthapati, “An Interesting Inscription from Uttarameru,” Seminar on Inscriptions, 1966: Speeches and Papers (Madras: Department of Archaeology, 1968), 178–88.

58. Tripurāṃtaka, who lived in the thirteenth century, was one such erudite traveler. After completing his education in Gujarat, he undertook an epic pilgrimage to sacred centers in India’s four corners. For a discussion of his itinerary and motivations, see Nachiket Chanchani, “From Asoda to Almora, the Road Less Taken: Māru-Gurjara Architecture in the Central Himalayas,” Arts Asiatiques 69 (2014), 3–16.

59. Dimunitive medieval South Indian bronzes of Drāviḍa vimānas and sejant lions are preserved today in museum collections, including the Ashmolean Museum at the University of Oxford (fig. 25) and the Metropolitan Museum of Art in New York (fig. 16). Although I have not encountered such statuettes in the Central Himalayas, I know that several medieval Kerala bronzes of Vaiṣṇava deities have been excavated in this region.

60. I have encountered both practices along old pilgrim routes in the Central Himalayas. For estimates of such processions, see Sax, Mountain Goddess, and Alka Hingorani, Making Faces: Self and Image Creation in a Himalayan Valley (Honolulu: University of Hawai‘i Press, 2012).

61. M. A. Dhaky, The Indian Temple Forms in Karṇāṭa Inscriptions and Architecture (New Delhi: Abhinav Publications, 1977), 41–42.

62. These inscriptions give the regnal years of the kings who issued them, but not the Vikram Samvat year in which they were issued. F. Kielhorn, who translated Lalitasuradeva’s charter, the oldest of four copperplates preserved in Pandukeshwar, melded astrological information given in the inscription with a paleographic analysis of the script to posit that it was issued on December 22, 853 CE. His well-informed speculation has led subsequent scholars to date remaining inscriptions preserved at the site to the ninth and tenth centuries CE. For details, see F. Kielhorn, “Pandukesvara Plate of Lalitasuradeva,” Indian Antiquary 25 (1896), 177–84.

63. Annual Report on Indian Epigraphy (New Delhi: Archaeological Survey of India, 1960–61), 113.

64. See Kielhorn, “Pandukesvara Plate,” and D. C. Sircar, “Three Plates from Pandukesvar,” Epigraphia Indica 31 (1956), 277–98.

65. For a transcription and preliminary translation of Padmaṭadeva’s and Subhikṣarājadeva’s copper plates preserved at Pandukesvar, see Sircar, “Three Plates from Pandukesvar.” The dvi-aṅga latina Nāgara temple that stands behind the brightly painted gatehouse at Badrinath has been rebuilt so many times since the reigns of Padmaṭadeva and Subhikṣarājadeva that it is difficult now to conjecture its medieval appearance.

66. Kielhorn, “Pandukesvara Plate,” 183, and Sircar, “Three Plates from Pandukesvar,” 283, 290.

67. Jageshwar and other Pāśupata centers in the Central Himalayas have not yielded inscriptions or other written documents from the early medieval period that testify to their patronage by early medieval rulers.

68. For a more detailed reconstruction of Chamba’s history, see Mahesh Sharma, Western Himalayan Temple Records: State, Pilgrimage, Ritual and Legality in Chambā (Leiden: Brill, 2009), 19–64.

69. Epigraphic records indicate that from the early thirteenth century onward, the rājās of Chamba and other Hindu kingdoms in the western Himalayas began to visit Haridwar, Badrinath, and Kedaranath. For the motivations of some of these rulers and their activities at these pilgrimage centers, see Bühler, “The Two Praśastīs of Baijnāth,” 97–118, as well as “Mhesa Plate of Bhot Varman” (pp. 23–24), “Lakṣmī-Nārāyaṇa Temple Plate of Gaṇeśa Varman’s Time” (pp. 48–50); and “Chamba Plate of Pratāpsiṃha” (pp. 73–75) in Antiquities of Chamba State: Part II: Medieval and Later Inscriptions, ed. Bahadurchand Chahbra (1957; repr. New Delhi: Archaeological Survey of India, 1994).

70. For an account of the position of the Mahābharata in the lives of communities living in the region today, see William S. Sax, Dancing the Self: Person and Performance in the Pāṇdav Līlā of Garhwal (New York: Oxford University Press, 2002).

71. For an account on the emergence and development of the idea and iconography of Bhārata Mātā, see Lise McKean, “Bhārata Mātā: Mother India and Her Militant Matriots,” in Devi: Goddesses of India, ed. John Stratton Hawley and Donna Marie Wulff (Berkeley: University of California Press, 1996), 250–80, and Sumathi Ramaswamy, The Goddess and the Nation: Mapping Mother India, (Durham: Duke University Press, 2010).

72. John Strachey, India (London: Kegan, Paul, Trench & Co., 1888), 2.

73. David Ludden, “Introduction: Ayodhya: A Window on the World,” in Contesting the Nation: Religion, Community, and the Politics of Democracy in India, ed. David Ludden (Philadelphia: University of Pennsylvania Press, 1996), 6. For an incisive critique of this perception, see Diana L. Eck, India: A Sacred Geography (New York: Harmony Books, 2012).

74. M. C. Joshi, “The Khasas in the History of the Uttakarkhand,” in Himalaya Past and Present, vol. 3, ed. M. P. Joshi, A. C. Fanger, and C. W. Brown (Almora: Almora Book Depot, 1990), 193–200.

75. Sircar, “Three Plates from Pandukesvar,” 283. The three other copper plate inscriptions from Pandukeshwar also contain similar lists. Although such listings of ethnicities and communities had become a standard component of eulogies and land grants composed in Sanskrit in the Indian plains by the ninth century, they are conspicuous by their absence in the Central Himalayas prior to this historical moment.

76. See Bhardhwaj, Hindu Places; Sondra L. Hausner, Wandering with Sadhus: Ascetics in the Hindu Himalayas (Bloomington: Indiana University Press, 2007); Andrea M. Pinkey, “The Sacred Share: Prasāda in South Asia” (PhD diss., Columbia University, 2008); and James Lochtefeld, God’s Gateway: Identity and Meaning in a Hindu Pilgrimage Place (New York: Oxford University Press, 2010).