Even the ‘unfinished’ Buddhist, Hindu and Jain temple caves in Ellora have stories to tell

It is likely to come as a surprise to many to hear that the rock-cut site of Ellora, renowned for the grandeur of its richly sculpted monuments dedicated variously to the Buddhist, Hindu, and Jain faiths, carries a considerable degree of “unfinished” work. What exactly do we mean in this context by “finish” or the lack thereof? Finish implies firstly that none of the component parts of a monument are lacking or deficient. A finished monument is thus one which has all of its requisite parts – gateways, halls, porches, shrine, roofs, towers – and which depicts carved imagery appropriate to its dedication to one or other deity. Finish also implies that the various portions of a monument, and all its accompanying decorative and figural carvings, are crafted to a state of refinement so that nothing rough or inadequate remains. Recurring instances of lack of finish in both these senses are evident at Ellora, and they occur across the site and in caves dedicated to all three faiths. There is, for instance, a cave in which rock-cutters have not yet reached the stage of creating uniformly vertical walls or flat horizontal ceilings, but in which sculpted images have already been commenced. There are caves that are usable, but which reveal major areas of unfinished sculptural and architectural work. And there are caves that are almost complete but lack the element of refinement and finish in several details.

Why should “unfinished” work matter to us or be important in any way? Firstly, the substantial body of unfinished monuments at Ellora, and elsewhere in India, demands that we reassess pre-modern India’s attitude towards “finish.” The modern eye is so focused on finished work, especially in the case of the richly decorated monuments of India, that we tend to easily overlook the lack of finish. More often than not, such unfinished work may not be attributed to geological reasons such as bad strata of rock, or the sudden appearance of an underground spring of water, nor entirely to external factors such as natural disasters, wars, the death of a monarch, or the bankruptcy of a patron. It would appear rather that we need to come to terms with the fact that pre-modern India had a remarkably flexible and pragmatic approach towards the concept of finish.

I hope to demonstrate that once the shrine of a rock-cut sacred monument was ready for consecration and worship, patrons and worshipers appear to have found it unimportant, even irrelevant, to concern themselves with the detailed finish of the overall structure. As long as the sanctum was complete and subsidiary sacred images and/or sacred myths were clearly decipherable by devotees, it mattered little if figures within a panel remained somewhat incomplete, or if the panel’s framing pilasters were roughly cut. In Buddhist, Hindu, and Jain contexts, the patron who initiated the work, the rock-cutters who extracted the structure from a mountainside, the stone carvers who worked on the decorative details, as well as the devotees who worshiped therein, seem to have been able to ignore the fact that pillars remain incomplete, walls carry rough tool marks, and images remain partly obscured by the living rock. The very use of the word “unfinished” requires rethinking as it applies to pre-modern India, and this in turn necessitates a critical evaluation of what constituted an acceptable “sacred space.”

Another reason to study unfinished work is that it is a statement of process, emphasising that stone carving is a subtractive process in which the artist/craftsperson is faced with a large mass of stone from which they have to remove matter to create their final product. Anything that is inadvertently or carelessly cut away cannot be put back or remodelled. Because of this, stone workers have always developed processes by which they gradually work upon and into the stone, thinking their way, one might say, towards the final surface. The process of stone carving has been likened to peeling an onion.1 Artists carve stone away in layers, and when work remains unfinished, one may trace the process involved in carving. At Ellora, there is visible evidence of the three sets of tools that they used. The preliminary tool is a pointed axe and/ or a pointed chisel with which 85% of the stone removed is cut away, leaving behind long, smooth lines as evidence of its work. All finer work was done with either a flat-headed or a rounded chisel, after which abrasives were used to create a smooth and even surface on the stone. Unfinished work also provides clear evidence of the top-down process of carving, in which the upper portions of a carved panel, an image, or a wall tend to be completed to a fine degree of finish, while the lower areas are often untouched. Presumably, rock-cutters developed this system because it minimized the use of scaffolding in the preliminary phase of work; in sculptural work, completing the head and face first would have enabled the artist to visualise clearly the proportionality of the figure.

It is useful to emphasise that what we call rock-cut architecture is, in fact, a sculptor’s technique in which the entire mountainside is the piece of rock to be sculpted by carving out portions to create interiors, and shaping other portions to form exteriors. There were none of the architect/builder’s problems of getting blocks cut to fit precisely together and to function safely in structural terms. Nor were there the many different tasks that need to be executed in construction, like quarrying, transport, lifting, and placing. But rock-cutting is by no means a simpler technique. How, for instance, do you measure from exterior to interior of a rock-cut monument, and maintain a row of columns in alignment, when the desired endpoint is deep within the yet unpenetrated rock of the hillside? In construction, it is possible to rectify errors while carrying on with the building, but in rock-cutting, stone that is carved away cannot be put back. Since there are no definitive examples in India where remedial work was attempted, it is clear that the architect and rock-carvers had developed an unknown technique of successfully working their way into a hillside with the barest minimum of mistakes.

Equally crucial in rock-cutting is the issue of spatial organisation, which involves fitting a group of workers into a confined space where there is no possibility of moving things around. In the preliminary stages, rock-cut architecture necessarily confines work to a space that is significantly smaller than the final spaces, and it requires careful planning to figure out the crucial issue of the organisation of the stone carvers. Another important issue is the amount of stone waste that is created by the process of excavating into a mountainside. A vast quantity of stone waste must be extracted to create the spacious interiors of halls of worship; in caves, where space is tight, stone waste needs to be cleared continually in order for the rock-cutters and the sculptors to proceed with their work.

In this brief exploration of unfinished work at Ellora, we will consider one monument dedicated to each of the three faiths represented at the site. We will follow the order of their arrangement along the low scarp of rock at the site where caves numbered 1–12 are dedicated to Buddhism, caves numbered 13–29 are Hindu in focus, and caves numbered 30–34 are of Jain affiliation.

Cave 2 is a structure intended purely as a hall of worship. It does not incorporate monastic residential cells, which are common in the context of Buddhist viharas (monastic residences). Its badly damaged veranda opens into a central courtyard-like space demarcated by twelve columns. Cut into the far wall is a sanctum whose wide and tall doorway is flanked by a monumental standing bodhisattva and a female attendant. Within the shrine is a finely finished haloed image of the Buddha of impressive dimensions, seated upon a lion throne. Against each side wall is a tall standing Buddha and an attendant bodhisattva, while its inner corners are occupied by a set of five kneeling devotees with hands in the anjali mudra (gesture of respect and adoration) and carved one above the other in vertical formation. Thus far, the cave is complete.

But Cave 2’s unusual plan includes two side wings or galleries that are raised slightly over four feet above the floor level of the rest of the cave, though without any steps to provide access to them. Each “gallery” features five deep bays cut into its side wall and separated from the next by deeply cut attached pillars. Each bay features a large image of the Buddha, seated on an elaborately decorated lion throne, with legs pendent and placed on the floor. Each Buddha is flanked by bodhisattvas, placed so as to face each other across the niche of the Buddha image, while a set of flying celestials complete the assemblage. All five Buddhas in the gallery to the right of the viewer entering the cave are complete, together with their attendant bodhisattvas and the bays in which they are seated. By contrast, the Buddhas in the left gallery, together with their attendants and the niches that contain them, display differing degrees of unfinishedness that leave us with a number of unanswered questions.

Ignoring the right gallery with its finished imagery, let us turn to the Buddhas in the left gallery, moving inwards from the cave’s exterior wall. The first and third of the five Buddhas are finished images, seated within a carefully carved and completed niche, and each haloed Buddha sits with knees apart and feet planted on a chandrashila (crescent-shaped platform) carved on the floor. A standing bodhisattva attendant, holding a flower in one hand, faces his counterpart across the width of each Buddha niche. The second Buddha in this left gallery is close to complete; however, his feet are not fully released from the stone below, his throne has not been finished, and his accompanying attendants await completion.

The final two Buddha images towards the inner end of the cave are in differing states of unfinished. The niche of Buddha number four has not yet been cut back to create a vertical rear wall, but curves in at the top and sides almost as if it were a natural cave. Since the sides of the niche have not been cut back, there is no indication as yet of the flanking bodhisattva figures. The central figure of the Buddha has been carved in three levels of finish. The most highly finished area is the head, followed by the shoulders, chest, and upper arms, while his feet are still embedded in the rock. His throne has not taken shape, so the Buddha sits on rough stone which extends across the cave-like niche. Buddha number five is in an even rougher state of unfinished. The only thing that is at all clear is the head that has been carved so as to be clearly recognisable. The rest of the figure, as well as the niche, remains an uneven mass of stone, cut down and back to varying heights and depths. Both images confirm that rock-cutting, as well as finer carving, commenced from the uppermost levels and moved down.

Crucial questions remain to be answered regarding Cave 2’s unfinished work. Did the stone carvers and their patrons consider the cave to be usable, since they could indeed offer worship to the Buddha within the cave’s sanctum? Was the cave’s unknown patron perhaps satisfied with the level of work executed on the cave, and hence able to ignore the unfinished left gallery? Did they perhaps run out of funding for their work? In terms of time taken, had the stone carvers already committed all their time and were expected to be elsewhere for another project?

An excerpt from ‘Unfinished Work at Ellora’ by Vidya Dehejia in Ellora: Cross-fertilization of Style in Buddhist, Hindu and Jain Cave Temples, edited by Deepanjana Klein and Arno Klein, Mapin Publishing.