THE EXPERIMENTAL PHASE
Unitary Multiple Icons of the Kusana Period and Related Works of the Early Gupta Period

 There are several north Indian sculptures of the Kuṣāṇa period and other related products of the early Gupta period which represent an experimental phase in early iconographical composition It was during this phase that the unitary multiple icon as a free-standing sculpture was invented. These works are termed experimental for the following reasons. They do not continue any earlier iconographical tradition. Each sculpture is unique; none constitutes a fixed iconographic precedent which is later repeatedly copied. Each sculpture appears to be an original attempt to express in visual terms a complex abstract concept.

The sculptures to be discussed in this chapter which do not bear any definite brahmanical symbols are nevertheless considered to be brahmanical icons on the following- grounds. The metaphysical concept of which each sculpture appears to be a Visual expression cannot be traced in either Buddhist or Jaina theology; a reasonable explanation of each icon is, on the other hand, to be found in the brahmanical tradition. The images described and analysed in this chapter can be shown to be the prototypes form which iron types of the brahmanical cults were derived.

Of the few extant works which represent this experimental Phase in brahmanical iconography, I shall examine the following:

1.	Mathura Museum, no. 611.9 (stone fragment).
2.	Nānd Rajasthan (stow sculpture, in situ).
3.	Mathura Museum, no. 14.392-395 (part of a stone sculpture; assembled fragments).
4.	Mathura Museum, no. 00.F2 (part of a stone sculpture).
5.	Muktā Devī temple museum, Musānagar, Uttar Pradesh (relief panel on a stone railing pillar).
6.	Mathura Museum, no. 15.516 (damaged stone sculpture).
7.	Lucknow Museum, no. H4 (damaged stone sculpture).
8.	Mathura Museum, no. 14.382 (part of a stone sculpture).
9.	Mathura Museum, no. G-64.312 (part of a stone sculptor?).
I0.	Mathura Museum, no.00.E12 (damaged stone sculpture).
11.	Bikaner Museum, no. 228 (terracotta plaque).
12.	[Private collection] (part of a stone sculpture).

          While each of these sculptures is unique, as noted above, they may be classified for the purposes of analysis according to the type of iconographical structure employed in integrating the multiple elements of their composition to form a unitary icon. These structures are the following:

Type A. A vertical arrangement of elements in which each figure is so positioned as to appear behind and partly above the figure below it, as if all were seated upon each other’s shoulders, or standing upon the successive steps of a stairway.

Type B. A ramifying composition in which several partly represented figures are conjoined obliquely and vertically, somewhat like the spines of a fan, to a central upright figure.

Type C. An apparent compromise between these two types, consisting of one partly represented figure above and toward the back of the main figure (a modification of the method used in Type A) while the obliquely disposed figures of Type B Seem to have been abbreviated in such a way that only their heads are represented on either side of the face of the central figure.

Type D. Multi-headed unitary icons of different types.

Of the selected works which remain from this experimental phase in brahmanical iconography, listed above, Sculptures numbered 1 and 2 belong to Type A; Sculptures 3 to 5 are of Type B; Sculptures 6 and 7 are related types; sculptures 8 to 10 belong to Type C; and 11 and 12 are transitional images classed as Type D. The descriptive analysis and interpretation of each work within the four types which follows will provide the iconographical background to mature, established multiple icon types of the Gupta and post periods.

The first, and probably the earlier example of Type A, is a red stone fragment found in the course of fieldwork, in one of the storerooms of the Mathura Museum (Plate 1-2). It is an upright rectangular slab, 59.7 centimetres in height. Upon one of the wider sides is one anthropomorphic figure in low relief, another headless figure above it and a head only below it.  There are no gins of joints (for example, mortice sockets) on the narrow sides, which would have suggested that it was a railing pillar.  On the contrary, there are faint traces of carving, now badly eroded and notrecognizable as representing any particular figure or objects, on the sides where such joints would normally be found if it were part of a stone railing.  It may have been a cult object in its own right; this conjecture receives some support from the description of the Nānd column, which follows, as the only other known example of Type A.

The head at the base of the fragment (Plate 2) were the turban-with-rosette headdress which is common to Kuṣāṇa sculpture and is to be seen on the face of the Ekamukhaliṅga from Aghapur[1], on other Kuṣāṇa Mukhaliṅgas, and eleven times on the Nānd column. The now headless topmost figure, evidently male, retains no clear evidence of clothing or body adornment, and the right arm is broken, its remains suggesting that it was doubled back in the abhayamudrā. The left hand holds an object resembling a stick with a swelling outward curve at the top, which appears to have been a cāmara. The middle figure has a hairstyle pulled to the left of the head, which is surmounted by what appears to have been a peaked headdress, and wears pendant ear-ornaments and a necklace. The bosom is unmistakably female. The right hand is held up, as if in abhayamudrā, and some elongated object is held in the left hand, leaning across the body.

The second, and only other known example of Type A, is a sculpture which presents the appearance of a populated Śivaliṅga (Plate 3-9).  It stands in the open, implanted upright in the surface of a square mound faced with stone blocks among the low hills outside Pushkar not far from Ajmer in Rajasthan, near the village of Nānd.

The column stands approximately 152.5 cm high above the surface of the mound. It is carved with anthropomorphic figures on four sides and has an elongated hemispherical top. The main shaft of the column is demarcated by a rebate cut horizontally into the stone three-quarters of the distance from the base; the remaining portion constitutes the curved top of the column, having a smaller diameter than that of the shaft.

Four badly damaged figures are carved upon the top portion, each facing one of the four cardinal directions. At first glance, these seem to be grotesque, pot-bellied dwarf-like creatures of the type associated with Śiva. They appear in a squatting posture upon the narrow shelf formed by the rebate. Their hands and arms are mostly broken, but it seems from the remains that the right arm hung downward and that the left elbow rested on the left knee, with the forearm extended horizontally across the chest (Plate 6). They seem to have been depicted as completely naked, and at least two of them are ithyphallic, a feature which would suggest strongly that they are Śaiva.  Their formal posture and positioning in the four directions, however, suggests that they were intended to be more than mere Śiva-gaṇas.

Below each of these four is a vertical series of three figures.  The two upper figure of each series are depicted standing behind and partly above the one below it, in such a way as to be fully visible only from the waist upward. The lowest figure can also be seen only above the waist; unless the mound is excavated, it cannot be ascertained whether or not there are further figures on the shaft.  It is possible that there is a tenon or other plain projection beneath the lowest visible figures and that what is at present visible constitutes the whole sculpture.

Eleven of these twelve figures appears to wear the same headdress, consisting of a turban with a large ‘rosette’ on the crown, of the type frequently seen in Kuṣāṇa sculpture. The most immediately relevant comparative example is the headdress of the Ekamukhaliṅga from Aghapur already mentioned.  On one side of the shaft, two of these figure – the third is severely damaged – have a circular halo behind the headdress, a feature not shared with any of the remaining nine figures.  All the figures are represented with natural human anatomy and all are two-armed, although the forearms and hand-held symbols are in nearly every instance broken off, and none has more than one face.

My description of this complicated sculpture might suggest a Śaiva cult affiliation. But his assumption must be questioned at the outset set. The following points may be set against a straight-forward Śaiva identification, Frequently, Liṅga sculptures of early date which incorporate an anthropomorphic element are unequivocally phallic in shape. The Aghapur Liṅga and the Liṅga at Guḍimallam, Andhra Pradesh, are perhaps the most striking examples, in which the base of the top section of the Liṅga has a greater diameter than the shaft. In no type Liṅga which exhibits this primitive attempt at realism has the opposite of such a relative scale of proportions been used. That is, the ‘realistic’ phallic symbol does not have a top portion of smaller diameter than the rest of the shaft, which is exactly what the Nānd column does have, as may be observed in Plate 9 especially Even in the less realistic Liṅga types of the Gupta period and later, the brahmasūtra lines which demarcate the top section do not mark an abrupt reduction in the diameter. As the Nānd pillar is not a ‘realistic’ phallic symbol, it may not be a Śivaliṅga.

The second point is equally negative. If the figures which quart upon the top of the shaft are Śaiva-gaṇas, they are ritually displaced. For such minor grotesque figures are rarely, if ever accorded such high ritual status as to warrant a place against the very apex of the most sacred of Śaiva symbols, the Liṅga. The deformed creature upon which Śaiva stands in the Guḍimallam Liṅga sculpture is clearly a defeated personification of anti-Śaiva forces; it is debased, not elevated, and Śaiva does not emanate from it as the figures upon the shaft of the Nānd column emanate from each other.

The third point which allows one to query the phallic nature of this column is the fact that there were in existence, in the third century A. D. at the latest, non-phallic cult objects of comparable shape and content. These constructs are the Buddhist pillar carvings of the kind found at Sanchi on the uprights of the gateways and at Amaravati.[2]The best illustration is that type of pillar, to be seen in relief panels from Amaravati, which has populated panels upon the shaft. A good example is a pillar, 128 cm in height[3], which depicts three successive stages in the life of the Buddha, in ascending sequences, which is followed by a frieze of squatting gaṇas who support the top of the pillar which has the domed apex of a miniature stūpa. Thus far it is a direct parallel, although in a different idiom, of the Nānd column; above the stūpa.-apex there follow the familiar Buddhist appurtenances of harmikā and dharmacakra.

For our analysis, it is convenient to refer to the side of the Nānd column which bears the nimbate figures-and which may thus he regarded as the ‘front’ or most important aspect of the composition-as side ‘A’. The remainder may then he termed sides ‘B’, ‘C’ and ‘D respectively, in clockwise order. Similarly, individual figures from bottom to top are numbered 1 to 3, and the apical figures are referred to as number 4. Thus the base figure on side ‘A’ will he called A 1

Only one scholar, R. C. Agrawala, has published this sculpture;[4] shall discuss his observations and interpretations first as providing a useful point of departure from which to enter into my own interpretation and conclusion. Agrawala assumes from the outset of his study of the column that it is a Śivaliṅga, also referring to it, inaccurately, as a; Chaturmukha Liṅga’,5 Liṅga having four faces. The Śaiva character of the sculpture is taken for granted, apparently, because of the shape of the column and because of the four ithyphallic figures squatting upon the rim of the shaft. As these pot-bellied, naked figures are separated by the ledge formed by the rebate from the super-imposed, standing crowned figures below them, I find it an inaccurate statement that the column ‘is divided into four horizontal tiers’6 Agrawala is in no doubt as to the identity of these four apical figures, stating of the column that ‘It depicts Lakulīśa as ūrdhvaretas at the top and on all the four sides; the deity is shown squattish like the Kuṣāṇa Sūrya from Mathura; all his four heads and hands are completely broken and that is why it is not possible to form any idea of the head-gears and so also of the weapons held by Lakulīśa.,7 This seems to be a remarkable statement, simultaneously acknowledging that identificatory details are lost and asserting a positive identification. Agrawala bases his identification of all four figures as Lakulīśa upon their undeniable ithyphallic (ūrdhva-retas) appearance and, contradicting his description of the extent of the damage in the previous statement quoted immediately, above upon what he claims represents a’ “manuscript”… (Pustaka)’ held in the hand of one of the four figures and ‘touching his chest to the left’.8 In view of the severely damaged condition of (these figures, this can only be guesswork; Lakulīśa, Agrawala states holds the manuscript in certain later sculptures.9 By reverse reasoning he claims that these squatting figures represent what ‘appears to be the earliest extant representation of squattish and independent Lakulīśa.10

In 1965 Agrawala published an article11 which challenged a statement made by K. V. Soundararajan to the effect that no Icons of Lakulīśa survive which antedate the sixth century A.D.,12 the evidence adduced by Agrawala in contradicting Soundararajan consists of the squatting ‘Lakulīśa’ figures near the top of the Nānd column and an image upon a pillar13 bearing an inscription with the Gupta date 61, equivalent to A.D. 380, which is also said to represent Lakulīśa. Subsequently, in an article14 published in 1970, Agrawala attempted to prove that the figure upon the same Gupta pillar is not ‘Lakulīśa’ but ‘Śiva as Bhairava’; one of the grounds for this change of mind-which is not acknowledged-is that no icons of Lakulīśa are found in Kuṣāṇa sculpture, including the Nānd column, with a flabby belly (lambodara)’,15 which is a prominent feature of the Gupta pillar figure. This is not only an unacknowledged reversal of a previously advanced identification, but also inaccurate description. For the four highest figures on the Nānd column are distinctly pot-bellied, as will be evident from the photographs, in marked contrast to the three superimposed figures beneath each of them; this contrast is most clearly to be seen by looking at the profiles of these figures, especially in the detail of the top of side C (Plate 6). On the grounds of insufficient identificatory details due to damage and the lack of comparative iconographical evidence from the Kuṣāṇa period, I do not accept the identification of these four figures as Lakulīśa.

Agrawala16 makes no Śaiva identifications on the remainder of the shaft; this fact in it self might lead one to suspect that the sculpture is not a Śivaliṅga. Most of the figures on the main shaft, it is stated here, ‘have raised up their right hands in abhaya pose.’ This is inaccurate. The abhayamudrā can, in fact, be positively identified only in the case of one figure, namely B.2. The right hands of seven of the eleven remaining figures, moreover, definitely held an object, now damaged; they could not have been posed in the empty-handed abhaya gesture.

Of the figures of level 1, Agrawala writes: ‘The lowest (i.e. fourth from the top) portion of course depicts Vaiṣṇava divinities very clearly on all the four sides; the main figure is that of Vāsudeva with a halo behind his head, the lower right hand is raised up in abhaya pose, the upper left seems to have carried a colossal mace with bulky portion on the top as also in the Kuṣāṇa sculptures from Mathura.’ There is no evidence to suggest that the latter figure (my A.1) was originally four-armed; like the other two haloed figures on this side, and indeed all the other figures on the sculpture, it appears to have had two arms. The abhayamudrā cannot clearly be distinguished in the broken remains of the right arm. None of the known Kuṣāṇa sculptures from Mathura which represent Vāsudeva or Viṣṇu as independent images support a mace on the left, which is the side upon which the cakra (disc) is held. The club appears conventionally at this time to have been depicted in the right hand. Moreover, the object to the left of figure A.1 (Plate 3, bottom right) is not iconographically connected with it. It is a continuation downward of the damaged object held on the end of a stick or stalk in the right hand of figure D. 1 (Plate 7, bottom left), as may readily be seen by, comparing Plates 3 and 7, probably as part of the block left in place as a strengthening bracket. Figure C.1 (Plate 5) is considered by Agrawala to be female, and he notes the large pot or drinking bowl held in the left hand; it is difficult to agree that it is a lotus which is held in the right hand of this figure. Agrawala suggests that ‘she may even represent Lakshmī or Ekānāṃśa [sic]’. Of the other figures at level 1, he considers B. 1 (Plate 4) to be an image of Baladeva, the object resting against its left shoulder being seen as a plough (Balarāma, Baladeva or Saṅkarṣaṇa being also known as Halāyudha. ‘He who is armed with a plough’). Since the ‘plough’ is severely damaged, this identification is uncertain; the ‘abhayamudrā that Agrawala sees the right hand of this figure displaying is a mistake, for there was some object, now unrecognizable, held in this hand, which is supported by a thickening of the shaft of the kind mistaken for the mace of Vāsudeva in the case of figures A. 1 and D. 1 Only in figure D. 1 (plates 7 and 8) is there one unmistakable iconographic symbol. Both arms of this figure are intact, the hands being raised to shoulder level and holding, on the proper right, the slightly curved stem of what appears to have been either a large flower or else a yak-tail (cāmara), and on the left a fully preserved, densely-spoked disc, the diameter of which is greater than the width of the crowned face. The figure thus represents, beyond question, an aspect of Viṣṇu, holding one of the standard emblems of that god, the cakra. Despite the erosion of The sculpture, the manner of holding the disc is clearly archaic in terms of brahmanical art, and may be compared with the upper left-hand of the earliest Kuṣāṇa Vāsudeva icons of Mathura,17 in which the disc is similarly held with the palm facing outward and the fingers hooked around the rim between the spokes. Tempting though it is to see, in the broken object at the end of the stick held in the right hand, the lotus blossom, another attribute of Viṣṇu it must be remarked that none of the Kuṣāṇa Vāsudeva icons in the Mathura style holds this symbolic flower. The only commonly represented object of this period which would conform to the shape of the broken cross-section remaining at the end of the stick would be the yak-tail fly-whisk. As this is held only by attendants of figures of divine Status in brahmanical sculpture, it must be presumed that the cakra-bearer represents either a minor Vaiṣṇava figure of unique iconography, or else an iconographically incomplete aspect or partial manifestation of Viṣṇu.

Regarding the figures of level 1 as a group, one sees (1) a crowned and nimbate figure whom Agrawala regards as Vāsudeva; for convenience, it may he presumed that this most important side of the column originally was oriented to face the east; (2) to the right of this Vāsudeva, figure B.1, which may represent Saṅkarṣaṇa, on the south face; (3) to the left or north side of Vāsudeva another Vaiṣṇava figure, D.1, wielding the disc; and (4) at the back, or on the west face of the column at this level, figure C. 1, which appears to portray an uncrowned female bearing a pot. If this latter figure is indeed female, as part of the first of three tetrads surrounding the axis, the fact will be seen to be of significance for my argument associating the Nānd column and two of the Mathura sculptures (infra, Sculptures 3 and 4) as manifestations of a theological system skin to that of the Pāṅcarātra.

Of level 2, it is stated only that it ‘may be Brahmabhāga(?)’, evidently in accordance with his assumption that the column as a whole is a Śivaliṅga, the three portions of which arc said to be associated with Brahmā, Viṣṇu and Rudra respectively from the base upward.18 This wholly unacceptable proposition has apparently been arrived at by Agrawala because of the irrefutably Vaiṣṇava character of at least one figure at level I; only by ignoring this fact and then counting the three divisions above it (levels 2 and 3, plus the four squatting figures) as the standard vertical series of divisions or bhāgas of a Śivaliṅga could level 2 he taken as the base-portion, dedicated to Brahmā, of a vertical axis.

On examination, the four figures at level 2 are seen all to he crowned with the crested turban and all definitely male. Figure A.2 (Plate 3) is, like A. 1, distinguished by a halo encircling the crown; but whereas the halo of figure A. 1 is decorated with concentric designs, that of A.2 above it is incised with curved lines, perhaps suggestive of lotus petals. The iconographical significance of this is not clear; hut the fact that there is a difference is in itself noteworthy. The left hand rests against the chest, while the right is raised to shoulder level; that this raised hand held an object is apparent from a supporting mass of the block, sloping downward to blend with the main shaft, its top upon which the hand-held emblem was represented being smashed. The figure to the proper right, B.2 (Plate 4) appears to have held its left hand in a similar position to that of the previous deity, while the right israised in aclear abhayamudrā, the edge of the hand turned forward as is common in Kuṣāṇa representations of this gesture. At the back at this level, figure C.2 (Plate 5) has its left hand lowered to waist level and holding an indistinct small object, rendered in low relief across the body, resembling a small flower. The right hand is raised to the level of the chest and holds a large object, damaged beyond recognition, by the handle, which rests aslope the shoulder; there is a marked contrast between the massiveness of this emblem and that held in the corresponding hand by the (probably) female figure below, which is broad hut supported by a very slender and curving handle or stalk. The last figure at this level, D.2 (Plates 7 and 8), holds two prominent emblems, one in each hand, of which that on the right is similar to that held in the same hand of the preceding figure. In the left hand, which is held at waist level, is a long staff which rises from behind the turban-crest of the figure below (D.1), across the shoulder, and on at an upward angle until cut off by a break in the stone. At the end of this staff there appears a large leaf-shaped object which may, perhaps, be intended to represent a spearhead; there is an oblong projection from the staff or haft, just below the leaf or spearhead, connecting with the inner shoulder or neck of the figure by whom it is held.

Turning now to the figures of level 3, it will be seen that on the front (Plate 3), A.3 is severely damaged; the broken remains of what may have been a halo are evident above the obliterated face, level with the rebate. One may assume that this figure was nimbate, since the two below it have very prominent-and clearly differentiated–śiraścakras. The left hand is again held across the body at chest level, and the broken right hand was evidently raised, with a broken surface beside it suggesting that it held some object. To its right, figure B.3 (Plate 4) appears to he in the same posture, but here again the hands are broken. On the rear or western side (Plate 5), C.3 has the left hand in the same position as the two preceding figures, and originally held some large object over the right shoulder which is also destroyed. Exactly the same may he said of the last figure in this series at level 3 (Plates 7 and 9), except that the left hand gesture against the chest is here seen clearly for the first time, the hand being undamaged. This strange gesture-which is repeated by so many of the figures upon the shaft that it attains almost to the conventionalized status of a mudrā–canhere be seen to take the form of what might be termed simply muṣṭi-karaṇa-mudr,19 “afist-making gesture’, for that is exactly how it appears. The knuckles are directed downward at an angle, the thumb invisible behind them; it might seem to be a pugilistic gesture, suggesting some heroic posture, or the formal stance of a man of power, somewhat like the Victorian holding of the lapel while posing for a photograph. It may derive from the practice, as seen in Gandhāra sculpture especially, but also at Mathura, of holding up the end of one’s robe over the left wrist, which, to judge from the frequency of its occurrence in sculpture, was considered stylish. These very Indian figures, bare to the waist, may be imitating such fashionable gestures to convey the impression of prosperity and self-confidence, in the absence of any emblem for the hand to bold. Its appearance upon this early sculpture, in the case of as many as seven of the twelve figures on the main shaft, is noteworthy for it is a transient convention in India, possibly of. Mediterranean origin, which is not perpetuated in later brahmanical iconography.

The four squatting figures upon the rebate which marks the top of the main shaft, their barks to the domed apex of the column, seem to he nearly identical in posture and in that they are all pot-bellied, ithyphallic and naked but for a cord around the waist. That on side B (Plate 4) suggests that they probably also wore armlets, at least on the left arm. These shared characteristics, which identify them as a distinct group, also serve to differentiate them utterly from the well-proportioned standing figures on the main shaft, each one of which appears to be different. This differentiation and their support, the ledge formed by the rebate, cuts off the four topmost figures from the ascending sequence of emanations below them.

Speculation upon their precise identity, as ‘Lakulīśa’ or any other established figure of the pantheon is, it may be suggested, a time-wasting exercise, in view both of their severely damaged condition and the formative phase through which brahmanical iconography was passing during the Kuṣāṇa period. It is rather, I propose, the ideology behind the complex composition of such early religious sculptures which is a useful subject for investigation with regard to the origins of the later, more numerous, standardized multiple iconographic types. The purpose of the four squatting figures at the top of the shaft of the present sculpture may be inferred from the details which remain visible and from their very difference from the twelve emanatory forms below them.

It may be said with some confidence that they represent ṛṣi-figure: their apparent yogic posture (somewhat similar to the utkuṭikāsana20), the ūrdhvaretas condition symbolized by the erect phallus21 and their nakedness (dig-ambara)22would all indicate that they are yogins enjoying the spiritual benefits of their exercises. Their paunchy appearance does not contradict this proposition: the Buddhist teaching of avoidance of extreme asceticism had long been known, and the Hindu tradition itself does not preclude sensual enjoyments on the part of enlightened seers. Indeed, Brahmā, the embodiment of wisdom and the seer, although not represented as ithypallic, is traditionally portrayed with a pot-belly.23 As ṛṣis, they are not divine figures-unless they represent apotheosised yogins such as the Lakulīśa figure whom these images represent in the opinion of Agrawala-but rather types or symbols of the enlightened state to which the devotee may aspire. To such a man the gods and their associated mythology become to a degree irrelevant, for the yogin who has achieved his aim (known as a sidha24) is in a beatific state and immortal as arc the gods; he becomes a maharṣi, inhabiting one of the highest worlds, or even attending the court of Indra.25 On earth, their worship was in epic times especially associated with groves of sacred trees which stood between earth and heaven.26 Similar notions were encountered in modern India during the field research for this work: in several villages, and especially on the hanks of rivers, was found a holy tree locally termed the mokṣ-pipplā (Sanskrit mokṣa-pippala, ‘tree of salvation’). In Kuṣāṇa times, the idea of the saviour finding enlightement under a particular tree was already well established in the Buddhist and Jaina traditions.