Prof. Dr. Klaus Bruhn
In the earlier periods, each specimen of the banana plant motif can be studied as an artistic realization in its own right. But as the number of specimens increase we wonder whether a classification (on morphological lines) might be possible. In discussing this issue we must distinguish between diversity (all related specimens form a polythetic class) and fluidity (lack of true formulas). We are faced with both, diversity and fluidity, but while the former is no obstacle to a classification (banana plant motifs a, b, c …) we observe serious taxonomic difficulties due to fluidity. There it is not possible to establish a set of distinct or enumerable formulas. Tentatively, we suggest a distinction between a more realistic type (fig. 2) and a more abstract type (fig. 5). Here, the realistic type is less fluid than the abstract type and it may not be altogether impossible to isolate a number of specimens which demonstrate the main varieties of the realistic type. But such a procedure would hardly make sense in the case of the abstract type.
Due to these circumstances, we have to look for non-morphological ways for the organization of our material (§§ 5-6). In the present section we shall study the banana plant motifs in four sequences of small niches, encircling the outer walls of four large early medieval temples: Telī-kā-Mandir (Gwalior Fort), Deogarh-12 (Jaina Temple No. 12 at Deogarh), Gadarmal Temple (Badoh-Pathari), and Mālādevī Temple (Jaina temple at Gyaraspur). The four temples belong to O. Viennot's periods IV-VI (TIC pp. 247-48). Absolute chronology causes great problems because dated inscriptions belonging to the relevant period (A.D. 750-900) are extremely rare. For published illustrations and plans refer to § 12.
Since we shall be concerned with niches it seems useful to say a few words about their architectural context. We subdivide the outer walls, vertical surfaces encompassing the temple and sometimes only interrupted by a doorway opposite the cella, into three superimposed horizontal zones: plinth, niche-zone and entablature. The niche-zone is dominated by niches but our tripartite analysis does not suggest that niches are actually confined to this zone. We can however say that, due to the niches, the intermediate zone shows a maximum of discontinuity and variety. Discontinuity (see below) is mainly relevant to the formal analysis of an individual temple, whereas variety is a “survey concept“. Zones of great variety (“niche-zones“, pillars etc.) require a maximum of documentation. A survey therefore lias to cover each and every original feature of a given temple and also as many different temples as possible. Refer in this connection to TIC 111 (Niche A versus Niche B) and to TIC 111 vs. TIC 133/135 (Temple A versus Temple B).
The most conspicuous elements on the outer walls of the Telī-kā-Mandir are big niches. One of them contains an ekamukha liṅga (Meister Ām, fn. 17), but all the others are now empty. Two niche images have been brought by M.B. Garde to the Gwalior Museum (Misra Gw p. 224). The big niches are either confined to the niche-zone or their base-line is so low that they cut into the plinth (TIC 96-98). In this connection we use the term “incision“ and we add that the whole scheme of the outer wall is, in our phase, largely dominated by a conflict between continuity (girdle effect) and discontinuity (vertical accents). Discontinuity is due to the dynamism of niches and doorways which operates by way of incision against the basic continuity of the three horizontal zones (e.g. compare Kramrisch In 107). We may add that, generally speaking, the tripartite wall scheme is most pronounced in those temples which are somewhat earlier than the four temples to be discussed presently (TIC 63 and 108). Even then, the tripartite pattern is - via affirmationis or via negationis - indispensable for descriptive purposes in the case of most early medieval temples.
One and the same temple may have large niches and small niches side by side. This is, however, not to say that “large niche“ or “small niche“ are terms which recommened themselves for general use.
The situation changes in many cases from temple to temple. This differs, however, when we consider the figures in the niches. Here we would like to introduce the term “small niche figures“ for figures that are found in small niches on the outer walls (Telī-kā-Mandir etc.) and which do not show much sonographic profile or characterization. Fully characterized figures and under-characterized figures may occur side by side in the same sequence but it seems appropriate to treat the under-characterized figures in the small niches as a category sui generis (which must also be distinguished from under-characterized figures in other architectural or sculptural contexts). Finally, we have to add that the distinction between “fully characterized“ and “under-characterized“ is not identical with the distinction between “identifiable“ and “unidentifiable“. Sometimes, figures with many arms and attributes cannot be identified while identification may be possible in the case of quite unpretentious two-armed figures.
Our first item pertains to the 113 small niches which appear immediately above the plinth of the Telī-kā-Mandir (TIC 96-98). A noteworthy decorative feature of the relevant zone (or rather “sub-zone“) is the “band of leaf scrolls“ (J.C. Harle: SUB p. 150-51; JID 283: panel above the pilaster capitals). The iconographic programme of the niches is simple: one figure per niche (exceptions are negligeable) and few if any motifs in addition to the figure. Most of the figures are Śaiva, but references to Śaiva mythology are minimal. A certain similarity with the stone panels of Paharpur seems to exist (cf. Dikshit Pa 30). We do not know the deity (deities) to whom the temple was dedicated but G. von Mitterwallner suggests, on account of the oblong cella, the Aṣṭamātṛkās. Naturally, we are here only concerned with the vegetable motifs accompanying the niche figures. These are either flanked by banana plants (“L, R“), or a banana plant appears on one side (“L“ or “R“). It seems that the banana plant is restricted to Śaiva themes but we are not absolutely sure because our photographic material is complete only for niches 1-56 (“N 1“-“N 56“). Here, and in the case of the other three temples, we number the niches clock-wise, and starting from the doorway (Deogarh-12: main doorway), this is our list:
In the case of N 7 (L?), N 11, N 25, and N 33 (“L, R“?) our photos do not permit a conclusive answer. L (R) = on our left (right).
The distribution (L: R: L, R) is not determined by any rational motives. So we merely mention that the duplication (L, R)-found already at Paharpur-foreshadows the developments in later (iconoplastic) art. For the study of the morphology of the banana plants we refer the reader to seven niches which we have illustrated below or, earlier, in JID: N 28 (JID 278); N 30 (fig. 7), N 50 (fig. 8), N 56 (fig. 5), N 107 (fig. 6), N 111 (JID 283, panel on our right), N 112 (JID 283, panel on our left).
Most of the banana plants follow what we would call the “abstract“ type. Our fig. 5 is a case in point. The “realistic“ type (figs. 1, 2 etc.) is hardly found. N 107 (fig. 6) is a rather weak example of the realistic type, but there may be a few bitter examples amongst Niches 57-113. By contrast, we do get forms of the motif which are partially realistic although they show the vigour and compactness of the abstract type (N 28).
Marginal in the context of the present sequence but interesting for general reasons are a few untypical specimens of our motif. In N 50 we have one or two elements in the lower corners which do not look like banana plants but betray the influence of the double-leaf as it occurs, apart from decorative employment, in some of the panels (including N 50). By contrast, N 30 shows a stylization which is vaguely reminiscent of the later isolated node. Finally, vegetable motifs like those in N 111 and N 112 are no models of clarity. It is not difficult to quote further examples (including some cases from our sequence) where vegetable motifs filling the corners of a panel are virtually out of focus. In Handa Os 102 we note banana plants looking like double-leaves (compare for true double-leaves and decorative double-leaves at Osian Handa Os 95 etc.). A standing Jina of the medieval period in Deogarh Temple 9 (Bruhn Di fig. 32; our Neg. no. 958) includes miniature compositions with standing Jinas where one and the same corner motif appears in all the four corners of the panels. A similar case (also medieval) is shown in JID 375. A primitive early medieval image of unsettled provenance shows Ṛṣabha and Supārśva (five hoods) along with three vegetable motifs (v) beneath: v-Ṛṣabha-v-Supārśva-v (Neven St 162). Whatever the evaluation of contemporary and later parallels, as a working hypothesis we would suggest that what looks in the Telī-kā-Mandir niches like an isolated node is actually some less common form of the banana plant (refer to N 30 = fig. 7; and to N 46 = Neg. no. 2187, niche on our left). As a consequence, we are at this stage only concerned with the “dynamism“ of two motifs, viz. banana plant and double-leaf. For interferences between different motifs (substitution etc.) refer also to the next para.
Deogarh-12 has a pradakṣiṇāpatha, and its śikhara rests on the walls of the sanctum. As a consequence, the facades of the prāsāda or mūlaprāsāda (portion below) and of the śikhara (portion above) do not form a continuous surface (JID 393, Section). The vertical walls in this case are only the outer walls of the pradakṣiṇāpatha. The old tripartite scheme is preserved. The vertical walls, more particularly the niche-zone, are interrupted by the main doorway and by three subsidiary doorways on the other three sides (JID 393, Plan). The niche-zone accommodates 26 niches (JID 393, Plan), and there are no other niches on the outer walls. All 26 niches follow the same type, but differences in width are more pronounced than in the case of the Telī-kā-Mandir. N 2 and N 25 are hidden from sight by small shrines (JID 392). N 26 is occupied by a seated Jina, while the remaining 23 (25) niches contain Jaina goddesses and other female figures. In JID we have reproduced all visible niches except N 24 (see JID 52-74). Out of the 24 visible niches, five show the banana plant: N 3 (L, R; JID 53), N 4 (L, R; JID 54), N 6 (L, R; JID 56), N 8 (L, R; JID 58), N 20 (L, R; JID 70). Niche 10 (JID 60) is not clear. All double motifs except that in Niche 8 represent what we would call the realistic type. The motif of Niche 8 represents a rare variety or an ad hoc fabrication. In one case (N 7: JID 57) we have as corner motifs (lower corners) the two rolled-up dupaṭṭā ends. This motif recurs inter alia in JID 90 (one dupaṭṭā end only). One gets the impression (JID 57) that one corner motif (banana plant) has been “replaced by“ some other motif (dupaṭṭā ends), or that the other motif has been “substituted for“ the current corner motif. Irrespective of the evaluation of the present case we have to emphasize that “substitution“ is an important form principle (the number of cases increase as we extend the meaning of the term). The same applies to “assimilation“. Again both form principles overlap considerably so that we end in an analytic quandary. Our recommentation for such cases is to isolate specific clusters of interferences (comprising in our case substitution and assimilation) and to replace logical analysis by documentation. In connection with substitution and assimilation, corner motifs would be of special interest.
Ignoring the confused and dilapidated upper portions of the outer walls of the Gadarmal Temple, we can isolate two lower zones which may be designated as plinth and lower niche-zone, respectively. The lower niche-zone is provided with 22 small niches, and a few additional small niches are attached to the plinth (TIC 185-87). The 22 niches contain couples (N 5, N 10, N 13, and N 18) or single figures (male or female). All four couples are flanked by banana plants, but there are no banana plants in the remaining panels. Refer for a niche with a couple to fig. 9 (N 10, backwall) and for a niche with a single figure to JID 75 (N 14, backwall). The general line is realistic. This applies probably also to specimens of the banana plant on the same temple which do not belong to the set of 22 niches. It is worthy of note that N 10 foreshadows the medieval calyx type (§ 7) and continues the formula of Weiner Gu 16 (Muṇḍeśvarī).
In the case of the Mālādevī Temple, we can isolate on the outer walls (partly replaced by the surface of the bed rock: Ali Pr, plan) a sequence of heterogenous niches. The niches differ in character and size but the base-lines seem to be the same in all cases (compare the Sūrya Temple at Mankhera, TIC 229). There are big and small niches. The latter fall into two distinct types which shall be designated as “broad“ and “narrow“, respectively. Seven small niches are broad and four are narrow. N 1 and N 2 decorate the south-east corner (N 1: east face; N 2: south face). This is a list of the eleven small niches (error and omissions excepted):
The four narrow niches accommodate two-armed figures with snake-hoods and with two banana plants (L, R). N 3 contains a male figure (Deva Mā 3), the figures in the other three niches being female. Two of the goddesses in the seven broad niches, those in N 2 and N 4, have likewise banana plants (L). All the small niches are framed by round pillars (TIC p. 101: colonnettes rondes), but there are no fabulous animals on the outer sides of the pillars as one might expect under the circumstances (TIC p. 106, lines 9-12: deux innovations). The narrow niches show technical peculiarities which cannot be discussed here. The banana plants of the narrow and of the broad niches are elaborate and more refined than those of Deogarh-12 and of the Gadarmal Temple.