Distinction in Indian Iconography [Part I, 2]

Posted: 02.01.2012
Updated on: 02.07.2015


The paper was published in Bulletin of the Deccan College Research Institute 20 (S.K. De Felicitation Volume), Poona 1960, pp. 164-248.


Distinction in Indian Iconography

Part I: The Phenomena


II. Standard Types of Iconoplastic Art

§ 10. The stereotype character of medieval art depends largely on two conventional systems of representation which have created what may be called the “attribute-type” and the “panel-type“.

In the case of the attribute-type all elements of the scene are transformed into accessories of the main figure, are 'attached' to it. Human figures (consorts, friends and foes), animals, even plants and mountains become attributes in the narrower or in the wider sense. A good many elements are completely dropped. In the final conception nothing remains separated from the main figure. The places for the attributes are also standardized: the symbols are mostly 'attached' to the feet, to the hands, and to the head. Where we find variations in the place, position and shape of an attribute we are generally only concerned with stylistic, not with iconographic variants (cf. the Introduction). It is for example irrelevant to the identification in which position a cakra is held and whether the face shows spokes or a lotus. The description of many icons can therefore be reduced to a mere formula giving only name and place of the attributes.

There is no strict correlation between the place and the attribute. The trident is normally a hand-attribute but it can also be attached to the head (§ 50c). The conch appears in the hand, as vāhana (i.e. with the figure seated or standing on it), [1] and on the head (§ 50c). The cakra is represented in the hand, on the chest (§ 50c), and on the head.

In theory we can distinguish between attributes which are attached to the body, attributes which consist in anatomical peculiarities (third eye and third leg, uṣṇīṣa), and attributes which consist in particular attitudes (raised leg, ūrdhvalinga). In epic representations only attributes of the second type are admissible, and Indra's flask as represented on the northern gate at Sāñcī (Vessantara-Jātaka) is strictly speaking already an unepic element. Non-epic representation associates a momentary feature permanently with a person (Kṛṣṇa's cakra) and makes a “permanent“ feature (the number of arms) subject to change. The time relation of the elements has become irrelevant.


§ 11. We do not know how far the concentration of the elements described in the previous § is an original feature of the art-motif and how far it is the result of later transformations. The Mahiāsuramardinī at Deogarh [2] (Gupta period) forms already a perfect unit with the overpowered buffalo: she places one foot on its head and clutches its tail. It is not impossible that an earlier representation showed the two opponents separated, as is the case with the Mahiṣāsuramardinī-panel at Mahābalipuram; [3] here the goddess is engaged in a fight with the buffalo-demon who appears in hybrid form (human body and buffalo's head) and stands in front of her. But since we have in the North no rendering of the motif earlier than the Deogarh panel, we cannot say whether this shows the original or a later form. Cases where we find in Hindu iconography one and the same motif in an earlier “un - concentrated“ and in a later “concentrated“ form are rare.


§ 12. The panel-type shows the main figure surrounded by small panels, more or less clearly separated by offsets and each accommodating a human or animal figure (fig. 15/16). In a wider sense, the panel-type includes every symmetrical, composite scheme which arranges different figures (or inanimate objects) round a big central figure: for example a central figure with two smaller figures to the left and right. The selection of the elements is different from the attribute-type. The latter prefers such elements which can be turned into attributes (weapons, implements in general, killed enemies, etc.) and increases the number of arms, which have to serve as the receptacles for these attributes. The panel-type, on the other hand, gives preference to figures which can be accommodated in separate panels: associated figures (avatāras in the case of Viṣṇu, Jinas in the case of a Jina), attendants (chowrie-bearers, parasol-bearers), or witnesses of some incident (cows and shepherds around Kṛṣṇa Govardhanadhara: fig. 16). - Our types are no doubt abstractions and only a part of the icons will answer in all respects to one of the two descriptions. But broadly speaking the images with a great number of arms are not the same as those which present an elaborate system of subsidiary figures.


§ 13. Just as in the case of the attribute-type, there is no strict correlation between the place (here the panel) and the occupant (here the figure); the latter is exchangeable (see § 5 on substitution). Thus Buddha and Sūrya can occupy the places by the side of the main figure (Hari-Hara) normally reserved for attendants like chowrie-bearers, parasol-bearers, etc. (Banerjea, Pl. XLVIII, 1: “Hari-Hara-Sūrya-Buddha“). The representation of attendant figures in these places is only the statistical maximum; it does not mean that the places are meant for such figures. Another image shows a female chowrie-bearer (?) to the proper right and the yakṣī Ambikā to the proper left of the Jina Pārśvanātha (fig. 14). [4] This case is slightly different from the first in so far as only one out of two figures which form a pair (the chowrie-bearers) has been replaced by a figure of different character (Ambikā).


§ 14. Sometimes motifs which were normally rendered in a different way, viz. in a more epic form have been adapted to the panel-type. They lost thereby their distinctive character.

The Kṛṣṇa Govardhanadhara from Mathurā which is reproduced in fig. 16 can be compared with a Jina from Deogarh (fig. 15). The date of the first specimen is controversial [5] but the piece is in any case much earlier than the Jina from Deogarh which is medieval. Amongst the early representations of the Kṛṣṇa-Govardhanadhara-motif this seems to be the only one that follows the panel-type. The correspondence of the elements of the two compared compositions is clear:

  • Shepherds with sticks (looking like maces) flanking Kṛṣṇa: chowrie-bearers flanking the Jina.
  • Cows (resembling sheep) by the sides of Kṛṣṇa: miniature-Jinas by the sides of the Jina.
  • Mountain and tree above Kṛṣṇa: triple parasol with tree above the Jina.

The tree above Kṛṣṇa is out of place; but this tree which is originally only associated with the yogi (Buddha, Jina, Nara/Nārayaṇa, Dakṣiṇāmūrti) is later on shown even above amorous couples. [6] Were the branches not clearly carved below the mountain Govardhana, one could conjecture that the tree suggested vegetation (just as we find leaves on the mountain in the case of an unpublished Gupta image of Kṛṣṇa Govardhanadhara at Deogarh. [7] It is however evident that the artist did not even try to bring the leaves into agreement with the motif. The identical plan of the two images demonstrates stereotype representation, but the presence of the tree in both cases is due to a contamination of the Kṛṣṇa-motif with the yogi-motif (§ 25).

An unusual medieval representation of the Kūrmāvatāra in the Gwalior Museum [7] can be compared for its plan with a Buddhist composition at Kanheri (cave No. 90) which shows the Buddha surrounded by attendants and by other Buddhas. In both cases all the represented figures are seated or standing on lotuses, and also the arrangement of the flowers is about the same. - The lotuses do not originally belong to the Kūrmāvatāra-motif; here they are so predominant that they have changed the character of the scene completely (cf. § 8). The myth is of course aquatic in character and would admit of a few lotuses as indication of the water. But that all figures (including the tortoise and the churning gods and asuras) are placed on separate lotuses, is an invention of the artist (cf. also § 7 on the lotus-parasol and § 28).


§ 15. At the end of the chapter a word or two may be said about over-characterization and differentiation. These two tendencies have strictly speaking nothing to do with our subject. But they add to the difficulties which we experience in the study of Indian icons and we shall therefore discuss them briefly.

For the characterization of a figure one attribute is sufficient. But only in the case of serial groups like the Jinas or Mānui-Buddhas were the theologians ready to restrict the number of attributes to one. Elsewhere we find a plurality of attributes and this is in keeping with the iconographic conventions of other arts. Still the great number of attributes in the narrower sense as well as of distinguishing features in general is a peculiarity of Indian icons and calls for some explanation. Three reasons can be suggested.

  1. Amalgamation. The full iconographic equipment of two or more figures was combined; an increase in the number of arms made it possible to concentrate the attributes of several figures on one figure (the amalgam). See § 37.
  2. Multiplication. The addition of further pairs of arms was probably not always inspired by syncretistic ideas. We find images with very many arms where each arm carries a weapon, but not an attribute in the strict sense. Later on, the great number of arms was extended to other figures where each arm had to be provided with a different symbol. See § 37.
  3. Variation. Variety is not only the result of certain historical developments. The theologians took an active interest in the enlargement of the number of characteristics. According to iconographic theory the gods are defined rot only by their head-, hand-, and feet-attributes but also by differences in their proportions, [9] by different colours (which are again different for different parts of the body and of the dress), by different ornaments, and so forth. These distinctions were, however, normally not observed by the artists.


§ 16. Over-characterization is an obstacle to the identification, because the distinction of icons is getting more difficult as the number of attributes increases. Lack of uniformity due to differentiation presents an additional difficulty.

The Vaikhānasāgama describes twelve varieties of Viṣṇu-images and gives for most of the varieties again three sub-varieties. Besides we get the so-called caturviśatimūrtis of Viṣṇu. These 24 variants are distinguished only by their names and by the different permutations of the four conventional attributes (conch, cakra, mace, lotus). The different avatāras of Viṣṇu can also be represented in more than one form, although the number of variants is smaller in these cases.

Leaving aside such cases of systematic differentiation we find accidental differences between different texts, between different monuments, and even between different figures of the same temple (see § 22 on Īśāna). The divergencies are so great that it is often not easy to find a relatively large number of images of the same deity which coincide in all respects. One cannot even be sure to find certain standard-elements on all images of a god or goddess. It follows from the iconographic tables prepared by U. P. Shah that the Jain Ambikā is by no means always represented with a child and/or a mango-bunch and that the Jain Sarasvatī carries sometimes neither her vīā nor her book. [10] Under such circumstances it is necessary to study the icons not in isolation but in their full (literary or artistic) context. The scanty and irregular evidence which is contained in the single icon must be supplemented by the evidence which can be derived from other figures in the same context.


§ 17. We also find the very opposite of over-characterization - “under-characterization“, which is a form of stereotype representation. Images of the Buddha and Jina in yogi-attitude are often not sufficiently characterized. The yogi-forms of Śiva, Brahmā, Viṣṇu, however, carry as a rule two attributes in an additional pair of arms (which does not display the dhyānamudrā). Under-characterization need not be older than the full characterization, but as time went on the latter type gained ground; in Hindu iconography under-characterized figures do not seem to have been in vogue at any time. - There are indeed cases where we find it difficult to distinguish a Jina from a Buddha. The identification may depend completely on the indication of the garment in the case of the Buddha; but the outlines of the garment are occasionally too faint to be distinguished from a streak in the stone.

Under-characterization is also found in the case of the consorts of the gods, i.e. in the case of female deities who are represented beside their husbands. But here the female figure is only a sort of attribute of the male god and not conceived as an independent being. The characterization of Rāma (and Bharata, Lakṣmana, Sītā, Hanumān) and some other gods in South Indian iconography is also moderate as compared with the normal Indian usage. But this is a later development; it can be compared with the system of characterization in Greek iconography.

Share this page on:



Bulletin of the Deccan College Research Institute 20 (1960)

Compiled by PK