Prof. Dr. Klaus Bruhn
The essay was published in Berliner Indologische Studien No. 13/14. 2000, pp. 273-337.
§ 4. The System
In the case of the 24 yakṣas and 24 yakṣīs we have texts and images. However, we cannot derive the images from the texts, nor (as is possible in other cases) text elements from the images. We maintain that the connection between both sides is limited to a few cases (out of a total of forty-eight). The actual nature of this type of non-agreement will be described in §§ 4-5.
The earliest list of the 48 YY (Digambara) is contained in six āryā stanzas in part iv of the Trilokaprajñapti (Tiloyapannattī):
934 (yakṣas):
Govadaṇa-Mahājakkhā Timuho Jakkhesaro ya Tumburao /
Mādaṅga Vijaya Ajio Bamho Bamhesaro ya Kumāro //
935 (yakṣas):
Chammuhao Pādālo Kiṃnara-Kiṃpurusa-Garuḍa-Gandhavvā /
taha ya Kubero Varuṇo Bhiuḍī-Gomedha-Pāsa-Mātaṅgā //
936 (yakṣas):
gujjhakao idi ede jakkhā cauvīsa Usaha-pahudīṇaṃ /
titthayarāṇaṃ pāse ceṭṭhante bhatti-saṃjuttā //
937 (yakṣīs):
jakkhīo Cakkesari-Rohiṇi-Paṇṇatti-Vajjasiṅkhalā /
Vajjaṅkusā ya Appadicakkesari Purisadattā ya //
938 (yakṣīs):
Maṇavegā-Kālīo taha Jālāmāliṇī Mahākālī /
Gaurī-Gandhārīo Veroṭī solasā Aṇantamadi //
939 (yakṣīs):
Maṇasi-Mahāmāṇasiyā Jayā ya Vijayāparājidāo ya /
Bahurūpiṇi-Kumbhaṇḍī Paumā-Siddhāyiṇīo //
The text mentions 24 (?) plus 24 names, but it is otherwise not in a satisfactory condition (gujjhakao in 4.936, cf. gujjhaga), and a systematic comparison with the other relevant Digambara texts (mainly the manuscript “Ka“ and the Pratiṣṭhāsaroddhāra) would not be helpful. The Trilokaprajñapti has no descriptions (no pratimālakṣaṇa.s), and we quote the six āryā.s only in order to give an impression of the character of what appears to be the earliest available list.
The date of the Trilokaprajñapti is not known (Upadhye Ti: 4-10). There are later texts (Digambara and Śvetāmbara), comprising different literary genres (lists, quasi-narrative accounts, stuti.s, etc.) and supplying names with or without descriptions. We can thus isolate a literary “YY corpus“. In the Śvetāmbara tradition, we find the 48 names and the 48 descriptions of the YY inter alia in the 24 Jina biographies of Hemacandra's encyclopedic Triṣaṣṭiśalākāpuruṣacaritra (A.D. 1160-1172). The author has included the relevant data in his samavasaraṇa descriptions. These legendary descriptions (long tracts found in several different versions) contain iconographic data of one type or another (e.g. the aṣṭa prātihārya.s); but the combined descriptions (samavasaraṇa cum YY) are possibly a peculiarity of the Triṣaṣṭiśalākāpuruṣacaritra.
Below we present as a textual specimen the data concerning the YY of Ariṣṭanemi (no. 22) and Pārśva (no. 23). The names are the following:
[Trilokaprajnapti:]
Gomedha/Kumbhaṇḍī and Pāsa/Paumā, i.e. Pāsajakkha/Paumāvaī;
[Ms. Ka:]
Sarvāhṇa/Kūṣmāṇḍinī and Dharanendra/Padmāvatī, see fig. 12;
[Pratiṣṭhāsaroddhāra:]
Gomeda/Āmrā and Dharaṇa/Padmāvatī.
Descriptions (based on drawings derived from the manuscript Ka and on the Pratiṣṭhāsaroddhāra) will be given in outline. The descriptions in the Pratiṣṭhāsaroddhāra (edition not available to the present author) are complete. The drawings derived from the manuscript Ka are often somewhat whimsical or at least unlikely, and it would be much better to have the manuscript itself. Refer for Digambara sources in general to Settar Ka: 25-26. Complete descriptions are easily found in the Śvetāmbara tradition (inter alia Hemacandra's Triṣaṣṭiśalākāpuruṣacaritra and the Nirvāṇakalikā). But these Śvetāmbara texts can, at least in principle, not be used for the interpretation of Digambara images. - The two Digambara descriptions are as follows:
In the Śvetāmbara tradition, the descriptions are as follows (Nirvāṇakalikā fol.37a).
The Pratiṣṭhāsaroddhāra of the Digambara tradition belongs, according to U.P. Shah, to the 13th century (Shah Ma: 123, fn. 38), while the Nirvāṇakalikā of the Śvetāmbaras can be attributed to “c. eleventh century“ (JRM: 214). The Jinaratnakośa (H.D. Velankar) mentions saṃvat 1285 as the date of the composition of the Pratiṣṭhāsaroddhāra.
We do not know what prompted the ancient authors to invent this system. They hardly expected that the artists would be guided by the fantastic descriptions in the texts (fig. 12). The whole concept of YY is mere theory with the exception of Ṛṣabha's yakṣa and yakṣī (Gomukha and Cakreśvarī; see § 6 for C). Ambikā and Padmāvatī confirm the system to some extent, but not in the same manner as Gomukha and Cakreśvarī. Ambikā is omnipresent in art, but she does not occur as an element of the system: Ambikā's partner (Sarvāhṇa, etc.: supra) is nowhere found in art, and Ambikā's connection with Nemi is in art probably restricted to a few serial representations of the Jinas.
The case of Padmāvatī is different. It seems to be far-fetched to deny that an attendant goddess of Pārśva is Padmāvatī, if she is provided with a hood-circle. However, the hood-circle is less specific than, say a bird vāhana, and we would apply in the case of Padmāvatī (and generally speaking) the following rule: an attendant deity can only be identified in terms of the system, if he or she agrees in a significant manner with the description of a figure in the texts, and if he or she is shown in the context prescribed by the texts. The correct context is established through the Jina (yakṣī shown along with the Jina who demands such a yakṣī) or through the partner (yakṣī shown along with the yakṣa who demands such a yakṣī). Dhaky Sā: pl. 25 (Dharaṇendra and Padmāvatī) shows a fairly satisfactory example (Dh. with three hoods, P. with one hood). The attendant deities of a dilapidated Pārśva image at Bhilsa (Mevissen Co: fig. 7) may also be mentioned (Dh. intact, P. blurred). Without going into details we add that the triad Y-Y-Jina cannot be identified in toto on the strength of one single member: Pārśva does not identify a female figure as Padmāvatī (and so on). A goddess with child (!) and hood-circle has been published in Mevissen Co: fig. 19.
The yakṣa list is even more artificial than the yakṣī list. Kubera was now rejected and Gomukha became an important figure as the yakṣa of the first Jina. Refer for Gomukha as yakṣa to JID: figs. 220-221. An independent Jaina Gomukha with bull cihna is found at Ellora (Pereira Mo: 172, plan; 178: drawing). Independent images of the Jaina and Hindu Gomukhas have been found at several places (Misra Ya: 127-128; Misra Ya: figs. 71-72 and 78-79; Nagar Wo: pl. 40 and 62). Eight Gomukhas appear on the facade of the Ādinātha temple at Khajuraho (Tiwari Ād: 139). See also Bhattacharya Na (Nandin in the Hindu tradition).
Jaina Gaṇeśas with a hood-circle may exist. If so, they are inspired by the Śvetāmbara description of Dharaṇendra as elephant-headed. But it seems that so far only a modern representation of this strange form of Gaṇeśa has been published (Cort Ga: 88). The evidence of Karnataka points to the reality of Dharaṇendra and Padmāvatī as a couple and probably also as parts of the system (Pal Ex: 182-183 [Dh], 187 [Dh-and-P]). An analysis of the available articles on Pārśva images with peculiar attendant goddesses and gods would be useful. Refer for the Jaina Gaṇeśa to § 6.
Kubera has disappeared. As we have seen, the companion of Ambikā is called in the system Gomedha, Sarvāhṇa, Gomeda (all Digambara) and Gomedha (Śvetāmbara). There is no connection with Kubera, unless the vāhana in the form of a man is viewed as reminiscent of the Hindu god (“Naravāhana“). The YY of Mahāvīra are, however, related to K-and-A on account of some iconographic elements (Nirvāṇakalikā fol. 37b). The mere name “Kubera“ is carried by the yakṣa of the 19th Jina Malli/Mallī (female in the case of the Śvetāmbara tradition). This Kubera has no connection with the old Kubera (the elephant vāhana of no. 19 being irrelevant). - The yakṣas nos. 2-22 and 24 probably never appeared in art.
No doubt, the mere idea that yakṣas and yakṣīs of different Jinas looked different, in opposition to the well-defined and wide-spread K-and-A formula, had to be accepted by the artists in principle. If two or more Jinas were represented together (series, etc.) it was unavoidable to differentiate the attendant deities. But this rule did not encourage the artists to differentiate according to the texts (or according to then-own ideas). Even the Jinas themselves were not distinguished regularly and systematically (cihna, etc.), and there was no attempt to give equal weight to all Jinas (a correctly characterized Jina no. 2 would in such a case be as frequent as any other characterized Jina - nos. 1, 3, 4, etc.). The project of representing the YY was thus doomed to failure right from the beginning. It was not in keeping with the spirit of Jaina iconography and it did not change this spirit.
The names of the mahāvidyās recur with two exceptions among the Digambara names of the yakṣīs (Deleu Ji: 265, 278). In other words, the names of the yakṣīs are largely derived from earlier traditions (mahāvidyās and Ambikā). Refer for the standard versions of the mahāvidyā list (Digambara and Śvetāmbara) to Bhattacharya Jn: 124-132 and Shah Ma: 119-120. There were even two Śvetāmbara hymns which linked the Jinas, not to the YY, but to the mahāvidyās (Shah Ma: 120). The mahāvidyās appear mainly in medieval Śvetāmbara art (Lohuizen He: 114-116; Singh We: 50a, 208-210), but they are also found on several earlier Śvetāmbara Jinas in bronze where they are shown in addition to K-and-A (§ 3 supra). The influence of mahāvidyās (following known descriptions or not) is also felt in Digambara art (§ 5). The numerous mahāvidyā lists, standard and nonstandard, and the earlier spate of vidyā.s form a subject in its own right (Shah Ma: 114-120, et al.).
Every reader can calculate for himself the total of all possible Jina-yakṣa-yakṣī and mahāvidyā data (Digambara and Śvetāmbara). A complete philological edition of all the lists is out of the question, but a short and systematic description of the textual evidence would be useful. As far as orientation on the basis of current research is concerned, we refer the reader to Glasenapp Jn (pp. 405-406, 532-533, pls. 24-27), to Tiwari El (pp. 124-137), and to Deleu Ji (264-266). - Misra Ya: 172-175 should be mentioned as an attempt to coordinate the relevant śilpa data (colour, cognizance, etc. of YY) as appearing in three different texts (Rūpamaṇḍana, etc.).
The artificiality and the largely theoretical character of the YY concept is nowadays alluded to in some publications. Sooner or later, books and articles on Jaina art will treat the entire YY ensemble as a quantité négligeable. It may be asked, why the shadowy character of the system was not noticed earlier. Everybody doing field-work relevant to Jaina art realizes the difference between his material and the system. On the other hand, there is not only the old mental habit of seeing unity (“or unity and diversity“) in Indian art and culture - images and texts must be “one“. There are also the above-mentioned points of agreement between Jaina art and the system. Three or four cases (Gomukha, etc.) can, however, not change the general situation. Late Śvetāmbara art (YY after A.D. 1200?) possibly shows a growing interest in the texts, and a pointed investigation may be useful. But positive findings would also not affect our evaluation of the general development.
What is perhaps more important than the exact extent of agreement is the principle to study, in the case of text-image differences, always both sides independently. Nor is this principle restricted to iconography. It would also be useful (and not less important) to compare the system of Indian poetics (theory of comparison, theory of rasa, etc.) to the deviating situation in the kāvya texts, and to develop for these texts a technique of analysis (poetic figures and other elements) which is independent of the existing theoretical works (Kāvyādarśa, etc.). In both cases (art and alaṃkāra) it is important to undertake “in toto comparisons“ of both sides. Otherwise, the individual cases of agreement will be overvalued: they are only sporadic and not paradigmatic. See also Bruhn Gr I (§ 1) and § 8 below.