4.3 Comparisons to the Arts of Central and South Asia (Continued)

4.3.3 Art from the Indian South: Amaravati and Nagarjunakonda

Amaravati is situated on the banks of the Krishna river in southeast India (Fig. 4b). The site witnessed an extensive occupation from the time of the Mauryas until the 14th century CE (Knox 1992:9). However, Amaravati art is most famous for carved reliefs on the remains of the so-called Great Stupa, which is divided by some scholars into an early style (3rd century BCE - late 1st/early 2nd century CE), of which there are few remains, and an High Period or classical style (2nd century CE - 3rd century CE) (Knox 1992:32-33).

Hackin compared the depiction of a purnaghata on one of the ivory plaques from Begram to the art of Amaravati, in particular to an execution of the asoka tree (Hackin 1939:19-20). However, in the same paragraph, he also remarks that there are more similarities with Mathura, so this particular Amaravati comparison becomes attenuated. Hackin goes on, though, to identify a strong analogy to Amaravati in the treatment of the female figurines on one of the most beautifully carved Begram ivory plaques: the top of a footstool or table called ‘coffret’ IX (Plate 57). The physical features -- the short nose and less (than other Begram figures) elongated eyes called ‘unique’ by Hackin -- and the slender curvature of the body described as ‘purely Indian in its gracefulness’, are compared by Hackin to the physical types within the art of Amaravati (Hackin 1939:20). Such comparison can be supported chronologically since, as mentioned previously, according to Stern (1954), the Begram ivories are to be dated between the ‘Old Style’ of Sañci and the later, and longest, phase of the Amaravati style.

Although other than the slender bodies of some of the Begram figures there are few similarities in the depiction of human figures between the Amaravati and Begram carvings, several other types of motifs on the reliefs from Amaravati show analogies to the art from Begram. A border on a railing pillar from Amaravati dated to the 2nd century CE (Knox 1992:38) shows a type of vine scroll, within which floral and faunal motifs are depicted (Fig.132), which is very reminiscent of the vine scroll depicted on a Begram plaque (Plates 410-413). The particular type of lotus flowers shown on another Amaravati fragment of a railing pillar (Fig.133) dated to the 1st - 2nd century CE by Knox (1992:46), too, is reminiscent of the lotuses shown on a Begram example (Plates 245 and 248). Another motif that compares favorably is the depiction of a full blown lotus: both the Begram (Plate 350) and Amaravati pieces (Fig.134) show tiny circular dots in the center kelk of the flower, as well as a spoked design emanating from this center. A garland spouting out of the mouths of makaras depicted on a relief from Amaravati (Fig.135), although not stylistically similar, is reminiscent of a similar garland consisting of a series of makaras from Begram (Plate 298). Other ivory and bone plaques from Begram depict small figures supporting wavy garlands on their shoulders, a motif also known from 1st century BCE Amaravati (Fig.136). Auspicious symbols such as the triratna, dharmacakra, and svastika, present on a few Begram plaques, can also be found in the early, 1st century BCE - 1st century CE, art of Amaravati (Fig.137).

Hackin was not alone in his view of analogies from Amaravati. When P. Hamelin was reconstructing the Begram objects, he, too, turned to the art of Amaravati for similarities. Stern noted several analogies between thrones depicted on the Amaravati reliefs and those from Begram. One such Amaravati piece depicts a man and a woman seated on a large throne, with one foot resting on a small stool in front of the throne (Fig.138). The decoration of the throne consists of horizontal bands, each of which alternatingly depicts a row of small banisters and a plaque, very much in the same fashion as Hamelin reconstructed the Begram furniture (Fig.72).

These connections, however, must all be considered rather general and indirect. Although the art of Amaravati shows analogies to the Begram carvings, it does so less consistently than the art of Mathura. Moreover, the analogies mainly concern subject matter and provide few stylistic correlations. Even if one could argue that most of the compatible motifs from Amaravati are from the early period, i.e. 1st - 2nd century CE, and thus contemporaneous with Mathura and Begram, this would provide no more than an indication of general artistic tendencies throughout Indian art, and not a distinct geographical placement of the Begram carvings or a direct connection to stylistic sources.

The art of Nagarjunakonda, situated close to Amaravati, is another ‘style’ that has been compared to the art of Begram (Fig. 4b). Nagarjunakonda is mostly known for its remains from the 3th and 4th century CE, when it was under the rule of the Iksvaku dynasty, and which includes secular, Hindu and Buddhist edifices (Rosen Stone 1994:1-20). According to E. Rosen Stone, the Nagarjunakonda style provides the intermediate stage between the Amaravati style of the 2nd century CE to the Gupta style of the 4th century CE (1994:21).

E. Rosen Stone published two works on the Begram ivories [1] (1974-75; 1994). In the first, she very strongly argued in favor of placing the Begram ivories in the latter part of the 3rd or early part of the 4th centuries CE (Rosen 1974-75:45). Rosen reached this conclusion by comparing the “degree of interplay” on the Begram ivories (Plates 96, 97, 98) with those of mithuna couples from Nagarjunakonda (Figs.139-140). Rosen Stone further cites a relief (Fig. 141) in which a man sits on a throne holding an elongated piece of wood in his hands; Rosen Stone identifies this piece as “Siddhartha stringing the mighty bow” (Rosen 1974-75:41), and it reminds her, even though she notes that the two pieces do not compare well at all, of a depiction from Begram of a woman, holding a necklace in front of her, and her attendant figure (Plate 1). In this first article, Rosen Stone even goes so far as to say that this Begram version is a direct borrowing of the Buddhist plaque described above (Rosen 1974-75: 41). In her second paper, however, Rosen Stone slightly retreats from this position and calls the relationship reciprocal (Rosen Stone 1994:94) .

The main comparison drawn by Rosen Stone between Begram and Nagarjunakonda is from the so-called ‘Coffret IX’ (Plate 57). She notes:

“Such figures (from Begram) would derive ultimately from types found in the middle phases of the art of Amaravati......The coiffures represented on Coffret IX are not found in the art of Sañci or Mathura. The chignon, appearing somewhat like a pony-tail, combined with the use of the bow in front is well known in the art of Nagarjunakonda and also appears in the art of Amaravati. The hip girdles, bracelets and ankle bangles are consistent both with Nagarjunakonda and its antecedents......On the corner of the main plaques are grilli or composite animals which have been identified as having an ultimate Iranian prototype, and which were used in the West until the third century A.D.”(Rosen 19:43-44)

In trying to provide a time frame for the Begram ivories which would fit -- that is, permit a direct relationship to -- the Nagarjunakonda dates of the 3rd - 4th century CE, Rosen Stone compares other elements, such as the vine-scroll on ‘Coffret IX’, to mid-third century Roman examples.

Despite these efforts at establishing a direct ‘reciprocal’ relationship between Begram and Nagarjunakonda, all the examples cited by Rosen Stone have antecedents, as she notes several times. The only comparative elements which Rosen Stone offers which seem to have no antecedents are the pony-tail type of hairstyle worn by the figurines depicted on ‘Coffret IX’ (compare Plate 57 and Fig. 142) and the ‘degree of interplay between the figurines of the Begram carvings’. As discussed below, however, variations of this type of hairstyle do indeed exist in the art of Gandhara, as in the art of Sañci. And, while there may be some similarity in the ‘degree of interplay’ between the figures, this may very well be the result of nothing more than the fact that these scenes belong to the realm of the secular mundane, unlike the reliefs adorning the religious edifices at Mathura and Sañci, and that such interplay can be seen in portrayals of daily life from other periods and styles as well.

The fact remains that all of the motifs cited from Amaravati and Nagarjunakonda have their antecedents in earlier Indian art. Even if one could make a much stronger case than has been made for the above-mentioned stylistic parallels, there is no evidence that the motifs on the Begram carvings were not in fact earlier prototypes for some of the motifs on the Nagarjunakonda reliefs or at least derived from common antecedents rather than existing simultaneously and ‘reciprocally’. Based on Begram’s archaeological context, the numismatic evidence, the two other categories of finds, Graeco-Roman and Chinese ware, and the comparative ivory and bone objects, the mid-3rd-4th century CE date of Nagarjunakonda is too late for the Begram ivory and bone carvings to have been produced within a contemporary and shared artistic sphere.

4.3.4 The northern connection: Gandhara and Bactria

When discussing the art of Gandhara, scholars generally refer to the period from the 1st to the 4th centuries CE when the so-called Graeco-Buddhist art styles came into being. But the region of Gandhara is actually quite a bit older, deriving its name as one of the conquered regions mentioned by Darius in the Bisutun inscription, and subsequently annexed as one of the Achaemenid satrapies from the 6th to 4th centuries BCE. The region includes parts of northern Pakistan and eastern Afghanistan; Begram lies at its extreme northwestern fringes (Fig. 4a).

The art of Gandhara was greatly influenced by a succession of styles and, eventually, underwent a wider and more intense fusion of Indian, Near Eastern or Iranian, and Greek and Roman motifs than did the more southern regions of Pakistan and India. To this extent the artistic history of Gandhara is more analogous to that of contingenous areas of Central Asia. With regard to this eclecticism and a similar fusion of styles in the Begram ivory and bone carvings, it is somewhat surprising that no scholar has thoroughly examined the possibility that at least some of the Begram carvings could have been produced in Gandhara. This is an omission which this section will begin to redress.

A number of factors support the view that the provenance of at least some of the Begram bone and ivory objects may have been Gandharan. In the first instance, the northern location of and probable point of distribution for the Begram ivories raises the obvious, but to this point ignored, possibility that since the market for the goods was located in the north, that is also where the craftsmen would have been situated. There should be nothing startling in itself about the suggestion of ivory artisans working in the northwest; Taxila, also located in Gandhara, has often been referred to as the site of an ivory carving school (Dwivedi 1976:94-95).

Secondly, although the majority of the Begram ivory and bone finds seem to have analogies with the art of Mathura, both Mathura and Gandhara were part of the Kushan empire and were connected by trade routes along which both artisans and their ideas could have moved. Most scholars seem to recognize without hesitation the intrusion of Greek and Roman motifs into the Indian heartland, but very few have addressed the movement of Indian styles to the north and west. This despite the fact that many clearly Indianesque styles are evident in finds as far north as Uzbekistan.

Finally, certain motifs from the Begram objects find their general placement in Gandhara rather than in Mathura or the Indian South. And even considering the possibility that these styles traveled down from Gandhara, i.e. Pakistan and Afghanistan, to India where they were incorporated into bone and ivory pieces which then traveled back north to Begram, this still confirms the general proposition that Gandharan styles undeniably contributed to the eclectic nature of the depictions on the Begram carvings.

The motifs on the Begram pieces that point to the area of Gandhara may not appear as often as do the motifs more generally compared to the art of Mathura, but those that do appear present a more direct analogy than the Mathura art does. The Kaniska reliquary discovered in Shah-ji-ki-dheri, Peshawar, presents perhaps a closer analogue to a motif on the Begram objects than any other representation in sculptural Indian art. The reliquary consists of a small circular box (12.7 cm dia.) made of copper and has been dated to the 2nd century CE (Fig.144). This date is by no means certain, however, and depends on whether it is considered to be dated in Kaniska’s time or dedicated by a later Kushan ruler in Kaniska’s memory. If one dates the reliquary in Kaniska’s time, then the object could be dated to the early 2nd century CE [2] , which makes it, more or less, contemporaneous to the Begram objects. Three small statues of the Buddha, Indra and Brahma are positioned on the top of the lid of the box and function as the handle. On the rim of the lid, a series of ducks or geese in full flight are depicted carrying large round wreaths in their beaks. These birds are almost identical in execution to the ones depicted on several of the Begram objects (Plates 227-232).

Another point of considerable similarity to the art of Gandhara can be witnessed in the execution of the bead-and-reel motif border of one of the Begram plaques (Plate 57). On a stone relief from the Peshawar region in Gandhara, a woman is depicted from behind, standing with her legs crossed and one heel raised, the latter a similar posture to many of the Begram figurines (Fig.145). Above her head flowers are depicted, to which her head is raised as if to smell their scent. The scene is set in a rectangular frame decorated with a bead-and-reel motif analogous to the bead-and-reel design on the carved ivory plaque from Begram. Other images from Gandhara or the northwestern regions display this bead-and-reel motif, including two stone reliefs from Shotorak in the vicinity of Begram (Figs.146-147). The way this bead-and-reel motif traveled may be easily understood when viewing Graeco-Bactrian coins, which display the same motif (Fig.148).

A signal motif present on the Begram ivory and bone objects is the depiction of a yak∑a figure with wings (Plates 398-399). The Begram examples depicted a small squatting figure supporting purnaghatas with their hands or on their heads. This type of figure is virtually unknown in India; indeed, wings are absent in humanoid figures and are only seen as an attribute for garuda and on fantastic creatures. The only exception of an example from India of a figure with wings is a figurine shown on a terra-cotta plaque from Basarh, dating to the 3rd century BCE (Coomaraswamy 1985:21). Unlike the regions farther south, however, the winged yak∑as can be found throughout Gandharan and Bactrian art, as well as in the rest of Central Asia, from Kucha and Miran.

Although architectural edifices found in the Begram plaques (Plates 86-91) may have their origin in India, the motifs are also present in Gandharan art. The alternating Indian arch and Sañci type torana are shown on edifices from the Sirkap mound in Taxila (Fig.149). A type of garden pavilion shown on a Begram plaque (Plate 128) is very reminiscent of edifices found on stone reliefs from Gandhara (Fig.150): a structure consisting of a horizontal roof with alternating sunken and jutting-out square pieces supported by columns with Graeco-Roman capitals.

The hairstyle of the women depicted on one ivory plaque from Begram shows tiny bows on top of their head (Plate 57) and is similar to the headdress worn by some Bodhisattva figures from Gandhara (Fig.151). Also in plaques from the same ensemble (Plates 54, 55, 56) are women shown with hair crowned by what looks like a twisted piece of cloth, from the middle of which stick out three pointed, leaf-like ornaments. This is a stylistic element known only from the Gandhara region, and the tiara is known mainly from sites in Gandhara and Bactria but is not a common jewelry type in Central and South India.

The adaptation of Hellenistic styles is also witnessed in the dress of one of the large ivory figurines from Begram (Plate 200). The short sleeve tunic is very similar to those found in Gandhara and Bactria. In addition, in one of the Begram plaques a woman is shown standing on a circular drum wearing a long Hellenistic tunic with sleeves (Plate 56). What differs in this case from both Graeco-Roman style and the Indian styles of Mathura and Sañci is that the woman is wearing trousers underneath her tunic. Women wearing pants are, however, frequently depicted in Gandharan art, from which fact one could deduce a direct local nomadic element in the region. Women interred in the somewhat more northern Tillya-tepe necropolis of the proto-Kushan period are also believed to have worn such pants (Sarianidi 1985:238). In this regard, it is interesting to note a particular depiction of a hunter on horseback from Begram (Plate 187): round appliqués seem to have been sewn on his trousers, a technique well known in Saka-Parthian and Kushan costumes from Bactria (Sarianidi 1985: 226), but not from the south.

Another interesting analogy is presented by the small supporting figures carved on the Buddhist monastery of Jaulian at Taxila (Fig.152) [3] . The squatting position and the position of the legs of the figure on the far left is very reminiscent of the posture of some vyala-yaksa supporting garlands, as depicted on the Begram carvings (Plates 295, 296, 297).

That styles in the Gandhara region were greatly influenced by Hellenistic styles existing in the early schools of Bactria is evident from comparisons between those two. In addition, the archaeological evidence from Bactria also indicates stylistic movement in the other direction, that is, the penetration of Indianizing motifs northward into the region from the 1st century BCE onwards. The Tillya-tepe necropolis, dated somewhere in the 1st century BCE or early 1st century CE, provides several objects which demonstrate the eclectic nature of the arts of the northern regions of the period and offers good comparisons to the art of Begram. Aside from the ivory comb discussed in section 4.1., the Tillya-tepe necropolis also yielded a gold coin depicting on one side a man turning a wheel, and on the other side a lion and a triratna symbol (Fig.153). The penetration at that time of the triratna symbol so far north is of significance for an early dating, i.e. 1st century CE, of the Begram objects, which also show the triratna on several plaques (Plates 389-393). In addition, the Kushan Aphrodite from Tillya-tepe shows the characteristic Indian dot on the forehead of the tiny figurine (Fig.154).

The objects from Tillya-tepe also offer stylistic analogies to the ivory and bone objects from Begram. One metal plaque depicts a figure clutching two dolphins in his arms (Fig.155): he wears a vegetal loin cloth and in one hands he hold a fly whisk. On his head is a motif of tiny dots very similar to those on the figure depicted on the ivory comb found in Tillya-tepe (Fig.111), which may indicate a regional workshop for both objects. The figure of the man and dolphins is analogous in composition to the vyala-yaksas shown on the Begram carvings (Plates 285, 286, 287) and also wears a vegetal loin cloth. It can be suggested that, although in the latter examples the dolphins were replaced by makaras, the Tillya-tepe example presents the earlier prototype for the Begram images.

The depiction of a man clutching two winged creatures on a piece from Tillya-tepe (Fig.156) can also be compared to the carving from Begram of a male figure clutching the bodies of two winged creatures flanking him (Plate 207). There is even a slight similarity in the upper part of the tunic. And another Tillya-tepe piece provides an actual object that appears on some of the Begram carvings (Fig.157). A pair of golden earrings the decorative part of which consists of a small square piece below which is a larger heart-shaped medallion, both inlaid with turquoise, is almost an exact duplicate of the earrings worn by some of the women in the Begram carvings (Plates 51 and 57).

Other motifs in the arts of Bactria which may have been the inspiration for some Gandharan, and thence Begram, motifs include the type of marks on two winged creatures from Begram (Plates 300-301) which are very reminiscent of the inlaid decoration of some metalwork from Bactria. Also, the depiction on Begram objects of eagles, which are not a common motif in Indian art, were present in Bactrian art and may have been transmitted through contact with the Roman eagle standard.

The Grylli motif on two Begram objects also was well known in Central Asia in pre-Kushan times, as witnessed by finds from sites such as Koj-Krylgan-Kala in Choresmia (Fig.158). They may have ultimately been derived from Near Eastern prototypes, such as the lion-fish or goat-fish figures, though they were also known in Hellenistic and Roman art.

4.3.5 Conclusion

The reasoning behind the use of stone relief comparisons for the Begram ivory and bone objects is obvious. The scarcity of ivory finds in relation to other types of more lasting materials certainly seemed, at least before the discovery of analogous ivories from Bactria (see section 4.1), to have pointed to stone reliefs as the only source for comparison. And this approach was somewhat supported by epigraphic references, such as an inscription from Sañci, which seemed to indicate that some ivory carvers may have worked in both media. In addition, the recognizable art schools of Mathura and Sañci demonstrated some commonalties, at least in terms of general, regional styles.

However, a number of problems exist with using stone reliefs as comparative material. Some have been previously noted, such as the fact that differences in style may be due to the lesser scope but higher precision afforded by ivory, and that the mundane aspect of the subject matter depicted on the Begram ivory and bone carvings may have left room for freer adaptation of styles and motifs. There are additional problems, though. First is the issue of whether stone reliefs or monumental art always precede smaller arts and crafts; all previous comparisons seem to assume that the stone reliefs are the seed which germinated the ivory and bone carvings. It seems reasonable to assume, however, that even if a particular monumental art piece preceded similar ivory pieces in a different region, since monuments do not travel, the long-distance diffusion of the motifs may have been the result of small portable examples being exchanged along trade routes. The decorated Graeco-Roman glassware from Begram dated in the 1st-early 2nd century CE (Whitehouse 1989) is a prime example of how Roman elements may have come to be incorporated into Gandharan art. Similarly, the diffusion of Indianesque motifs into Central Asia could have taken place through small, portable, objects such as the Begram ivory and bone carvings.

A second problem concerning stone reliefs as comparative material is whether one can assume that the prevalence or absence of a certain motif in contemporarily constructed stone sculptures is determinative of what was incorporated in other, smaller objects such as the ivory and bone objects from Begram. Stone monuments tend to last for a long time, and there seems no reason to assume that some long-standing motifs in stone reliefs might not have triggered the imagination of an artisan much later than the period of the construction of the monument. In this regard, the art from Amaravati and Nagarjunakonda seem to demonstrate that several motifs and subject matters continued to exercise stylistic influence for a long time after they were no longer dominant.

Yet another problem is that we know very little about the artisans who worked on ancient monuments in India. The sheer size of the monuments make it reasonable to assume that many artisans from different regions may have contributed to the creation of the edifices. Thus, terms such as ‘Mathura school’ to delimit a regional art style may be more complex than common usage suggests.

Notwithstanding all these caveats, it can be suggested that the styles of the ivory and bone carvings from Begram are not only closest to the general styles used in 1st-2nd century CE Mathura sculpture, but that they also have many analogies in the art of Gandhara. The distribution of Begram-type ivories in Gandhara and Bactria, plus the use of Kharosthi script on the pieces, suggest a Gandharan workshop with artisans well acquainted with the subject matter and styles of Mathura art. The dating of comparative ivory and bone objects and of the two other categories of finds, plus the architectural and numismatic evidence, all point to the 1st-2nd century CE (see Chapter 5) for the production of the Begram carvings, dating that would make direct comparisons to the art of Amaravati and Nagarjunakonda too late, but which do not obviate the probability that the Begram pieces shared some common antecedents with these later styles from the Indian South.

Website Maintained by:
Information Systems and Services,
International and Area Studies,
UC Berkeley

Electronic Cultural Atlas Initiative
URL: ECAI.org