A Prolegomenon to the Visual Language of Dance in Gandhāra

Abstract

Pre-modern Indian subcontinent provides a treasure trove of art historical data in the form of stone sculptures and reliefs to study dance. While significant steps towards understanding the literary and visual language of dance have been made, artistic production from Gandhāra (the ancient region broadly covering the northwestern part of the subcontinent) largely remains absent in scholarly discussions. Ancient Gandhāra readily lends itself to a global approach as an active participant alongside the so-called ancient Silk Roads connecting the Mediterranean regions with China. Furthermore, as part of the Buddhist pilgrimage routes, Gandhāra also developed ties with Buddhist sites located further east and participated in the spread of Buddhism to China. Within this context, this article discusses the most common dance depicted in Gandhāran art to understand how artists represented dance in the static medium. Using this dance as an illustration, this article also argues that the iconographic conventions of the Gandhāran artistic repertoire for dance are shared outside the region, notably in Kizil, which is located alongside the northern branch of the Silk Roads.

1. The Language of Dance in the Indian Subcontinent

It is customary to begin any systematic study on dance in the Indian subcontinent with the earliest authoritative texts codifying performances. As the most important text, the Nāṭyaśāstra or the ‘Treatise on Theater’ has been the object of dance-related studies since the 1930s.1 The Nāṭyaśāstra consists of thirty-six or thirty-seven chapters addressing different aspects of theater and can be dated roughly between the second century BCE and the second century CE. Its authorship is attributed to the mythical Bharata who introduced the art of nāṭya (sometimes translated as dance-theater, drama or play) to humans to combat the moral degradation of society. This nāṭya comprised different elements such as rituals, music, and dancing.

Dance or nṛtta is dealt with in Chapter 4 of the Nāṭyaśāstra using highly technical instructions and mythological narratives. Often placed in opposition to mimetic dance (nṛtya), nṛtta is defined as an abstract dance that does not contain narrative content.2 The Nāṭyaśāstra outlines how nṛtta consists of 108 karaṇas (units of dance) comprising limb movements and hand gestures (hasta/mudrā).3 The various combinations of the karaṇas provide larger choreutic sequences called the aṅgaharas, which consist of 32 linear measurements, poses and stances. These sequences, when they appear in other mediums, such as paintings and sculptures, can be identified as dance.4

The study of art historical evidence in the light of Nāṭyaśāstra’s karaṇas has certainly led to the development of nuanced methodological approaches. Traditionally, karaṇas were thought to be static poses; however, it is now accepted that they are part of a whole movement consisting of a beginning, the course of the movement, and an end. This is particularly useful when studying the representation of dance in sculptures where only static poses can be identified. On this understanding, Vatsyayan astutely remarked that ‘the plastic can capture only a single moment in a continuous flow of movements and only suggests through the arrested image the moment before or after’.5 If the karaṇas in plastic arts are a series of movements, then images represent a snapshot of different moments associated with dance.6 This allows for an unlimited number of representations of dance movements in sculptures that stem from the different stages of the 108 karaṇas.7

Such codified dance movements found in texts from the Indian subcontinent find a limited echo in the corpus of Gandhāran art, which is commonly associated with the ancient region of Gandhāra (Figure 1). The corpus of Gandhāran art consists mainly of schist and stucco reliefs and statues dating broadly from the first to the fourth century CE.8 The dance movements found in Gandhāran reliefs do not correspond to the descriptions found in Indic texts. However, viewing ‘petrified’ static poses of dance in art as part of dynamic performances is useful in our context where the primary source of information is stone reliefs. In order to systematically study dance, this article will show that the representation of dance in Gandhāran art was highly motivated by particular visual conventions. When studied together, reliefs belonging to diverse geographical and chronological periods present a limited repertoire of movements associated with dancing figures.

As a prolegomenon to dance in Gandhāran art, the present article examines one form of dance executed by both male and female figures on schist reliefs.9 Multiple forms of dance are depicted in Gandhāran art, such as the ‘Persian snap’, and this article engages with a dance that I provisionally label as ‘Indic’.10 The iconographic conventions related to the Indic dance studied in this article have a strong relationship to dance imagery from major sites such as Sānchi and Bhārhut. In comparing the representations of our Indic dance with images from other Indian Buddhist sites, we can see that the dancing figures wear similar costumes and perform variations of the same movement.

I wish to illustrate this striking similarity using two examples: Figure 2 and Figure 3. See the dancer centrally depicted in Figure 2, which is from Gandhāra, and likely from the Swāt Valley, and the two dancers on the bottom register in Figure 3, which is from the Bhārhut stūpa: these figures execute the same movement. Moreover, both these dances occur in courtly settings, although their purpose varies. In Figure 2, a female dancer performs in the palace of Siddhārtha and in Figure 3, female dancers perform in the palace of Indra. We will encounter Figure 2 once again in this article, so let us turn to an interesting detail in Figure 3. An inscription found on the dome of Indra’s palace identifies the scene as the cūḍāmaha (literally, the festival of the hair crest). It suggests that the dance is taking place as part of a festive celebration. The focus of the dance is the hair relic, indicated by Prince Siddhārtha’s turban, located under the inscription.11 Despite the different purposes of the dances, even a cursory glance at Figure 2 and Figure 3 reveals that the conventions used to depict dance in these images are nearly identical. Only stylistic and aesthetical elements related to the figures’ costume and jewelry differ, but these can be explained based on regional choices made by the artists.

Due to their similarities, I have tentatively labeled the reliefs depicting the variations of this dance movement as ‘Indic’. This label does not presuppose the existence of an authentic performance group that was associated with a specific Indian ethnicity. It simply allows us to define a distinct corpus of data and contrast it with other dance forms that have already received some attention in scholarly studies. Representations of this Indic dance are the focus of the first section, and so the label should be viewed simply as a shorthand to differentiate between numerous dance movements appearing in Gandhāran art whose origin and existence cannot easily be clarified using the available data.

In the second section, Gandhāran reliefs will be compared to a wall painting found in Kizil with a dancing female figure. This section will argue that at least some iconographic conventions related to the representation of dance from Gandhāra may have also been adapted by artists along the ancient Silk Roads. This comparative study provides yet another dimension to the dynamic interactions occurring between Gandhāra and Kuča via the long-distance religious and mercantile routes.

I am aware that the scope of this article only allows me to address a handful of examples. They are not meant to be representative of all dance movements found in Gandhāran art. They can only provide a small glimpse of dance based on the visual corpus of this astonishingly multicultural region. A complete inventory of dance movements in Gandhāran art is still a desideratum. While I eagerly await such a study, my modest goal in this article is to present a preliminary analysis of selected reliefs for further exploration.

Before examining the representation of dance in Gandhāran art, it is important to establish what I mean by this term. Dance is intentional, rhythmic, and stereotypical bodily movements that can be observed and identified as a special category of behavior. In Gandhāran art, the presence of abstract dance can be detected based on the position of the body and the limbs of the figures performing it. The dancing figures in art are usually depicted frontally; however, some images present the dancer from the back, perhaps indicating turns and perspectives. In visualizing our Indic dance, the figures commonly stand with their lower limbs crossed or one of them lifted upwards. For example, in Figure 4, a female dancer performs alongside a male musician with a harp. The female dancer has her lower limbs crossed in a posture that is commonly associated with dancing figures. The dance occurs on a relief depicting the Candakinnara jātaka, in which a kinnara (mythical beings) couple’s tranquil life is tragically interrupted by a lustful king. The kinnaras harmony in their forest abode is communicated by their dance and musical performances.

Regarding the upper limbs, dancing figures usually hold one or both arms away from the body. We must keep in mind that non-dancing figures also perform gestures that are similar to dancing figures. However, they are rarely accompanied by musicians, which is the case when dance is represented. Moreover, the arrangement of the dancer’s body normally deviates from its central axis based on their limb movements. Such an arrangement can be favorably compared to the tribhaṅga or the triple bend pose. With this background, we will see in the following sections how the Indic dance in Gandhāran art comprises a limited (re)combination of arm and leg movements that are repeatedly highlighted by the reliefs.12

2. Finding Dance in Gandhāra Art

The Indic dance in Gandhāran art commonly occurs as part of courtly entertainment, festive celebrations, and decorations of panels depicting the Buddha’s life story. The most common context within which this dance can be identified is the palace of Prince Siddhārtha, the future Buddha Śākyamuni. The visualization of Siddhārtha’s luxurious palace life, such as in Figure 2, often includes dancers and musicians who entertain the prince.13 Usually female dancers, accompanied by musicians, perform in front of the Siddhārtha and his wife, Yaśodharā, and form part of the sensual world that is later abandoned by the prince. Similarly, dance also occurs as a form of courtly entertainment for the nāgas (serpent divinities having a human form with a snake hood, Figure 5). Since the nāgarāja (nāga king) is a royal figure in his own realm, he is often entertained by nāginī musicians and dancers.

As we have already seen in Figure 3, which comes from Bhārhut, dance can also be identified in Gandhāran reliefs depicting festivals and processions. During such events, dance occurs around a range of objects such as the Buddha’s turban, stūpas (hemispherical monuments sometimes containing relics), reliquaries, and the Buddha himself (Figure 6).

Besides these two contexts—courtly and festive celebrations—dancing figures also decorate architectural elements such as frames of niches and door jambs. These figures, both male and female, as we will see, perform the same variations of the Indic dance movements found in the other two contexts despite their decorative function.

Some of the datable reliefs with dancing figures in Gandhāran art come from Butkara I, one of the largest and longest-used Buddhist sites in Gandhāra. Relative dating and stratigraphy suggest that the site was used from the early third century BCE to the tenth century CE. Its foundations may have been associated with the Dharmarājikā stūpas built by the Mauryan King Aśoka (c. 304–232 BCE), and its later enlargement may be related to the Oḍirājas, a group of local Gandhāran potentates who controlled the Swāt Valley in the first century CE.14 The site, built mainly using local schist, is dominated by a large stūpa called the Great Stūpa. The systematic development of the site can be understood based on the different building periods of this Great Stūpa.15 During these building periods, the Great Stūpa was surrounded by over 227 smaller monuments including stūpas, vihāras (shrines), and columns (stambha), some of which were decorated with reliefs.

In Butkara I, at least six bas-reliefs preserve figures performing the Indic dance with relatively few variations. They can be broadly dated between the first and the third century CE based on their stylistic features. Four out of six reliefs depicting dance preserve small-sized figures within architectural settings. We will see how this architectural frame enhances the impression of the movement and restricts it to the available space. In these cases, the figures are also placed in relation to musicians either within the same architectural frame or in associated registers.

The most elaborate example of dance from Butkara I is a relief depicting two male figures under an elaborate arch (Figure 7). It is part of the category of images belonging to the stile disegnativo or drawing style. This style can be dated to the first century CE which coincides with the Saka-Parthian era as well as the local Oḍirājas.16 The main characteristics of this style are the frontal depictions of figures, flat treatment of volume, and dense lines for drapery. According to Filigenzi, these characteristics ‘show the dependence on more or less coeval Indian iconographies, evidently held as authoritative models’.17 This may further explain the striking similarity between the representations of dance in drawing style and in other Indian Buddhist sites.

The relief depicts two male performers standing in front of an elaborate caitya arch, perhaps symbolizing a gateway. Gateways in Gandhāran art are commonly used to frame devotees, both monastic and lay, as they engage in a variety of veneration activities. They occur more frequently in the Swāt Valley, and reliefs from Butkara I and Saidu Sharif I preserve some of the most elaborate versions of these arches. However, the arches have only been identified in art and similar arches have not been unearthed at Gandhāran sites. The lack of caitya arches in the archeological record has led Brancaccio to convincingly argue that they were an imaginary space in art rather than a reference to contemporary architectural elements.18 The artists in Swāt were likely aware of central Indian architectural and artistic traditions and used certain elements to create new visual spaces with meaning. The caitya arches in Gandhāran art act as framing devices for devotional practices and make allusions to relic shrines located in the Buddha’s homeland. As zooming devices, they keep our vision focused on the figures and their actions while simultaneously placing them within a symbolic space located beyond the Swāt Valley. The figures, for their part, perform activities associated with contemporary praxis, and embed the local within the broader Buddhist landscape.

Returning to Figure 7, the two male figures are dressed similarly; they both wear a turban and a pleated lower garment.19 Their elaborate jewelry, such as the long-weighted earrings and necklaces, as well as their rich drapery, reveal their wealthy status.20 The male figure on the left holds an arched harp with a plectrum and strokes the strings of the instrument. The figure to the right performs the dance that we will also encounter in subsequent examples. Moreover, it can also be favorably compared to the representations of dance found in other Indian Buddhist sites such as the dancing figure accompanying the Bodhisattva on a relief from Amarāvatī (Figure 8). While performing this dance movement, the figure’s right arm is arched above his head and his hand reaches over his turban. His left arm is bent close to his torso and his lower limbs are crossed. His body is not haphazardly depicted but is balanced by distributing his weight to the side toward which the face is turned. Such a posture in response to music does not seem to be accidental. We will see how figures represented this way are not in the throes of movement but carefully executing a topical dance.

Several relief fragments from outside of the Swāt Valley dating from roughly after the second century CE present the same dance movement. For example, a relief panel of unknown provenance depicts a musician and a dancer under an elaborate pavilion (Figure 9). The courtly setting of the dance can be inferred based on the architectural elements and the relief may be related to scenes of Siddhārtha’s sensual palace life.21 Moreover, the performance occurs in front of an audience that is also part of the figural space. The presence of well-dressed female figures observing the dance from their balconies further emphasizes that the dance was performed for members of the royal palace. When compared with Figure 7, only a few negligible differences such as the dancer’s gender and musical instrument emerge. Despite these differences between Figure 7 and Figure 9, the dance remains strikingly identical. Their similarity suggests that artists deliberately used the same iconographic conventions to visually depict dance over a long period of time.

The other five reliefs from Butkara I dated roughly between the second and third century CE take up the same movement, using limited variations, to depict dance. Apart from the gender and clothing of the figures, little to no difference can be identified in the arrangement of the limbs and gestures from a technical perspective. For instance, a relief panel divided into two registers depicts a dance in one of them.22 The two registers likely depict two distinct stories from the Buddha’s biography; however, the present state of the relief only allows for a tentative identification. On the upper register, figures stand around an enthroned mound in veneration. We can suggest that the mound represents the relics of the Buddha and is perhaps related to the events after the mahāparinirvāṇa. On the bottom register, a variation of the Indic dance is depicted. On the left, a seated royal couple watches female musicians and a dancer perform. This scene can be favorably identified as courtly entertainment happening in Siddhārtha’s palace. The dancer, depicted on the right, is viewed only from the back and is wearing a pleated lower garment. Based on the perspective of the scene, the dancer faces her audience, the royal couple. Despite the poor condition of the relief, we can reconstruct some aspects of the dance movement. She is depicted with her left leg slightly raised above the ground, and the position of her right shoulder suggests that her right hand was raised above her head. A strikingly similar movement occurs on a relief in which the Buddha is welcomed by a group of musicians and a dancer. Based on its stylistic and iconographic features, the relief likely comes from the Swāt Valley and can be dated to the first century CE (Figure 10). At the center of the relief, a male figure wearing a pleated lower garment performs the same dance movement found in the aforementioned relief from Butkara I. Moreover, the perspectives are also identical; both dancers are turned away from the viewer.

The repetitive iconography of dance can be further illustrated based on other reliefs from Butkara I such as a door jamb relief.23 In its present state, the relief preserves three pairs of figures in different registers. In the middle of the fragment, a pair of dancers perform within an elaborate architectural setting. The pair comprises a poorly preserved male figure with a half-wrapped mantle and a female dancer wearing a pleated lower garment and long earrings. The female dancer’s right hand is raised above her head and her upper body is arched by leaning towards the left. Her left hand is held close to the body. Her lower limbs are also carefully arranged, and her feet are crossed to further accentuate the gracefulness of her movement.

The variations of s this dance are numerous, for example, a door jamb relief which depicts two dancing figures from Butkara I (Figure 11). The relief is divided into three registers. A female figure plays a long flute on the upper register as two female figures perform the same dance in two consecutive registers. Despite the difference in perspective, the leg and arm positions seem to be derived from the Indic dance. To explain this, let us look closely at the actions performed within both registers. The dance, read from top to bottom, is executed by the right side followed by the left side. It is identical and the only difference between the two registers is the side of the limbs that are activated. The two figures have their feet crossed and one of their hands is raised above their head.

The presence of the same dance in two consecutive registerssuggests some effort was made by the artists to depict a sequence. Repeating the same dance starting with the right side and then followed by the left creates the impression of bilateral symmetry. By representing it side by side, the visual language emphasizes the continuous yet uniform actions of the figures.24 In comparison, this symmetry is also highlighted by the two dancers in Figure 3, who execute the same movement on either side of the seated nāgarāja. Such representations are particularly interesting and, as in the case of Butkara I relief, reveal the existence of multiple perspectives of the same dance on smaller decorative figures.

The role of dance as a form of entertainment is evident, but why does dance appear as part of decorations? Buddhist texts identify dance as part of festivals organized to commemorate important events in the Buddha’s biographies.25 Already in the biographical texts, the use of dance to venerate the Buddha shows a blueprint for how it may have become an important element of relic rituals. Following the Buddha’s mahāparinirvāṇa, festivities (mahas) for his relics occurred and the Mallas performed sarīrapūjā26 with dance, song, music, garlands, and perfumes.27 Once his relics were interred within a stūpa, devotees were encouraged to perform similar acts of reverence and gain spiritual merit.28 The visual representation of such dances performed by the Mallas around the Buddha’s relics can also be found in Bhārhut (Figure 12). The relief uses the visual conventions for dance that we have already examined in Gandhāran art. Female figures, presumably belonging to the Malla clan dance and perform music in front of elephants transporting the Buddha’s relics. The dancers and musicians can be seen as part of an elaborate procession organized around the first distribution of relics following the Buddha’s funeral.

While images show us dance, the texts usually do not describe the dance performed by the devotees. However, they repeatedly mention it as an activity related to lay devotees during festivals. In one such festival described in the Avadānaśataka, singing and dancing were so enthusiastically conducted by the devotees that the stūpa was covered by dust after the festival and required further cleaning.29 Our Gandhāran reliefs such as Figure 7 could speak of the joyous and uplifting mood associated with such passionate performances. In light of these literary descriptions, if we consider that Buddhist sites were locations for veneration and performances, figures under arches and other architectural spaces may refer to dance as an offering albeit in a highly idealized manner. Similar to figures making offerings such as lamps and flowers, the male performers in Figure 7 and Figure 11, for instance can be viewed as making an offering of their music and dance. Through their performances, figures in stone venerate the Buddha and his relics perpetually and imbue the sacred area with a festive ambiance.

The use of such evocative festive imagery within the sacred area may also be considered as part of the broader communicative strategy used by the saṃgha. Pagel’s study of Buddhist texts such as avadānas has convincingly demonstrated that well-organized song, dance and theatrical performances as part of Buddhist festivals were also exploited for commercial reasons.30 Festivals were used as a tool to strengthen the ties between the lay and monastic communities through merriment and economic exchanges. Their success was determined by their ability to attract crowds which are usually described in the range of ‘several hundred thousand’. Merchants used the opportunity provided by the festivals to sell their goods to the gathering. Some of the goods sold in the festival markets and the resulting profit, may very well have been donated to the saṃgha. In these cases, texts emphasize the religious and commercial opportunities created by festivals that could be exploited by the saṃgha to enrich itself. Viewing our images in the light of these descriptions of festivals, one cannot help but suspect that the reliefs depicting dance recalled the substantial religious and commercial interest of the broader community that viewed them. Dance imagery in this context could have acted as a concise visual reminder of the opportunities for merry-making and merit-making created by the festivals.

3. Visual Convention of Dance at Kizil Cave 83

Even though we cannot definitively connect the dance movements depicted in Gandhāran art with contemporary praxis, other related evidence shows that the iconographic and stylistic conventions related to dance resonated within a wide geographical and cultural milieu. One such evidence comes in the form of a wall painting depicting a dance performance and found in Kizil Cave 83, the Treasure Cave or Schatzhöhle C. The square cave with a doomed ceiling is part of a compact group of grottos comprising Caves 82, 83, 84 and 85. The caves derive their name from a treasure found in a deep pit.31 Caves 83 and 84 possess paintings of the Sonderstil or Special Style due to the delicate contours, expression and slit eyes of the figures.32 This style closely follows Indic painting traditions and can be dated to the early development of the site, around the fourth century CE. The paintings also provide a consistent overlap during the period in which Buddhist activities in Gandhāra continued to develop. Keeping in mind the active role of Gandhāra and Kuča within the early trade routes connecting Central Asia and the Indian Subcontinent, it is not surprising to see the same dance depicted at Kizil in Cave 83.33

The domed ceiling of Cave 83 was destroyed, and the paintings on the side walls were poorly preserved by the time of Albert Grünwedel’s expeditions in 1906. The most well-preserved painting was on the back wall, and it depicts a dance (Figure 13). The painting, commonly interpreted as the legend of King Rudrāyaṇa, is dominated by an enthroned royal figure who watches a female dancer. The story is present in the Divyāvadāna or Divine Stories, a compendium of Indian Buddhist stories in Sanskrit that were compiled as early as the first century until the eighth century CE.34 The popularity and influential nature of individual āvadānas amongst Buddhist communities in various parts of the subcontinent is also attested by their representations in art. Moreover, the āvadānas were also included in the Mūlasarvāstivādavinaya, a Buddhist school that flourished in the Indic northwest during the first centuries CE.

For the present study, a general synopsis of the Rudrāyaṇa-avadāna is sufficient. This avadāna consists of connected sub-stories that recount several events related to King Rudrāyaṇa ruling the great city of Roruka with his favorite queen Candraprabhā, his son Śikhaṇḍin and his chief ministers. It starts with the friendship between Rudrāyaṇa and Bimbisara in which the latter sends the former an image of the Buddha. These events are followed by Rudrāyaṇa allowing his harem women to listen to dharma talks. Every day, the dharma talks were delivered for an unknown period. One day, when the King expertly played the veena while the Queen danced, he saw a sign that foretold the imminent death of Candraprabhā. Upon hearing the signs, Candraprabhā became a bhikṣuṇī and was born in the Cāturmahārājika (Four Great Kings) heaven. After some important religious breakthroughs, she showed her divine self to the King and he decided to renounce everything including his kingdom to his son. The avadāna continues with several events related to their son Śikhaṇḍin and ends with a moral lesson against evil.

For the identification of the painting in Cave 83, one of the most important moments of the story with the Queen Candraprabhā dancing in front of the King is persuasive. Under an elaborate architectural frame often related to decorative motifs in Roman art, King Rudrāyaṇa watches as Candraprabhā dances. Several courtiers and well-dressed women surrounding the royal couple also watch the performance. The painting likely shows us the moment in which the King saw signs foretelling the Queen’s imminent death. The bodily posture of Candraprabhā is depicted in a type of tribhaṅga that we have already come across in Gandhāra art. This posture lends the body a certain flexibility and emphasizes her movement. Her left leg is bent at the knee and is lifted as it is about to be placed behind her right leg. Her left hand is also raised above her head. Her body is curved toward the right and her head is bent down. Within her palms, she holds a long garland that is undulated as she performs her dance. Her lower body is covered by a rich, transparent, diaphanous garment that further amplifies her dance. The texture of the garment also provides a glimpse of her bejeweled girdle clasped around her hip. Besides the girdle, she wears many ornaments including bracelets, earrings, anklets, and necklaces, all of them fitting for a dancing Queen in the palace.

The dance in the Kizil painting finds close parallels with dancing female figures in Gandhāran art, for example, with Figure 14. On this relief, we are confronted with a similar movement with little variation. A female figure, who is partially preserved, is depicted with an arm extended along the length of the body and another arched above her head. Despite the damage to the surface, we can trace that the figure likely held a long garland between her hands. Moreover, the lower limb movement is not fully preserved. However, based on the painting and other representations of dance in Gandhāran art, we can suggest that the dancer’s leg movements were derived from conventional imagery and so, her feet were likely crossed. She performs this dance under a pavilion supported by a Gandhāran-Corinthian column and for an audience. Besides the iconographic conventions of dance, the architectural context and the elaborately dressed figures surrounding the dancer reveal that we are looking at a performance at the court, perhaps within the palace of Siddhārtha.35

In Gandhāra and Kizil, we find the Indic dance executed within the luxurious courtly setting. Whether it was Siddhārtha’s splendid palace or King Rudrāyaṇa’s court, both the stone reliefs and the mural painting present strikingly similar bodily posture in association with this dance. Based on their similarity, we can suggest that artists in Kizil relied on a popular iconographic convention to represent dance. They may have looked for an existing iconographic model that was already established in Gandharan art within the courtly context to represent the dance of Queen Candraprabhā. Thus, the iconographic conventions typically associated with courtly dance performances in both Gandhāra and Kuča suggest, in my opinion, that some elements of the visual language related to dance were likely shared by artists over a large geographic and chronological zone. They reveal that dance imagery conformed to artistic conventions and the dancing figures were symbols for dance in general rather than an illustration of a particular dance performance. Based on the striking uniformity between Cave 83 and Gandhāran reliefs, we can identify the presence of similar conventions, and add dance imagery to the growing list of similarities between the two regions.36

4. Preliminary Conclusions

Our knowledge of dance in Gandhāra relies entirely on the iconographic record. Based on the images alone, I would not dare to conclude that dances in art were observable in reality since we lack any other type of evidence to support such a claim. However, the analysis of a specific type of dance, here associated with Indic conventions, allows us to arrive at some preliminary conclusions. The repetition of similar dance movements in visual art over four centuries and within different regions in Gandhāra and elsewhere, in my opinion, suggests that artists used specific visual conventions to represent dance. This suggestion is further strengthened by the fact that the artists focused on a limited combination of arm and leg movements to convey dancing. These movements, I suspect, became topical and began to be associated with dance as a general category of performance in visual art. The representation of such movements were likely visual cues used to evoke the world of dance performances in general, rather than a particular event, in the mind of the viewers. The devotee, as an audience familiar with the visual language and perhaps influenced by contemporary experiences, would have understood these ‘petrified’ dances despite the limited recombination of movements.

The dance imagery commonly used in Gandhāran art may have also influenced the artists at Kizil, as suggested by the mural painting in Cave 83. The painting depicting Candraprabhā’s dance reveals that the artistic conventions associated with dance in the Indian subcontinent may have also been utilised in the Tarim Basin. The use of the Indic dance at Kizil sheds light on the constructed nature of dance imagery in both sculptures and mural paintings in the early centuries CE. Their striking similarity may have stemmed from the fact that dance postures in art were part of prescribed elements, lending dance imagery a level of uniformity within a wide geographical area. The artists at Kizil may have adopted the pre-existing elements associated with dance and deployed them to enhance the courtly setting of the narrative. In the case of the painting in Cave 83, the choice made by the artists was likely facilitated by recombining a limited set of arm and leg movements that were already consistently associated with Siddhārtha’s palace in Gandhāran art.