The territory of modern Kaśmīr was the homeland of famous Buddhist monks who had very important roles in the buddhization of the Himalayan region and Chinese Turkestan. The Kashimro-Kuchean monk Kumarajiva (344-413) was so famous that the Chinese Emperor Yao Xing (393-415) wanted him at Chang’an, while Padmasambhava was expressly invited sometime between 770-780 by Khri srong lde btsan (756-797) from Uḍḍiyāna (a region bordering Kaśmīr) in order to win over the last resistance to the adoption of Buddhism in Tibet. Later, Rin chen bzang po (958-1055) studied in India and Kaśmīr before the foundation of his famous Buddhist school in Tibet. However, the relations between Kaśmīr and Tibet were not always friendly especially during the rule of the Kārkoṭa (c. 622-855) and Utpala (c. 855-1003) dynasties.
1.1. Kaśmīr under the Kārkoṭas and Utpalas
During the reign of Muktāpīḍa Lalitāditya (c.725-760/61)1, Kaśmīr experienced an exceptional period of cultural and artistic flourishing. Despite the Kārkoṭa adhesion to Hinduism, Buddhists did not suffer much under Lalitāditya who apparently even protected the followers of the Dharma. The monuments dated to his reign, which are still visible, reveal clear signs of contact between Kaśmīr and the Classical and the Iranian world possibly due to the actual presence of architects and artists originally from these distant regions (Goetz, 1969.b) although it is not possible to exclude that such Hellenistic elements were just borrowings from the art of Gandhāra (Siudmak, 2007: 45).
The extension of Lalitāditya’s domains into parts of North-Western India and Bactria-Tokharistān and the submission of the Turki Śahī (c. 665-843, but in Zabul until 870) as it was presented by Hermann Goetz several years ago have been reconsidered in the light of recent re-examination of Chinese written sources and local coinage. The background of the “golden age” of Kaśmīr presents difficulties especially for the relations with neighboring kingdoms: Bactria-Tokharistān was under Arab rule, while Kapiśa and Zabul were governed by Turkish dynasties (the Turki Śahī) who were friendly with the Kārkoṭas (Sen, 2004: 152-154).
Very soon Lalitāditya entered into conflict with the Tibetans who – during the period of the sPu rgyal (or Yarlung) Dynasty (649-850 c.) – wanted to expand their domains into the Southern Hindukush. For this reason, he was considered a good ally of the Tang Empire (618-906) (Wink, 1990: 243-244; Sen, 2004). Chinese chronicles and other literary sources of the Heavenly Empire recently collected in a very interesting study by Tansen Sen, give quite a clear image of the geopolitical situation in the Southern Hindukush region around 700 (Sen, 2004). The position of the Arabs in Bactria-Tokharistān and the Turki Śahī in the area between modern Eastern Afghanistan and North-Western Pakistan has already been considered above. Kaśmīr (Gushimi or Jiashimiluo according to Chinese sources) formed an alliance with the Tang in order to contain the advance of the Tibetans into India. The Chinese army was headed by Gao Xianzhi, a general of Korean origins who, later, was defeated in 751 at the famous battle of Talas (South-Eastern Kazakhstan) by a Turco-Arab coalition.
It does not seem that Muktāpīḍa (Muduobi in Chinese sources) fought alone as in Goetz’s reconstruction although the expansionist intentions of the Tibetans should have certainly alarmed him. During the war against Khri lde gtsug brtsan (commonly known as Mes ag tshoms, 712-755) and Khri srong lde brtsan, the figure of Lalitāditya assumed legendary proportions and for this reason it should be considered more cautiously (Wink, 1990: 243-254; Sen, 2004).
The main direct literary source about the life of Muktāpīḍa Lalitāditya and other Kashmiri sovereigns is the Rājataraṅgiṇī. This work was composed by Kalhaṇa in 12th century and it represents the first example of a “chronicle” ever written in India. Interesting information can be obtained in the same source such as regarding the history of Kaśmīr in relation to Iranian-culture. In the Rājataraṅgiṇī, the Kuṣāṇ annexation of Kaśmīr under Huviṣka and Kaniṣka (2nd century) is mentioned together with the coming of the Kidarites and, possibly, the Hephtalites in the 4th-6th centuries (Rosenfield, 1967: 49-50; Dani, 1996: 167-172)2. Both these peoples, in fact, adopted the culture of Bactriana but their ethnic affiliation still remains a mystery. However, there are no Kashmiri monuments to be safely attributed to the Kuṣāṇas or other Iranian-culture peoples although, as we shall consider below, some doubts arise about the Harwan complex.
In a passage of the Rājataraṅgiṇī (in book VI, 192), we read about an interesting association between the Sun (Sūrya) Temple at Mārtand and a city not far away whose main activity was the cultivation of the grape (Stein, 1906: 141). As P. Pal argued, the city was possibly populated by a colony of Persians who had escaped during the Arab invasion (Pal, 1975: 42). His hypothesis is based both on the presence of grapes such as on the Indian cult dedicated to the sun by the Maga Brāhmaṇa (Wink, 1990: 252; Panaino, 1996), a sect settled in this part of India long before the fall of the Sasanians (224-642) who could have given refuge to the exiled Persians.
However, as already observed by this author (Compareti, 2000: 338), an identification of the people of the grape-city with the Sogdians seems more likely. As in Chinese sources, Sogdians are described as people fond of wine (Chavannes, 1903: 134; Trombert, 2005). According to one Chinese literary source, then, around the city of Shanshan (Xinjiang Province) there was a settlement called the City of the Grape (Putao Cheng) because of the plantations of grapevines there by Sogdians (Giles, 1930-1932: 829-830; de La Vaissière, Trombert, 2004: 950). Moreover, Mithra was a god venerated in Sogdiana itself as well (Grenet, 2001) and, according to the Muslim author al-Idrīsī, some parts of Kaśmīr were “inhabited by people and merchants from all parts of the world” (Wink, 1990: 247). Since al-Idrīsī was writing in 12th century, it is hard to imagine that he was not referring to Sogdians too. In any case, it is not possible to state for certain if the construction of the Sūrya Temple at Mārtand had some connection with the Iranians living in Kaśmīr since this god had many followers in medieval India as well. Moreover, the grape could have been introduced into Kaśmīr much earlier, even during the Indo-Greek or Kuṣāṇa period (Pal, 2007: 23).
1.2. Iranian elements in Kaśmīr
Some specimens of Kashmiri art display very clear Iranian borrowings. The much discussed terracotta tiles from the Buddhist complex of Harwan (not far from Śrinagar) have always attracted the attention of students of Iranian art for decorative elements such as pearl roundels containing single flowers (fig. 1) or birds (fig. 2), and archers hunting animals according to the so-called “flying gallop” style while ribbons attached to the body are floating in the air (fig. 3). These elements, in fact, call to mind a typical Sasanian decoration which, however, obliges us to revise the chronology proposed for the site (Paul, 1986: 53-62).
Fig. 1. After: Kak, 1933: pl. XL fig. 43.
A date as late as possible would fit better for Harwan, as, in the past, a chronology was proposed that was obviously too early, thus rendering impossible the presence of the (Iranian) pearl roundels pattern (Kak, 1933: pls. XX, XXII.1, XXII.3, XXX.22, XXXIV.31, XL.42-43; Fisher, 1987.a). In the most recent study dedicated to Kashmiri art, a date to 5th century is proposed (Paul, 1986: 44), although even the 6th century could also be considered likely: this was the period of invasions from the north-west which have been attributed to the Huṇas by Indian sources.
Fig. 2. After: Kak, 1933: pl. XXX fig. 22.
Of the Huṇas much needs to be discovered (Parlato, 1990; De La Vaissière, 2003) but it is not possible to deny that the invaders of North-Western India could have brought Iranian elements although their affiliation was to a different ethnic group (for example, even Altaic). In any case, it is worth remembering that pearl roundels containing single lotus flowers appeared around 1st century in Indian art at Bharut (?) and Sañci (Bénisti, 1952). So, a pure Indian component should be taken into consideration when studying the Harwan tiles.
Fig. 3. After: Kak, 1933: pl. XXXIII fig. 3 (detail).
On the other hand, the pearl roundels reproduced on the pillows on which two bronze statuettes of Buddha are sitting in maradhāsanamudrā and dharmacakramudrā can not be considered proper Sasanian despite the hypotheses expressed by some scholars who consider such an element indisputably Sasanian (Tucci, 1974: 300). The statuettes are kept respectively in the Norton Simon Museum (fig. 4) and the Lahore Museum (fig. 5) and they can be dated quite precisely to 8th century Kaśmīr (Pal, 1975: cat. 22.a-b; Catalogue Naples, 1964: cat. 329).
Fig. 4. After: Pal, 1975: cat. 22.a.
The vegetal elements represented inside the pearl roundels have a very precise parallel in the figurative textiles found around Turfan (Gao, 1986: 129, fig. 91 and 161, fig. 76). Although it is not clear if these textiles were actually produced in the area of Turfan or in the region of Shu (modern Sichuan Province), the presence of Sogdians at both Chinese sites represents an historical fact (Compareti, 2006). A third 8th century metal statuette recently showed on the occasion of an exhibition in Germany could be included in this group: it is possible to observe a Buddha in dharmacakramudrā sitting on a pillow on a high pedestal wonderfully embellished (Catalogue Berlin, 2006: cat. 11). It was kept in the Potala in Lhasa and its state of preservation is excellent: traces of color can be clearly observed on the whole statue especially on the face and hair of Buddha and on the pillow embellished with pearl roundels of the same kind of the two other metal statuette in the Norton Simon Museum and the Lahore Museum.
Fig. 5. After: Catalogue Naples, 1964: cat. 329.
One last 8th century Kashmiri bronze with silver and copper inlay, now part of a private collection, presents interesting decorations too (Heller, 2006: figs. 125-129). The sitting Buddha in dharmacakramudrā is curiously dressed with precious clothes, a pointed crown and ribbons while at his sides stand two bodhisattvas and, below, three smaller donors. Pearl roundels embellish the frontal side of the pillow on which the Buddha is sitting and the garments of one donor, although the latter is only partially visible (fig. 6). A very important detail of this statue is the three-pointed camail on Buddha’s shoulders. In fact, usually, just bodhisattvas and minor Buddhist deities have such precious garments since they are not supposed to have done any renunciation. This kind of small cloak reflects the fashion of the people living in the North-Western Indian regions and in the territory of modern Afghanistan and, sometimes, it can be observed worn by Buddha himself according to the typology called “Buddha parée” (Compareti, 2007).
Fig. 6. After: Heller, 2006: fig. 126 (detail).
A typical Iranian dress, the camail, can be observed often in Kashmiri statues of Sūrya as, for example, at Martand, in the Lahore Museum (fig. 7), in the Srinagar Museum and in several bronze statues kept in museums and private collections (Goetz, 1969: pl. XXI; Fisher, 1987.b: fig. 7; Siudmak, 1987: 51; Harle, 1987; Pal, 1992: figs. 3, 6-7). The camail is a three-pointed poncho-like cloak worn over the other clothes. Figures of donors wearing the camail can be observed in Gandhāran statues but also, later, in Sogdian paintings from Panjakand (Marshak, 2002: fig. 60). According to J. Siudmak, the camail would have been introduced into Kaśmīr during the Kārkoṭa Dynasty, possibly, by the Buddhist communities protected by Lalitāditya (Siudmak, 1987: 51). One of Lalitāditya’s ministers, in fact, was a Tokharian and a patron of Buddhist works too. His name was Cankuna, possibly a corruption of the Chinese title jiangjun (general), and, most likely, he was a follower of the Dharma (Goetz, 1969.a: 11-12). This figure as well has been critically reconsidered by Tansen Sen and, if his suggestions (as it seems likely) are to be considered correct, then Cankuna should be considered to have come from Bactria-Tokharistān and not from the region of the Tarim Basin as supposed by H. Goetz (Sen, 2004: 151-152). Since in all the territories just mentioned the pearl roundels pattern was very well-known and appreciated, it was proposed to attribute to Cankuna its introduction into Kaśmīr (von Schroeder, 1981: 107). The hypothesis seems to be likely but it is not clear if Cankuna himself adopted particular decorations directly from Sasanian or Sogdian traditions. The second hypothesis would seem more convincing since, in the 8th century, the Sasanians did not exist anymore.
Fig. 7. After: Pal, 1992: fig. 6.
Architectonic decoration under the Kārkoṭas show Iranian elements as well. The Śiva stone temple at Pandrethan has the ceiling embellished by large pearl roundels containing lotus flowers of the same kind of the Harwan ones (Brown, 1955: 48; Fisher, 1987.b: fig. 9, 14). The same shape of the ceiling has a clear parallel in the so-called “lantern” typology which was very widespread in Central Asia such as at Varakhša and Qal‘a-e Qahqaha (but also in the Caucasus). Strangely enough, pearl roundels do not appear at Mārtand.
After the death of Lalitāditya, the Kārkoṭa Dynasty lasted one century more although the territorial boundaries and the splendor of the court were not the same. The Utpalas did not favor Buddhism as their predecessors. Such a situation mirrors a general trend of the whole of India with the exception of the Bengala region. In the Utpala tributary territory of the Hindū Śahī (c. 843-1026)3, between modern South-Western Afghanistan and Northern Pakistan, Hinduism was the main religion too, as the same dynastic name of those sovereigns points out. Hindu sovereigns were not always tolerant with other religions spread in India which, during some periods, suffered persecutions (Verardi, 2002; Verardi, 2003; Verardi, Paparatti, 2004: 97-102). Among the few protectors of Buddhism in Central Asia there was Tang China that was present in the area until 750. This is considered to be the main reason for the sinicized features of the face of Buddhist statues on a great area extended from the borders with China to Margiana (Rhie, 1988; Verardi, Visconti, forthcoming). Kaśmīr does not seem to constitute an exception and, in fact, the famous Kashmiri Buddhist bronzes display very strong mongolic traits such as almond eyes (Goetz, 1955: 72; Heller, 2008: 30).
The Rājataraṅgiṇī is not too explicit about the material presence of Iranians in the territory of Kaśmīr although the decoration of some temples in the area of Śrinagar shows evident signs of Iranian borrowings. The temple of Avantisvāmi (or Avantiśvāra) in Vantipur has columns embellished with vertical pearl roundels containing animal, vegetal and geometric elements. The repertoire at Avantisvāmi is very rich: here, in fact, all the typology of pearl roundels just described can be observed (fig. 8). The roundels at Avantisvāmi also present square elements in the points where they should have been tangent to another roundel. This is a typical solution observed in Sasanian stuccoes and, occasionally, textiles, while in Sogdian textile art it is much more widespread (Compareti, 2004.a). Certainly, many Persians escaped from the Arabs who invaded the Sasanian Empire and it is also probable that they settled in some parts of India and Central Asia4. As it is well-known, typical Sasanian motifs were accepted by the Omayyads (661-750) and the Abbasids (750-1268) but it is very strange to observe them in Kaśmīr during the Utpala period, that is to say, approximately two centuries after the end of the Sasanians. Moreover, the decorations of the Avantisvāmi temple comprehend images of facing animals also which were not appreciated at the Sasanian court (Compareti, 2000: 338-339)5. On the contrary, facing animals represent the main subject inside the pearl roundels of the textiles improperly called zandaniji, which were woven in a non-determined region of Central Asia after the Arab conquest (Marshak, 2006.a; Raspopova, 2006). So, it seems more probable that the Iranian motifs adopted in the decoration of the Utpala temples were coming from Central Asia and not from Sasanian Persia.
Fig. 8. After: Kak, 1933: LXXXII.
At the time of the Kārkoṭas and the Utpalas and even later, at least until 12th-13th centuries, Kashmiri art deeply influenced the entire Himalayan region (Pal, 1987; Siudmak, 2000; Henss, 2002; Béguin, 2002: 246-247; Heller, 2008: 28-30). Most likely, Iranian elements observed in the artistic production of Tibet arrived there through Kashmiri artists. Not only central Tibet but also Ladakh and Guge – that is to say, the western outskirts of the Tibetan Empire – accepted many Iranian elements especially in the decoration of the garments of Bodhisattvas and, quite strangely, Buddha too. The main Buddhist centers of Ladakh are represented by the monasteries of Mangyu and Alchi, dated to 12th-13th century (Pal, 1988; Linrothe, 1994; Goepper, 1996). The latter site is particularly interesting for the presence of pearl roundels reproduced on the ceiling of the Sumtsek (gSum-brtsegs), a three-storied temple built in early 13th century (fig. 9). Many patterns of pearl roundels can be observed and scholars agree in recognizing not only Iranian motifs but also painted imitations of textiles (Flood, 1991; Goepper, 1993; Goepper, 1996: 225-264). The same decorations which embellish the garments of important people in the paintings at Alchi display clear Iranian borrowings: and also for them a Sogdian origin can be argued (Singh, 1991: 517; Sims, 2002: 23-24; Pal, 2007: 147-149). Very recently, new Tibetan paintings have been discovered in the region of the ancient kingdom of Guge which is nowadays under Chinese administration too. According to the preliminary investigations, there are clear Kashmiri borrowings in those paintings which could be dated between 11th-15th/16th century (The Institute of Chinese Tibetan Learning of Sichuan University, 2007: figs. 10-13, 21-24; Pritzker, 2008).
Fig. 9. After: Flood, 1991: fig. 3.
Something more could be added about the paintings on the dhotī of the bodhisattva Avalokiteśvara on the ground floor of the Sumtsek at Alchi. Approximately in its center, slightly above the knees of the statue, two scenes attracted the attention of scholars for two reasons: they are the only ones not to be interpreted as religious scenes and “clearly reflect stylistic elements that are different from the earlier Buddhist paintings of the eleventh-century monuments. They evoke a cultural milieu that was both sophisticated and cosmopolitan” (Pal, 2007: 147). On the right leg a royal couple sits in a palace while musicians and attendants can be observed outside on both sides. On the left leg there is the representation of a parade of horsemen and the king appears too holding clear royal symbols like the axe (fig. 10). Not only some of the textile decorations could be associated to the Iranian world but the royal scene itself calls to mind similar Islamic book illustrations which could be contemporary with the paintings in Alchi.
Fig. 10. After: Pal, 2007: fig. 157.a-b.
Many specimens of Islamic book illustration could be mentioned in order to find a parallel with the Alchi paintings under exam but there is a frontispiece in the Topkapı Saray (Istanbul) known as Hazine 2125 which resembles surviving parts of an important 7th century cycle of Sogdian paintings at Afrāsyāb (ancient Samarkand) whose origins are definitely rooted in pre-Islamic traditions of Central Asian art (Esin, 1977; Pugačenkova, 1987). As I attempted to prove in another paper, at least at Afrāsyāb there is the representation of the local Nawrūz (the most important Iranian festival) while in the Topkapı and in many other Islamic book illustrations it is not possible to be that specific: what is really important to remark is the common origins of such a stereotyped scene despite its meaning which could be different according to the cultural milieu where it is found (Compareti, forthcoming 2009). Musicians appear always in this group of images together with dancers and attendants offering a flower (or a plant), a bow with arrows, and also horses and other animals used for hunting (dogs and falcons) to the king or the royal couple (fig. 11)6. Something similar could be guessed for Alchi itself. In fact, even if we do not know the exact meaning of that royal scene it is interesting to note that the Kashmiri artists who executed it chose an Iranian model which was spread for a very long time and accepted also by other Central Asian peoples like Turks and Mongols.
Fig. 11. After: Compareti, forthcoming 2009: figs. 4-5.
In Ladakh some Sogdian inscriptions have been discovered too. They cannot be dated precisely but, while the 4th-6th century inscriptions do not represent an unicum since there are other contemporary ones in the Hindus Valley (Sims-Williams, 1989-1992), the 9th century inscription, on the contrary, is among the latest evidence concerning the presence of Sogdians along the Himalyan trade routes (Sims-Williams, 1993; Vohra, 1994).
The main monuments in the territory of Guge (which is nowadays divided between India and China) are: the monastery of Tabo and the grottoes of Dung dkar, both dated to the 11th century (Wandl, 1999; Neumann, 2000). The paintings from Tabo are definitely the most interesting and they were rightly associated with the paintings at Bāmyān for their numerous Iranian decorative elements (Klimburg-Salter, 1996). A unique group of manuscript covers from the area of Gilgit possibly dated to 7th-8th century constitutes another parallel with paintings produced in the areas of modern Afghanistan where Iranian Buddhism was spread. Not only are the figures of Buddha and Bodhisattvas on these covers extremely similar to the models at Bāmyān, Kakrak and Fondukistān but the features of the donors represented kneeling besides them also have clearly common roots (Klimburg-Salter, 1990; Klimburg-Salter, 1991).
2.1. Tibet under the Yarlung
During the period of the Yarlung Dynasty, Tibet was one of the main powers in Central Asia and the contacts with the Sogdians are attested in several sources. The Tibetans extended their conquests on the Tarim Basin between 666-692 and, more firmly, between 760-850. They also conquered some parts of Gansu (Haarh, 1969; Beckwith, 1987; Hoffmann, 1990). Until a first attempt of establishing good relationships with the Omayyads and, later, the Abbasids (Dunlop, 1973), the btsanpo Khri lde gtsug brtsan (712-755) – one of whose wives was a princess from Samarkand (Twitchett, 1979: 432) – fought against the Arabs together with his Turkic allies in order to expel them from Sogdiana (Beckwith, 1987: 108-110). It is not completely clear what was the reason for such an intervention. Possibly the Sogdian traders represented a good source of gain for the Tibetans who controlled the access to some vital routes in Central Asia. The relationships between Sogdiana and Tibet were in general friendly (Li, 1957-58; Hoffman, 1971; Uray, 1979: 282-283; De La Vaissière, 2002: 152-153), even though it is reported that there was at least one episode of an incident when a Tibetan official was captured by some Sogdians in 694 (Li, 1957-58; Hoffman, 1971: 443-446; Beckwith, 1987: 56).
The Yarlung conquest of Khotan – an Iranian Buddhist kingdom in strict relations with Kaśmīr and Sogdiana (Bailey, 1982: 4, 9, 57: Kumamoto, 1996: 84-86; Mu, Wang, 1996) – heralded the immigration of many Khotanese monks, craftsmen and merchants into Tibet (Hofmann, 1971: 451-453; Gropp, 1974: 36-37). A Kashmiri bronze statue is said to have been found at the site of Domoko, in the Khotan Oasis (Heller, 2006: 185, n. 25; Heller, 2008: fig. 11). The word for merchant in Khotanese was sūlī (plural sūlya, sūlīya) and, although it is probably connected with an Iranian root sau– (to earn), most likely, it is also associated to the ethnonym sūlya: Sogdian. This is probably due to the fact that the Sogdians were considered the merchants par excellence in Central Asia (Bailey, 1982: 23; De La Vaissière, 2002: 64, 130).
The Sogdians had trade relationships with Tibet (Beckwith, 1965: 100-103: De La Vaissière, 2002: 303). Since the main items traded by Sogdians were luxury goods, it is possible to consider that they imported into Tibet precious silks and metalwork. The court of Lhasa, on the other hand, provided the Sogdians with the famous perfumed musk which was very much requested by the Muslim courts. According to Mas’udi (10th century), some merchants that he met in Eastern Persia arrived there “from Sogdiana through