Demons in the detail: temple architecture of the Hoysala dynasty

Fig 1. Narasimha at Chennakeshava temple, Belur

I recently visited the state of Karnataka, India, where I fulfilled a long-standing ambition to see the ancient temples of the Deccan region. But not everything translated well into print or pictures: the famous site of Hampi, for example, is a vast complex of ruined temples dotted across a rocky landscape which quite defeated my pathetic phone-camera.

Instead, it was the less famous though equally beautiful temples that provided good blog material. And this image (Fig 1)  has always been a favourite of mine, partly because it’s deliciously gory but also because of its underlying legalistic humour.

What we see is the god Vishnu disembowelling the demon-king Hiranyakashipu (those two long chains are his intestines, meticulously carved in stone). As told in the Bhagavata Purana (an 8th-9th C religious text), the demon persuaded the god Brahma to make him invincible. The terms of the gift, however, were that Hiranyakashipu could not be killed (a) by man or beast, (b) indoors or outdoors, or (c) by night or day. 

Vishnu overcame these parameters easily by assuming the form of the man-lion Narasimha (neither man nor beast ✓) and killing the demon on the threshold of his home (neither indoors nor outdoors ✓) at the point of sunset (neither night nor day ✓). Which goes to show that even a demon-king needs to check the small print.

This particularly gruesome version comes from the Chennakeshava temple at Belur, a 12th C temple dedicated to Vishnu, built by the kings of the Hoysala dynasty. The Hoysalas ruled a large part of the Deccan from the 10th to the 13th C, but are not especially famous today: they never achieved the domination of the Cholas (whose rule extended to South-east Asia) or the glory of the Mughals. 

Fig 2. Approximate area of Hoysala kingdom, 10th-13th C

Their place in history is secured, however, by their temple architecture which is ornate to the point of obsession. The level of detailing stands out even in a milieu that is not known for restraint. Italian Rococo is perhaps the only other school that comes close.

In Fig 3 for example we see the 12th C Hoysaleswara temple at Halebidu, dedicated to Shiva.  As at Belur and the other great temple at Somnathpur (which I didn’t visit), the façade is densely layered and segmented, providing multiple surfaces for the display of religious sculpture. The material used was chloritic schist, also known as soapstone (not to be confused with the soapstone in fake jade jewellery). This relatively soft but supple stone can be carved intricately without crumbling, a quality of which the Hoysala sculptors took full advantage.

Fig 3. Hoysaleswara temple, Halebidu

Each of the carved panels looks like it took a lifetime to complete. And from the architectural plans (below), it seems that each of the linga sanctuaries (the star-shaped sections on the left) was designed with as many facets as possible to provide space for ornamentation.

Source: Temples of Deccan India, George Michell

The two door-guardians at the main entrance (Fig 4) are perfect examples of the microscopic detailing typical of Hoysala sculpture.

Fig 4. Door-guardians at Halebidu

The outer walls are covered with carved panels with borders of leaves, flowers or animals. The one in Fig 5 depicts Ganesha (right) and a rare image of Brahma riding a peacock (left). Each bead in their jewellery and head-gear is carved in three-dimensional detail. Even Ganesha’s pet rat (under his feet) wears a jewelled harness and has sharp little claws on each toe. 

Fig 5. Brahma (L) and Ganesha (R) at Halebidu

Stories from the Bhagavata Purana, the Mahabharata or the Ramayana were the sculptors’ favourite subjects. The panel in Fig 6 depicts the ace archer Arjuna (from the Mahabharata) shooting an arrow into the eye of a fish-shaped target suspended high above. In the story, Arjuna uses only a reflection in a bowl of water to take his aim. The sculptor depicts Arjuna’s right arm pulled back and his hand wide open as if he has just released the arrow. His right leg is bent so that he can shoot upwards but his eyes are downcast, looking only at the reflection in the water.

Fig 6. Arjuna the archer, Belur

In Fig 7 at Belur, Vishnu outwits another demon-king, named Bali, who has taken over the universe. Vishnu, disguised as a dwarf, begs Bali to give him a little piece of land, so small that he can cover it in three steps. The demon-king agrees. Vishnu then reveals himself and takes two gigantic steps covering the earth and the universe. Bali, admitting defeat, offers his own head as the last step and Vishnu takes it, thus securing his victory. 

Fig 7. Trivikrama panels at Belur

The story is carved on two neighbouring niches: the dwarf on the left (with his little umbrella) making his request to Bali, and Vishnu on the right, taking one of his giant steps which makes him look like he's dancing the can-can. This popular image, known as Trivikrama (the three steps), appears at Vishnu temples across southern India. 

It was touching to see these stories being re-told to groups of local visitors (we were the only foreigners there), as they probably were hundreds of years ago. 

More trip-based posts to come when I finish sorting out my 300+ photos.

God or goddess? Guan-yin at the Burrell Collection

 

Fig 1  The bodhisattva Guan-yin. Approx 1100-1200 CE, Song Dynasty, China
Wood with paint and gilt. Burrell Collection, Glasgow

A year after my Berlin fiasco with British Airways, I spent my measly refund (thanks, BA!) on a flight to Glasgow to see the Burrell Collection, a huge stash of art amassed by Scottish millionaire William Burrell in the early 20^th century.

After a major refurbishment, the museum re-opened in 2022. The renovated spaces are lovely (and warm – important in Scotland!) but it was the new display labels that set off a Twitter-storm. In fact, that was the main reason for my visit because I couldn’t believe what I had heard.

Fig 2  The renovated Burrell building in typical Scottish weather

Attention centred on the labelling of two figurines of the Chinese deity Guan-yin. This deity, widely worshipped in East Asia as a goddess, evolved out of the Indian male bodhisattva[1] Avalokitesvara after Buddhism spread from India into China. This gender evolution is one of the great mysteries of East Asian Buddhism and worthy of close study. But the Burrell went with:

Figs 3-5  Figure of Guan-yin and accompanying labels. Burrell Collection, Glasgow

This was the most notable, but not the only, example of “new” labelling. Elsewhere, a Tang Dynasty dancer figurine was labelled: “Do you like to dance? This girl does. She’s moving her body and clothes to show off her dancing skills.” And other examples in a similar vein.

All British museums are under pressure to be more “inclusive” if they want to retain public funding, so the Burrell’s labels may be driven by this. Or it may be an attempt to go viral to boost visitor numbers. The museum has defended its approach (see this link) and was even named as the UK Art Fund’s 2023 Museum of the Year. It pushed me to research Guan-yin’s evolution in greater detail, which turned out to be quite complicated (surprise!) so this post is a bit long.

The name “Guan-shi-yin”, later shortened to “Guan-yin”, first appeared in Chinese translations of Buddhist scriptures in the 3^rd century CE. In one of the main scriptures, the Lotus Sutra, the Buddha praises the compassionate nature of a bodhisattva named Avalokitesvara. When this was translated from Sanskrit, the name Avalokitesvara (“lord who sees from high”) became “Guan-shi-yin” (“one who perceives/senses the entire world”), reflecting his mission to hear and assist all persons in need.

The original sutra does not specify Avalokitesvara’s gender. The “svara” component of his name typically refers to a male aristocrat or lord. But the sutra lists 33 different forms which he has assumed in order to carry out his tasks, of which 7 are female, so it seems a female persona was not ruled out.

Indian religious art invariably depicts Avalokitesvara as male, as in Fig 6 below where he sports a luxuriant moustache, a bare chest and an impressive set of abs.

Fig 6   Figure of Avalokitesvara. Approx 200 CE. Grey schist stone. Gandhara, Pakistan. Source: Los Angeles Country Museum of Art

But in China and amongst the Chinese diaspora, the same deity is worshipped as a goddess and is typically depicted as below. How did the Indian version evolve into this?

Fig 7   Guan-yin. 18th C, Qing Dynasty, China. Dehua porcelain. Source: National Gallery of Australia

Early Buddhist imagery in China focused on the Buddha himself. This changed in the 5^th- 6^th centuries when the religion prospered under a series of devoutly Buddhist rulers. Guan-yin began to appear in sculpture as a slender figure dressed in elaborate robes and jewels (Figs 8 and 9). The clothes and jewellery probably came from Indian religious imagery, as well as from the Lotus Sutra which says that Avalokitesvara once received a magnificent necklace as a reward. But sculptors seemed to prefer a slimmer physique, unlike the Indian model. A modern viewer might already perceive these figures as female, though one cannot assume the same of 6^th C viewers.

Fig 8 (L) Guan-yin. Northern Qi Dynasty (550-577). Sandstone. Source: British Museum
Fig 9 (R)  Guan-yin. Northern Zhou or early Sui Dynasty (approx 580 CE). Limestone. Source: Museum of Fine Art, Boston

By the 8^th-9^th C, Guan-yin’s role as a saviour deity had won him a major following in China, especially at the Mogao cave temples in Dunhuang. Devotees commissioned paintings of him in various sitting and standing poses: someone even requested a double portrait (Fig 10 below).

Depictions from this period give him a softer, more fluid appearance: apparently, the 8^th C painter Han Gan sometimes used local courtesans as models for his bodhisattvas.  Nonetheless, the deity often had a pencil moustache and sometimes a tiny beard (as in Fig 10) as a reminder of his gender. In some paintings, this looks odd against the languid posture, full lips, plush robes and jewellery. Were artists already pushing the gender boundaries? 

Fig 10  Two Guan-yins painted on silk banner. 8-9th C. Mogao caves, Dunhuang. Source: International Dunhuang Project, British Museum 

In the 10^th century, the gender issue became more apparent with two new depictions of Guan-yin. One was the Water-Moon Guan-yin, showing him seated in a relaxed pose by a bamboo-fringed pond against a full moon. He still has masculine features: in the earliest known version from 943 CE (Fig 11), he has a vestigial moustache and beard, while in Fig 12 he is clean-shaven but has broad shoulders and muscular arms.

When this image was popularised in sculpture a century later, the deity still looked androgynous – see for example the Burrell’s own 12^th C piece (Fig 1). A variation known as Guan-yin of the Southern Seas often showed the deity looking positively Amazonian (Fig 13).

Fig 11  Water-Moon Guan-yin painted on silk. 943 CE. Mogao Caves, Dunhuang. Source: Musee Guimet, Paris

Fig 12  Water-Moon Guan-yin. 925-975 CE. Painting on paper. Mogao Caves, Dunhuang. Source: British Museum

Fig 13  Guan-yin of the Southern Seas. 11-12th C. Wood with paint and gilt. H: 2.4m. Source: Nelson-Atkins Museum of Art, Kansas City.

The other new image to emerge was the White-robed Guan-yin, which historian Yu Chun-Fang suggests was the first overtly feminine depiction. An early example is at the Yanxia cave temple near Hangzhou (Fig 14) from the 10^th or 11^th C. This sculpture has much in common with modern Guan-yin images: the flowing hooded robe, the soft posture, demure downward-looking face and the prayer-beads held in one hand.

Fig 14  Guan-yin in a hooded robe. Mid-10th C. Limestone. 1.85m. Yanxia Caves, Hangzhou. Source: Broeskamp, 2014

But then the evidential trail starts to run cold. The decline of the Mogao cave temples meant fewer religious paintings to show us developments in the depiction of Guan-yin. With the rise of Confucianism under the new Song Dynasty, painters preferred subjects like landscapes and nature. Images of Guan-yin from that period are more often found in sculpture, where the Water-Moon or Southern Seas versions were popular.

Artists never lost interest in the White-robed Guan-yin though: in this rare 13^th C painting (Fig 15), the artist has fused two forms: he has painted the bamboo and full moon from the Water-Moon but the deity is White-robed with a feminine appearance. 

Fig 15  White-robed Guan-yin. 13th C, Southern Song Dynasty. Ink on paper. Source: Cleveland Museum of Art, Ohio 

And in this early 14^th C painting (Fig 16), the full moon has become a halo but the deity wears a billowing hooded white robe. I think these images show how the foundations of modern Chinese Guan-yin imagery were laid.

Fig 16  White-robed Guan-yin. Approx 1300 CE, China. Ink on paper. Signed by Jueji Yongzhong. Source: Cleveland Museum of Art, Ohio

By the 13-14^th centuries, images of Guan-yin with masculine features had become rare in China. Instead, worship centred on the belief that the deity could help women to bear children. This “child-giving” ability was highlighted in a locally-penned Chinese sutra which (in modern terms) went viral. Unsurprisingly, when artists wanted a template to portray a Child-giving Guan-yin, they chose the feminine persona of the White-robed Guan-yin. Yu Chun-Fang argues that this was the deciding factor in Guan-yin’s gender transformation in China, which I think is plausible.

To emphasise her raison d'etre, the Child-giving Guan-yin was portrayed carrying a child, looking unmistakeably female. This appeared as early as the late Song/Yuan period as in Fig 17 below, and remained popular for centuries thereafter.

Fig 17 (L)  Stele of Guan-yin holding a child. 11th-14th C. Grey stone. Source: Christie's
Fig 18 (R)  Guan-yin with infant. Late 17th C. Dehua porcelain. Source: Royal Collections Trust, UK

The feminine Guan-yin received further impetus when Christian missionaries arrived in China in the 16^th century. Some of them (it is alleged) used the visual similarity between the Child-giving Guan-yin and the Virgin Mary to attract converts. The Virgin and Child certainly became one of the most popular Christian images in the region. A major producer of these was the Dehua kilns in Fujian who were famous for their white porcelain Guan-yins; only a small tweak in iconography was needed for their Christian clientele.

As an interesting side-note, the Guan-yin/Virgin link gained significance in Japan in the 17^th-18^th C when Christianity was banned. Christians could, however, acquire a “Guan-yin” figure (known in Japan as Kannon) and secretly worship it as the Virgin Mary. This gave rise to the Maria Kannon (Fig 19), often made at Dehua in China and smuggled into Japan. The similarities with Fig 18 above are obvious, though Maria Kannons usually had a concealed Christian icon (e.g a tiny crucifix). 

Fig 19  Maria Kannon. 17th C, Ming or Qing Dynasty. Dehua porcelain.Source: Tokyo National Museum

In both Japan and Korea, where the child-giving aspect was not so wide-spread, images of Guan-yin with male features persisted well into the 19^th C, as in the two examples below where he has a moustache and beard. A 17^th C moustachioed figure of Gwaneum (as he is known) is listed as a national treasure of Korea and worshipped at the Beopjusa Temple to this day. In China and the major centres of Chinese migration (Singapore, Taiwan and Hong Kong) however, it would be nigh impossible to find a non-female Guan-yin now.  

Fig 20 (L)  Bodhisattva Kannon on a dragon, by K Kyosai. 19th C. Japan. Source: Tokyo National Museum

Fig 21 (R)  White-robed Avalokitesvara (Gwaneum). 19th C. Korea. Source: Philadelphia Museum of Art

This post is a ruthlessly summarised account of one issue (gender transformation) within a complicated topic where a lot of key historical facts are missing. To sum up: Avalokitesvara as Guan-yin was portrayed in China as male until at least the 10^th C, when depictions began to diverge. The popularity of the Child-giving Guan-yin weighed in favour of a female image, and the Guan-yin/Virgin connection supported this preference. I haven’t had space to discuss other avatars of Guan-yin, like the Thousand-armed Guan-yin or the intriguingly named Fish-basket Guan-yin.

Researching this subject, I’ve been struck by how fluid religious beliefs can be over time. In this troubled era, Guan-yin’s role as a non-judgemental saviour of humanity rather suggests that his/her gender is irrelevant. So although I'm still dubious about the Burrell’s labels, the Burrell Guan-yin’s advice (after many years in Glasgow) would probably be to get over it and be grateful that such an impressive collection is accessible for free. Definitely worth a visit if you are in Scotland.

References:

Guanyin: the Chinese transformation of Avalokitesvara (2001), Yu Chun-Fang

"The Construction and dissemination of a new visual idiom - the White-robed Guanyin and the Upper Tianzhu monastery in Hangzhou" (2014). Bernadette Broeskamp

Chinese sculpture - a great tradition (2007), Ann Paludan

[1] In Buddhism: an enlightened being who chooses to serve the mortal world instead of ascending to Nirvana. Often depicted in art as attendants of the Buddha.

Imperial clocks from the Palace Museum, Beijing

Fig 1. Clock in the form of a pagoda with extending mechanism. Height 126cm, to 158cm fully extended.
Gilded copper with coloured stones. 18^th C. Made in England (maker’s name unknown)

This summer, London’s Science Museum plays host to 23 magnificent clocks from the Palace Museum, Beijing. Among the precious objects collected by emperors of the Qing dynasty (1644-1911), clocks were a great favourite, especially those with automated features known as zimingzhong 自明钟 (“self-chiming clocks”). The Palace Museum holds hundreds of them.  

But as you might expect of the Qing (not known for their restrained taste), these are not any old clocks. Time-keeping seems to be a secondary function: their main purpose is to knock your eye out with dazzling adornment and mechanical trickery.

Fig 2. Clock in the form of a crane with pavillion. Gilded copper, enamel and coloured stones.
18^th C. Made in England by James Cox, possibly with additional decoration in China. 

The one used on the exhibition’s publicity material (see above) is a modest example. Standing 38cm high, made of gilded copper and coloured glass, it takes the form of a crane carrying a pink-and-gold pavilion on its back, with a small clock-face attached. The crane turns back to offer an auspicious lingzhi fungus to the residents of the pavilion. It plays a tune every hour, or so we’re told: none of the exhibition’s clocks is active. 

While this is one of the least ostentatious pieces in the show, it embodies all the key attributes of the collection: the influence of foreign clock-makers (its mechanism was made in England), automated functions, the use of natural and religious imagery, and dazzling ornamentation.

A more typically bling example is the pagoda clock (Fig 1) which greets you as you enter the exhibition. Standing 1.26m high, elaborately wrought from gilded copper inset with deep blue stones, its nine tiers extend upwards like a concertina every three hours, and then retract three hours later, while the music box plays a Chinese folksong. It is one of a pair of pagoda clocks previously owned by the Emperor Qianlong. You can see its twin in action in this video from the Palace Museum (go to 02:00).  Have a look too at other examples of automation at 06:30 to 06:50. The museum workshop obviously takes great pride in restoring these devices.

Fig 3. Clock in the form of a lotus jar with opening flowers, swimming ducks, Daoist figurines and
musical device. Gilded copper, glass and enamel. Height 118cm. 18^th C. Made in the Palace workshops, Beijing.

I was very taken by this lotus jar clock (Fig 3), whose blossoms open to reveal three Daoist figurines (see below), while the little ducks swim round the pond to the sound of music.   

Fig 4. Detail from the lotus clock. The blossoms open to reveal (from left) a white ape,
the Queen Mother of the West, and a boy holding a peach. The blossoms close again when the music ends.

I could go on about the other pieces on display but here are some of the blingiest:

Fig 5. From L to R:

1) Elephant with clock on a plinth borne by four lions. Gilded copper with coloured stones. 18^th C. Height approx 100cm. Made in England. 

2) Double-gourd shaped clock mounted on two layers of automated scenes decorated with auspicious symbols. Gilded copper and enamel. 18^th C. Height 112cm. Made in Guangzhou, China.

3) Clock with knight and horse under a jewelled canopy. Gilded copper with coloured stones. 18^th C. Height 143cm. Made in England by Stephen Rimbault. The curtain rises and falls to music while a dragon flies across.

While the clocks are dazzling, the overall purpose of the exhibition is a bit vague. Is the aim just to astonish us with sparkly toys? In a display of “self-chiming clocks”, it’s disappointing to find none of them operational (a small notice explains that they are too old to be played regularly). There are a couple of videos showing clocks in action, and another video in the hall outside showing an automated farmyard scene with pecking chickens and a barking dog. Otherwise you have to read the labels and use your imagination.  

Perhaps there is a scientific purpose? The exhibition offers some models of cogs and wheels to play with but these time-keeping mechanisms were known in Europe two centuries before they arrived in China. The Science Museum has a Clockmakers gallery so perhaps this seemed like a good fit, but there’s no explicit link to it, which is strange. Also strange is the absence of an exhibition catalogue. A well-illustrated catalogue would have sold by the truckload, giving the museum some much-needed cash.

I found all this a tad frustrating because the rise of the Imperial clock collection links a number of important historical threads: notably the infiltration of foreign missionaries into China and the adoption of foreign technology by Chinese artisans. 

The Jesuit missionary Matteo Ricci (see my earlier post) first gave a pair of mechanised chiming clocks to the Wanli emperor of the Ming dynasty in 1601. Prior to this, Chinese timekeeping relied on cumbersome devices like sun-dials or water clocks. One famous water-driven clock, invented in 1092, was about 6m high (see below). Some foreign visitors were bemused that the Chinese, who invented printing and gunpowder, hadn’t figured out automated time-keeping.

Fig 6. Water-driven clock invented by imperial engineer Su Sung in 1092. Source: Cambridge University Press

The advent of European clocks running on weights and springs, much smaller and prettier than any existing devices, must have resembled the arrival of the iPhone. Eventually, any visitor seeking favour with the imperial court would include at least one luxury clock in their array of gifts, preferably sourced from elite makers like James Cox of London (French and Swiss makers were also in vogue). 

As the imperial collection grew, the Jesuits helped to repair and service these delicate toys. This, plus their skills in fine arts and astronomy, helped consolidate their place at court. They also practiced “cultural accommodation”, allowing their Christian converts to observe Confucian practices like ancestor worship. All other Christian missionaries forbade this and were duly thrown out of China by the Kangxi emperor in 1721. I like to think that the emperor didn’t want to lose the only people who could fix his precious clocks. 

Ever responsive to new trends, Chinese artisans began making their own zimingzhong, both within the Palace and around the city of Guangzhou where the Jesuits had their first Chinese mission. Their work shows a distinct blend of Chinese and European design features, as in Fig 7 where the double-gourd (symbol of health and longevity) sits on a platform borne by four rams. The animal-mounted platform (featuring lions, rhinos, bulls, etc) was a favourite detail used by British clock-makers. 

Fig 7. Double-gourd clock with mechanical features including automated scenes of Chinese rural life. Gilded copper with enamelling. Height 94cm. Dated 1790. Made in Guangzhou. Source: Liao (2002)

The Guangzhou clockmakers also catered for wealthy local residents for whom a European-style clock was a prestige item. This appeared in contemporary literature like the 18th century novel Dream of the Red Chamber, where the ownership of clocks and watches highlights the Jia family’s elite status. Eventually, of course, they became an everyday item for all levels of society.

Given this rich background, the Science Museum could have done more to flesh out the history of these devices (here's where a catalogue would have been useful!). But I’m grateful that they were able to bring these amazing pieces to London. As long as you’re happy to be blinded by bling, this exhibition is well worth a visit. 

Silk Road treasures of the Humboldt Forum

 

1. Uighur princes and princesses. Wall paintings from Cave temple 9, Bezeklik, Xinjiang, est. 8^th-9^th C.

The Humboldt Forum is the newest of Berlin’s elite group of museums, all clustered on an island in the River Spree. These include the Pergamon (with the Ishtar Gate) and the Neues Museum (home of Nerfertiti), so the Humboldt has some tough competition. But none of the others can boast of two Buddhist cave temples, carried from the Taklamakan desert and re-assembled on-site. It was to see these, and the rest of the Humboldt’s world-famous Silk Road collection, that I visited Berlin in May. 

The Humboldt is the successor to two previously separate museums: the Museum for Asian Art and the Ethnological Museum which stood in a quiet suburb of Berlin. In the early 2000s, plans were launched to unite these collections with the other treasures on Museum Island. To house them, a replica imperial palace was built from scratch amidst much controversy and huge expense. The Humboldt Forum finally opened in 2021. Look at the size of it!

2. View of Humboldt Forum from the west. Source: © Asio otus / Wikimedia Commons

Utilising the 42,000 sq m of space has not been a problem. The collection holds over 500,000 objects from Africa, Asia, Oceania and the Americas: only some 20,000 are on display but they fill up the exhibition floors easily. Elsewhere, the bright modern interior is designed to accommodate workshops, seminar rooms and vast hordes of visitors. Curators from small cash-strapped museums (i.e. the vast majority) will be prostrate with envy.

What’s on display is beautifully presented, particularly in the galleries containing artefacts from Silk Road sites in Xinjiang, China. These were found and brought home by four German archaeological expeditions between 1902 and 1914, known as the Turfan Expeditions. The most “fruitful” forays were those involving Albert von le Coq, whose name like those of Aurel Stein and Paul Pelliot is synonymous with Silk Road exploration. Their stories are told in Peter Hopkirk’s book Foreign Devils on the Silk Road, which I can’t recommend enough.

Von le Coq was only an ambitious museum volunteer when he was allowed to lead the 1904 expedition, after the boss Albert Grünwedel fell ill. He grabbed his chance and found an amazing set of deserted Buddhist cave temples at Bezeklik, near Turfan. On hearing this, Grünwedel hauled himself over to Central Asia to rejoin the team, launching the third expedition. But being a quiet scholarly type and still unwell, he was often overridden by von le Coq, which had unfortunate consequences for some of their discoveries.

Soon after Grünwedel’s arrival, the team (on a local tip-off) uncovered the great cave temple complex at Kizil.  

3. Tarim Basin, Xinjiang, China

Ancient Buddhist cave temples are found across India, Central Asia and China. The ones at Ajanta, western India, date as far back as the 2nd century BCE. In structure, they range from tiny monastic cells to multi-room spaces with hallways, niches and shrines, decorated with sculptures and paintings of religious scenes.  

Kizil comprises more than 200 cave temples dating from the 4th to 7th century CE, dotted along a cliff-face in the northern Taklamakan. The largest are up to 20m deep and 15m in height. A Japanese team tried to explore them in 1902 but were hit by earthquakes. For the happy Germans (who came back again in 1913-14) it yielded a rich store of artefacts including the so-called Cave of the Ring-bearing Doves (image 4 below), a 6th C cave which the team named after the bird motif on its walls (image 5). In the world of Silk Road scholarship, its more mundane name is Cave 123. This is one of the two cave temples reconstructed at the Humboldt.

   4. Reconstruction of Cave 123. The central niche held a statue of the Buddha, which was gone before the Germans arrived. The two side doorways to the rear cell allowed devotees to walk round the statue (circumambulate) while praying. Source: Staatliche Museen zu Berlin, Museum für Asiatische Kunst

             5. Ring-bearing dove motif, Cave 123, Kizil. Source: DW.com (left, remaining section in Kizil). Sketch by Grünwedel (right), from his book Ancient Buddhist Places of Worship in Chinese Turkestan (1912).

The bird motif is quite hard to see, especially in the Humboldt’s soft light. What catches the eye immediately is the cave’s domed ceiling, divided into eight segments depicting the Buddha with bodhisattvas. On each of the side walls is a large image of the Buddha surrounded by his disciples. Many of the colours are still vivid, derived from elements such as lapis lazuli (blue), copper (green) and gypsum (white). The cave has been rebuilt according to precise measurements from Kizil itself, so you can walk through the narrow door as worshippers did 1500 years ago and marvel at the densely decorated space, only 3m wide. 

The domed ceiling is unusual for the time and place (barrel-vaulted ceilings were more common), suggesting a Middle Eastern influence from the nearby trading hub of Kucha. The Silk Road was an international trade route, so these caves were not as isolated then as they seem now. 

6. Domed ceiling of Kizil Cave 123. Coloured segments are originals; brown patches are filled in

7. Moving the two-tonne dome of Cave 123 to the Humboldt in 2018. For full story see this link.
Source:  Staatliche Museen zu Berlin

The reconstructed cave sits in the middle of a large hall whose walls are hung with other painted stone slabs brought back by the expeditions. Von le Coq called them frescoes but technically they are murals, painted on dry (not wet) surfaces. Some of these are over 2x2 metres, comprising several pieces fitted back together. Von le Coq and his assistant Theodor Bartus (the muscles of the team) were proud of their prowess at cutting out these slabs and packing them for transport; the method is carefully described in von le Coq’s memoirs. Grünwedel would have preferred to copy the images and leave them in place but was often over-ruled. 

Von le Coq and Bartus had had plenty of practice cutting up cave temples on their earlier foray at Bezeklik. Their haul from there included images of Uighur princes and princesses (image 1) whose colours were perfectly preserved under centuries-old layers of dry sand. What isn’t on display, though, is a set of fifteen huge Buddhist wall-paintings (each about 3x2 metres) from the same cave, which von le Coq also dispatched to Berlin. Because of their size and weight, the paintings could not be moved to safety in WWII, and all were destroyed when the old museum was bombed.

 

They were however reproduced in von le Coq’s book Chotscho (1913), which I consulted in the British Library. The book itself is like a concrete slab but it was well worth seeing its breathtaking images like the one below. The lines where the segments were stuck back together are visible. Chotscho was the German spelling of Khocho, now Gaochang, the nearest town to Bezeklik.

                8. Buddha in a pranidhi scene from Cave temple 9, Bezeklik, Xinjiang, 8-9^th C. Original dimensions 3.25m x 2m.                    Pranidhi scenes, popular in Uighur Buddhism, depict past buddhas foreseeing the enlightenment of the present Buddha Sakyamuni. Source: Chotscho (1913), von le Coq

Back in the Humboldt, the adjoining gallery houses the second of its two caves: the Cave of the Sixteen Swordbearers (Cave 8), also from Kizil, dated 5th– 6th C. In this room the curators have tried to suggest the cliffs and sand dunes of the Taklamakan by building tall sloping blocks with doorways and narrow passages. The huge space is very dark, emphasising the dim light from the doorways.  Affixed along the passage walls are four paintings, each depicting four ornately dressed sword-bearing men who were probably the sponsors of the temple. These were found on the passage walls leading into the cave’s main chamber. 

9. One of four segments from the Cave of the Sixteen Swordbearers (Cave 8), Kizil, 5^th-6^th C. Approx 1.2m wide.

The men’s appearance has provoked much discussion due to their reddish-brown hair, light-coloured eyes, and Persian-style clothes with Sassanian motifs (see third figure from left). Men with similar features are seen in other paintings here, such as the monks in the Cave of the Painters (Cave 207, image 10). They are commonly described as Tocharians, which some scholars find unsatisfactory because it derives from an early analysis of their language which mistakenly linked them to the Tokharoi people of Afghanistan. Some scholars claim that they are Sogdians or Hephthalites. But there has been no conclusive identification so the term “Tocharian” is still used.

                     10. Buddha with bodhisattvas and monks, Cave of the Painters (Cave 207), est. 6^th C. The monks in the lower left have reddish hair and pale round faces.

The two Kizil cave temples are the jewels in the Humboldt’s Asian collection. But like other major museums in the Western hemisphere, the Humboldt is being challenged as to why it still possesses these treasures. Their provenance is obviously suspect: the argument that archaeologists “rescued” them is only partly true. Some were decaying or looted, but others were in good condition and could have been copied and preserved on-site as recommended by Grünwedel. When you stand in the Cave of the Ring-bearing Doves, you wonder how von le Coq had the audacity to cut it up and send it 6000km overland to Germany. The fifteen great paintings from Bezeklik might have survived if they had not gone to Berlin: they were doing very well before von le Coq and Bartus turned up. 

Equally, it is certain that many caves had been looted or destroyed by locals even before the Germans arrived; the British and French teams reported the same. Aurel Stein, working in the southern Taklamakan, re-buried a set of stucco Buddhist statues which were too large to move, hoping to study them later. But before his return, the statues were found and destroyed by tomb-robbers looking for treasure. Elsewhere, poor farmers used the coloured pigments from cave murals as plant fertiliser. Xinjiang was China’s wild west then: the Qing dynasty was dying and the average citizen of this remote corner was just trying to survive. It was not until the 1930s that the Chinese government officially banned exploration by foreign teams.

A more constructive question is whether the Humboldt should repatriate these treasures, as it is doing with its Benin bronzes from Africa. China’s vast modern museums could house them easily; in fact the Kizil and Bezeklik caves are now well-developed tourist sites. But this raises the question of whether the artefacts should go to Beijing/Shanghai or more correctly to Xinjiang, just as the Benin bronzes have gone home to the kingdom of Benin within Nigeria, and not to Lagos. Perhaps the Humboldt would rather not wade into the hot mess of Beijing-Xinjiang relations.

My trip to Berlin ended in another hot mess (British Airways flight cancellations) but I’m delighted that I made the trip. Apart from anything else, entry to the Humboldt is completely free, thanks to Neil MacGregor, the ex-British Museum boss who headed the project from 2015 to 2018. Berliners who were horrified by the project’s €680mn price-tag may not be happy but I am likely to visit again, just not flying BA.

Bibliography

Buried Treasures of Chinese Turkestan (1928), Albert von le Coq

Chotscho (1913), Albert von le Coq

Altbuddhistische Kultstätten in Chinesisch-Turkistan (Ancient Buddhist Places of Worship in Chinese Turkestan) (1912), Albert Grünwedel

Foreign Devils on the Silk Road (1980), Peter Hopkirk

“A Place of Safekeeping? The Vicissitudes of the Bezeklik Murals”, Susan Whitfield, from Conservation of Ancient Sites on the Silk Road (2010), Getty Centre

The best things I learned about ceramic glazes

The study of ceramic glazes was a challenging summer project, to say the least, but I came away even more impressed by the ingenuity of ancient Chinese potters. The project made me re-assess my views on some of their wares, one category in particular. And I only covered a few Tang and Song ceramics: the others like Yue, Yaozhou, Qingbai and all the post-13th century wares will have to wait, as my new academic term begins in a few days.

I wrote four detailed posts (not counting this one) and I'm sorry if they don’t make light reading: as I said in July, this was something I needed to get my head around. To make up for that, here is a "Best Of" list to give you an idea of the most interesting features. There are embedded hyperlinks in case you’re unexpectedly overcome by a thirst for knowledge!

1. Si, Al, Colour and Flux
Not the members of a Cantonese boy-band but the four things you need in a glaze. Silicon (chemical symbol Si) makes it glassy, aluminium (Al) makes it clingy, and various minerals like iron, copper or cobalt give it colour. But the key component is the flux (such as lead, potassium or soda ash): it is the go-between agent that encourages the other elements to melt easily and coat the pot smoothly without cracking. This is why glazes are often referred to by their fluxes: lead glaze, lime-ash glaze, lime-alkali glaze, etc. More about glaze chemistry here (about halfway down).

2. Iron makes the most amazing colours

From left: Tang sancai horse and rider, Ru dish (upper), Longquan dish (lower), Longquan vase with iron-spot decoration.

We think of iron as a heavy black metal, but all the above objects owe their colours to varying levels of iron in the glaze. The amber-brown shade of sancai ware, the blue-greens of Ru, Guan and Longquan, and of course the iron-spots of Longquan tobi seiji, are all derived from iron. The final colour depends on various things: the percentage content (as little as 2% iron can be enough), the level of titanium dioxide in the glaze (too much makes the iron go yellow), and the temperature and atmosphere of the kiln. Next time you admire a dreamy green celadon, remember that its colour comes from the same mineral later used to make railway engines.

3. Ru is rare, Guan is brilliant…

Ru ceramics are famously rare (fewer than 100 pieces remain), beautiful and staggeringly expensive. But it turns out that the glaze, while lovely beyond doubt, is not unusually complex from a technical POV. It’s a reworking of ancient Shang ash glazes but with a bit of iron and a lot of obsessively careful firing. You could argue that this simplicity is part of its charm but I suspect that the enduring popularity of Ru is more about its rarity and historic links to the tragically deposed Northern Song. More details in the middle section of this post.

Ru ware narcissus bulb planter with uncrackled glaze. China, Northern Song Dynasty, 12th C. 

National Palace Museum, Taipei.

Guan is partly inspired by Ru ware but it’s a very different creature. The abstract patterns of its crackled glaze are bold and modern, yet they also reflect the dynasty's sad past. More importantly, the potters’ idea of using low-silica materials to make this crackle (when the received wisdom was that silica prevented cracking) was a brilliant bit of experimentation, especially when working with unfamiliar southern clays. More info in this post.

Guan dish, D: 22cm. Southern Song, 12-13th C. Met Museum NY

4. But Jun is astounding

The glaze that most impressed me was Jun, a blue ceramic from the Northern Song period. Jun ware has a misty shimmering quality known as “opalescence”, sometimes enhanced with purple splashes of copper pigment. It is not especially rare, and I always thought it was a bit florid by Song standards.

Jun ware dish, Northern Song period, 11-12th C. British Museum (Percival David Collection)

But I hadn’t realised that Jun glaze is not really blue at all. Its colour comes primarily from its molecular structure which scatters blue light in the same way as the Earth’s atmosphere makes the sky look blue. The glaze on its own is straw-coloured but if applied thickly, it can make the underlying object look anything from milky blue (if under-fired) to purplish blue (high-fired). It was not until 1983 that two scientists worked out the chemistry behind this, introducing the phrase “liquid-liquid phase separation” to the world of art history. For more info, see the last section of this post.

While researching these posts, my most frequent question was “how did they work that out?”. How did potters know that adding crumbly white stone to the glaze would make it come out green? I tried asking a couple of modern ceramic artists about their glazes and wow, they clammed up fast, as if I’d asked for their PIN numbers or email passwords. Because of course it took years of work and repeated failures before the glaze finally came right: they weren’t going to just tell me all. Likewise with the potters of the imperial kilns (more to the point, the emperor would have been furious if they revealed the secrets of imperial wares). But I hope at least they would have been entertained by my efforts to study their work. 

Main sources used: Chinese Glazes (1999), Nigel Wood; Science & Civilisation in China, V5 Pt 12 (2004 edition), Rose Kerr and Nigel Wood; Song Dynasty Ceramics (2004), Rose Kerr.