December 14, 1999
originally published in Orientations, April 1990
In The Newark Museum collection is a small ivory figure representing a form of Avalokites'vara of a type that has long puzzled historians of Himalayan art (Fig. 1). Figures displaying the stylistic eccentricities of this bodhisattva are, as evidenced here, relatively common. These stylistic eccentricities can be briefly catalogued as: a high three-lobed crown of rather simple design; the hair in an elaborate chignon which spills in two long buns on either side of the head and crown, and bell-like earrings. The images also show a remarkable lack of ornamentation; they stand on a small, square base in a relatively stiff pose, with, when complete, the right hand in varada mudra (gesture of bestowal) and the left close to the thigh in a gesture of holding a (missing) lotus.
Figures of this distinctive type can be found in many private and museum collections and almost every broad-based collection of Himalayan art has an example. Several of these figures are illustrated here and it is worthwhile examining them as a group.
Although the Newark bodhisattva is unusual in that it is made of ivory, it is not the only ivory example known. A larger figure in the Victoria and Albert Museum (Fig. 2) has many characteristics in common with the Newark version. It differs only in being more complete, as the Newark example is missing the left hand, which in the Victoria and Albert figure is present. Spink & Son in London also has a figure of ivory with a face that appears to have been slightly recut (Fig. 3).
More common are figures in wood. One of the most complete, where all the deity's iconographic characteristics are displayed, is in the Los Angeles County Museum of Art (Fig. 4). Here the right arm in varada mudra, missing in the Newark and Victoria and Albert examples, is present. The Los Angeles version also has what would appear to be a standing Buddha or bodhisattva in the central lobe of the crown, the smaller figure's hand gestures mirroring those of the larger. The face of the Los Angeles example is painted in the 'cold-gold' technique, in which a mixture of powdered gold and size is applied, a reliable indicator of the figure's Tibetan provenance (if not necessarily origin).
Other wooden examples include a complete and well-preserved figure strikingly painted in gold in the collection of Wesley and Carolyn Halpert in New York (Fig. 5); a figure in the Musée Guimet (Fig. 6), a painted wooden figure at Spink & Son (Fig. 7) and a particularly elegant piece in a private collection (Fig. 8). The Zimmerman Collection in New York contains a torso of the same bodhisattva, unusual for its unadorned crown and worn appearance, suggesting an age greater than the examples already noted (Fig. 9).
Although wood is the most popular material for the manufacture of these images, metal figures are not unknown. The Pritzker Collection in Chicago has a fine example, of a heavy cast copper alloy (Fig. 10), and another, seemingly of later manufacture but with the same stylistic characteristics, is in the National Museum in Kathmandu, Nepal (Fig. 11). The latter is notable in that, unlike the other figures, it has a seated figure of Amitabha in the crown rather than a standing Buddha.
The possibility that all the known examples were patterned after a single, particularly sacred image is strongly suggested by the fact that certain peculiarities have been reproduced almost exactly in images of divergent materials and levels of craftsmanship. Such patterning after a physical (rather than theoretical) original is not common in Himalayan art, but there is at least one other known example of such a custom. The sacred image of Avalokites'vara popularly known as Bumga dyah of Bungamati and Patan in the Kathmandu valley has often been copied in various materials for household altars and shrines. The peculiar mask-like visage of this figure (Fig. 12) and the related Avalokites'vara image of Jambaha dyah of Janbahal in Kathmandu (Fig. 13) is instantly recognizable in the copies (Fig. 14; see also Slusser, Nepal Mandala. Fig. 596). It is not entirely coincidental that these images share similar inconographic peculiarities with the others under discussion, notably the mudras, as well as stylistic peculiarities, such as the unusual head-dresses, lack of ornamentation, and small, square or rectangular bases.
0n a trip to Tibet in the summer of 1986, this author encountered similar bodhisattva figures in many of the major monasteries in the Lhasa valley, including Drepung ('Bras spungs), Sera (Se-ra) and various 1hakhangs (shrines) in the Potala Palace. The answer to the riddle became clearer when a monk attending one of the 1hakhangs in Sera provided the information that an image on display in that lhakhang was known as 'Potala Lokes'vara'. A visit to the Potala appeared to confirm that the central image of Avalokites'vara known as 'Phagpa Lokes'vara' ('Phags-pa Lo-ke-svara, Sanskrit:, 'Arya Lokes'vara'; 'Noble Lord of the World') of the Phagpa Lhakhang, the oldest and most sacred of the Potala shrines (Fig. 15), is the original of which all the known examples are copies.
This deduction might have been obvious to the first visitor with an interest in art history to the Phagpa Lhakhang were it not for the Tibetan custom of draping and decorating images, particularly sacred images, to the point of entirely obscuring everything but the face; in this case even the figure's peculiar earrings were concealed. Unfortunately, according to Gyatso, the monk attending the lhakhang, the vestments and late repoussé crown are never removed from the image of Phagpa Lokes'vara except when the image is to be repainted, an event that occurs only when a donor comes forward with the requisite gold and an artist to perform the repainting; this was subsequently confirmed by the scholar Lobsang Lhalungpa, who worked in a nearby office in the Potala in the mid-1940s and remembers this ceremony. Although it was impossible to view or photograph the image unencumbered, parts of the high-peaked crown and unique hair-style were still visible. Further, Gyatso confirmed that the central image, when disrobed, does very closely resemble several reproductions kept in the same lhakhang (Figs 16 and 17) and in other shrines in the Potala, including an ivory figure in the Dukhor (Dus-'khor) Lhakhang. Copies of the Phagpa Lokes'vara are in fact widespread throughout the Tibetan cultural domain and not confined to the Lhasa valley; two examples have been found on an altar in Chemre (ICe-'bre) monastery in Ladakh (Fig. 18) and another was photographed at Nako monastery near Tabo in an area culturally part of western Tibet (Fig. 19).
After several adventures the mendicant found the sandalwood tree:
This story of the 'four brothers', found with variations in many Tibetan histories, is a fascinating glimpse into the legend/history of the origins of Buddhism in the Himalayas. Two of the other figures are relatively easy to identify. The noble 'dBu-gang' is Bumga dyah/Karunamaya/Rato Macchendranath of Bungamati/Patan, a figure which has been copied in much the same way as Phagpa Lokes'vara; the noble IJa-ma-li' is Jambaha dyah, the so-called 'White Macchendranath' of Janbahal in Kathmandu, who in some Nepalese origin stories is said to have been stolen at one point and carried away by a rival king, perhaps to the 'India-Nepal border' mentioned in the story, before being returned to Kathmandu. One Tibetan version of the four brothers' story places IJa-ma-li in the temple of Khojarnath, on the Karnali river at the western border of Nepal and India (Wylie, p. 14, n. 20). It may have been to here that IJa-ma-li was temporarily taken from his home in Kathmandu, perhaps by one of the audacious Malla kings of medieval western Nepal/western Tibet.
There is no doubt that, as Kramrisch pointed out, the copies of Phagpa Lokes'vara examined here 'adhere to an ancient type'. There is no telling argument why we should reject, on stylistic grounds, a seventh-century date for the Phagpa type (until there is an opportunity to view the image uncovered we must refer to the 'type'' rather than the image itself). Certainly the simplicity of the figure, similar in many respects to the earliest known figures of the standing bodhisattva, would suggest such a date. An examination of South Asian sculpture from this period clearly shows that the early figures that most resemble this type are Nepalese. One of the earliest known bodhisattva figures from Nepal, a seventh-century bronze Vajrapani in The Metropolitan Museum of Art (Fig. 20), exhibits a similar simplicity of design (Lerner, Fig. 16).
The stylistic peculiarities already mentioned seem so archaic as to further support an early date, for often in images that predate the formation of a mature tradition experimental features are found. As examples of such experimentation we may cite the mitre-type crown on the seventh-century figure of Vajrapani noted above; another unusual crown on a sixth-century stone figure of Avalokites'vara (Fig. 21) near Gana Bahal in Kathmandu (this image also displays the simplicity and lack of ornamentation that bespeak its considerable antiquity, and has the same Buddha figure found in the crown of Jambaha dyah), and the elegant and unique hair-style displayed by a small, early image of Vajrapani (Slusser, vol. 2, Fig. 464). The mitre-like crown appears to be characteristic of Bumga dyah and a modified, three-lobed version appears on Jambaha dyah in Kathmandu, although the crown of Phagpa Lokes'vara is very different. Mary Slusser notes the occurrence of this mitre-like crown, which she mentions was normally reserved for Indra, on several later Nepalese portrayals of Avalokites'vara, and it may well be that these were patterned after the crown-type of Bumga dyah (see Slusser, vol. 2, Figs 462 and 463).
The location of Phagpa Lokes'vara is consonant with an early date for, although the present Potala Palace was not built until the time of the Great Fifth Dalai Lama (1617-82), there is little doubt that certain parts of the site are far older, including both the Phagpa Lhakhang and the cave containing images of the great emperor and his two queens. When The Clear Mirror of the Royal Genealogies was written, the Phagpa Lhakhang may well have been situated 'at the top of the Po-ta-la'.
If we do accept a seventh-century date for the original Phagpa Lokes'vara, this does not mean that the numerous, widely dispersed copies are of this early date. Any attempt to date these images without the benefit of scientific analysis would be an exercise in futility. It is also evident that, although the original Phagpa Lokes'vara may well be of Nepalese origin, the majority of the copies would logically be Tibetan, a dichotomy which has added considerably to the confusion surrounding these figures.
This article has been little more than a brief introduction to a subject rich in unexplored possibilities. There are still many questions regarding the Phagpa Lokes'vara, including the implications of his specific iconographical identity as Khasarpana, a youthful form of Avalokites'vara (the authoritative Sadhanamala notes that Khasarpana 'resides in the womb of Mount Potalaka'; Bhattacarya, p. 129). His relationship to the two figures on either side of him in the lhakhang is also mysterious. (According to Lobsang Lhalungpa, the three together are known as "phags-pa-sku-mched-gsum' ['three noble brothers'], somewhat confusing the story of four brothers found in The Clear Mirror of the Royal Genealogies).
In spite of the mystery which shrouds these figures, the legend and history surrounding Phagpa Lokes'vara and the available stylistic evidence suggest that Phagpa Lokes'vara, and at least one of his brothers in Nepal, are indeed precious relics of the early years of Mahayana Buddhism in the Himalayas. These two at least share characteristics unique to their cults: stories of royal introduction, peculiarities of iconography and style, and the otherwise unknown custom of manufacturing copies not of an iconographical type but rather a specific sacred image.
Ian Alsop has published extensively on the cultural history of Nepal and is currently working on a lexicon of classical Newari.