Tibet Archaeology

and all things Tibetan

Flight of the Kyung

November 2017

John Vincent Bellezza

Welcome to Flight of the Khyung and the world’s highest land: Upper Tibet! This month and the next two issues feature a new monograph presented here for the first time, leapfrogging hardcopy publication. The work describes ancient ceremonial monuments portrayed in the rock art of uppermost Tibet. These stepped shrines were carved and painted in many different styles, a tribute to the artistic and religious sentiments of Tibetan ancestors. The shrine depictions tell us much about how the higher reaches of the Tibetan Plateau changed religiously over the course of the first millennium of our Common Era. From the time in which archaic cults held sway until the introduction and consolidation of Buddhism, stepped shrine rock art highlights cultural milestones in the development of Upper Tibet.

Other recent publications by the author:

2017: “The Swastika, Stepped Shrine, Priest, Horned Eagle, and Wild Yak Rider: Prominent antecedents of Bon figurative and symbolic traditions in the rock art of Upper Tibet”, in Revue d’Etudes Tibétaines, no. 42, October 2017, pp. 5–38. Paris: CNRS.

2017: “Zenpar: Tibetan Wooden Molds for the Creation of Dough Figures in Esoteric Rituals”, in Arts of Asia, September-October, 2017. Hong Kong.

2017: “The Rock Art of Spiti: A General Introduction”, in Revue d’etudes tibétaines, vol. 41, pp. 56–85. The Spiti Valley: Recovering the Past and exploring the Present (eds. Y. Laurent and D. Pritzker). Paris: CNRS. http://www.himalaya.socanth.cam.ac.uk/collections/journals/ret/pdf/ret_41_03.pdf

2017: “Kailas Histories: Renunciate Traditions and the Construction of Himalayan Sacred Geography, by Alex McKay”, book review in The Tibet Journal, vol. 42 (no. 1), pp. 103–111. Dharamsala: LTWA.

2017: “Discussion of “A 5500 Year Model of Changing Crop Niches on the Tibetan Plateau Current””, article review in Current Anthropology, vol. 58 (no. 4) pp. 537, 538. Chicago: University of Chicago Press.

A Comprehensive Survey of Stepped Shrines in the Rock Art of Upper Tibet: With Archaeological, Cultural and Historical Comments*

Dedicated to Theo Bellezza

This monograph is organized as follows:

Introduction to the stepped shrine rock art of Upper Tibet (November 2017)

List of rock art sites with stepped shrines

Part I – The origins of stepped shrines in Upper Tibet (November 2017)

Elementary stepped shrines

Part II – Typological illustrations and data on stepped shrines in Upper Tibetan rock art (December 2017)

Key to typological inventory
Group I: elementary two-tiered shrines
Group II: elementary three-tiered shrines
Group III: elementary four-tiered shrines
Group IV: elementary multi-tiered shrines
Group V: elementary segmented shrines
Group VI: elementary shrines with spatulate or crescent-shaped finials
Group VII: elementary shrines with multi-foliate finial
Group VIII: elementary shrines with tricuspidate finial
Group IX: idiosyncratic elementary shrines
Group X: shrines with small bulbous upper section
Group XI: twin shrines sharing common base
Group XII: simple style chortens
Group XIII: chortens with forked finial
Group XIV: chortens with cross-piece spire
Group XVI: flat-topped chortens
Group XVII: chortens with simple spires
Group XVIII: intricate non-Buddhist chortens
Group XIX: intricate chortens of Brag gyam
Group XX: Buddhist chortens

Part III – Cross-cultural influences acting upon stepped shrines in Upper Tibetan rock art (January 2018)

Northern Pakistan

Conclusion (January 2018)

Bibliography (January 2018)

Catalogue of photographs of stepped shrines in Upper Tibetan rock art (January 2018)

The delivery of my lecture at the IATS XIV conference in Bergen, Norway in 2016 (based upon the findings of this monograph), was made possible by a travel grant awarded by the Lumbini International Research Institute (Lumbini). For the article produced from that lecture see Bellezza in press-b. The drafting of this monograph in 2016 was enabled by a grant from the Shelley & Donald Rubin Foundation (New York). My documentation of rock art in Upper Tibet in more recent years was supported by grants conferred by the Unicorn Foundation (Atlanta), Tise Foundation (Chicago) and by Joseph Optiker (Berglen). Most of the black and white drawings of stepped shrines in this work were commissioned from the Tibetan artist Lingtsang Kalsang Dorjee and his atelier. Another 40 drawings of stepped shrines as well as the thog lcags appearing in this work were kindly done by Rebecca C. Bellezza. Stepped shrine specimen L7 was produced by Martin Vernier. My heartfelt thanks to all the institutions and individuals who contributed to the publication of this work.

Introduction to the stepped shrine rock art of Upper Tibet


This work furnishes a comprehensive survey of stepped shrines in the rock art of Upper Tibet (Byang thang and Stod), a region of some 600,000 km².* Stepped shrines represented in the rock art of the region are of two major types: elementary tiered structures and complex tiered structures. The latter architectural type is known generically in Tibetan as chorten (mchod rten).†

Situated in what is now known as the Tibet Autonomous Region (TAR; Bod rang skyong ljongs).

Mchod rten literally means ‘support/receptacle of offerings’ (cf. Tucci 1932: 13). The spelling ‘chorten’ has emerged as the most common usage in the English language, a convention followed in this article.

Stepped shrines were documented by the author over the course of eighteen years (1995–2013), on twenty different expeditions to Upper Tibet.* A total of 232 Upper Tibetan stepped shrines in graphic form are included in this typological and analytical study. Rock art shrines were recorded at 37 sites, spanning the region from the southeastern corner of Gnam mtsho to the western portions of Ru thog and Sgar abutting the international ‘Line of Control’ with India.† Four-fifths of the stepped shrine rock art sites are clustered in just three areas of Upper Tibet: Gnam mtsho, Nag tshang-G.yag pa (Shan rtsa) and Ru thog.

For earlier presentations of my findings, see Bellezza 2017d, pp. 14–20; 2002, pp. 128, 129, 144, 145; 2014f, pp. 189–193; 2013g.

I have surveyed 76 sites in Upper Tibet with rock art consisting of more than just swastikas, thus nearly half of the total number of sites have stepped shrines depicted. For a list (incomplete) of rock art sites and their locations in Upper Tibet, see Bellezza 2008, pp. 683–686; 2014b, pp. 589– 591. Despite my extensive survey work in Upper Tibet, this study should not be construed as exhaustive. There is always something more to discover.

The most salient graphic feature of stepped shrine rock art in Upper Tibet is the stacking of different sized rectangles and squares upon one another to create tiered arrays. These strata are typically graduated with each successive quadrate reduced in size.* In more intricate examples, there are two or more sets of layered rectangles interspersed by squares and other geometric forms. The towering appearance or dominant vertical element of stepped shrines is often enhanced by the addition of spires and masts rising above the tiered base.†

Stepped structures are known throughout the world, an almost universal architectural form. Among the most spectacular examples are the stepped pyramids of Egypt, the ziggurats of Mesopotamia and western Iran and the pre-Hispanic temples of Mesoamerica.

The verticality of chortens is magnified by the central axis (srog shing) that runs along its entire height, its most fundamental component. According to a cross-cultural mythological perspective, the central pole represents the axis mundi and is related to the world mountain, Ri rab lhun po (Meru), and the tree of life (Ricca and Lo Bue 1993: 33, 34). Regarding the architecture of the early Indian stupa, however, Fussman (1986) does not find clear evidence of this cosmological theme.

Both chortens and simpler stepped shrines in Upper Tibetan rock art are portrayals of monuments with doctrinal, mythological and ritual significance, as widely attested in Tibetan literature and common practice. Nevertheless, it is very unlikely that the same mix of functions and purposes assigned stepped shrines in the extant Tibetan tradition also characterized the cultural scene in antiquity. As is well known, Tibetan religious traditions underwent fundamental developments in the last 1200 years, altering or effacing earlier intellectual and material traditions.

The Tibetan chorten, especially that of the Buddhists, is conventionally seen as the equivalent of the Indian stūpa (hereinafter: stupa).* The extant functions of the Tibetan chorten can be summarized as follows:†

  1. Model of the Buddha, particularly his mind (thugs)
  2. Monument commemorating the life events of the Buddha and the deeds of great Buddhist teachers
  3. Physical symbol of the cosmos and five elements
  4. Auspicious and meritorious addition to the landscape
  5. Ritualized instrument to prevent harm and misfortune
  6. Paradigm for teaching religious doctrines
  7. Reliquary for the mortal remains of high lamas and saints
  8. Receptacle for sacred substances

As is well known, the prototype of the Indian stūpa was the burial tumulus, which evolved into a cult monument to enshrine the relics of the Buddha. For studies of the Indian stupa with reference to its funerary and reliquary antecedents, see, among others, Combaz 1933; Pant 1976. Also see De Marco 1987; Thewalt 2000.

For a discussion of the basic functions and symbolism of chortens, see Tucci 1932, pp. 23–59; Cayton 1998; Harvey 1984. Also consult Anagarika Govinda 1976; Beer 1999, pp. 127–135.

In this monograph, the stepped shrines of Upper Tibetan rock art (petroglyphs and pictographs) are divided into twenty different groups according to appearance. These groups are categorized beginning with the simplest forms and progressing to the most intricate, charting the stylistic development of these ritual monuments in rock art. There is a continuum in the styles of stepped shrines, making it difficult in certain cases to distinguish a chorten from alternative kinds of tabernacles. Indeed, both types of monuments are closely related culturally and historically. As this work will demonstrate, the Tibetan chorten is as much an architectural product of pre-existing indigenous monuments as it is the Indian stupa.


Dates for the introduction of simple stepped shrines and chortens in Upper Tibetan rock art have not been determined with any surety. Elementary stepped shrines were not only carved in rock but also erected as stone structures in the region. Some of these ruined constructions occur at residential complexes predating the 7th century CE but precise dates are not yet available. Chortens were introduced in Upper Tibet with the advent of Buddhism and the cosmopolitan cultural influences accompanying it. Unlike Indus Kohistan and Ladakh, which had early contacts with Buddhism, chorten rock art did not reach Upper Tibet until this religion gained a foothold there in the 7th to 9th centuries CE. The designs of stepped shrines in rock art cannot be used in isolation to ascertain the age of individual examples. That is because elementary forms continued to be created alongside more complex varieties in Upper Tibet.

Unfortunately, there is still no scientific method for dating rock art reliably and consistently. This forces researchers to fall back on informed methods, especially when collateral archaeological evidence is unavailable for analysis. I have devised a relative chronology for Upper Tibetan rock art, serving as a provisional means for dating in broad terms (subject to verification whenever appropriate technologies and methods become available). The major criteria I employ in an assessment of the age of rock art, including stepped shrines, can be outlined as follows:

Stylistic and thematic categorization of motifs, subjects and scenes
Appraisal of the general characteristics of the contents of sites
Cultural and historical analysis of complementary textual and ethnographic materials
Cross-cultural comparisons with rock art in other regions
Application of collateral archaeological data
Gauging ecological conditions depicted in rock art
Assessment of the techniques of carving and painting
Examination of the degree of erosion and re-patination of carvings and the degree of browning and ablation of pigments.
Determination of the placement of palimpsests

Relying on a chronological analysis of rock art based on the above criteria, I have developed a dating regime for the rock art of Upper Tibet presented in this work as follows:

Late Bronze Age (ca. 1200–700 BCE)*
Iron Age (ca. 700–100 BCE)
Protohistoric Period (ca. 100 BCE to 600 CE)
Early Historic period (ca. 600–1000 CE)†
Vestigial period (ca. 1000–1300 CE)‡

Differential dating of Late Bronze Age and Early Iron Age is omitted due to lack of chronological data.

The Early Historic period is subdivided into two phases: Imperial period (ca. 600–850 CE) and post-Imperial period (ca. 850–1000 CE).

On the dating of Tibetan rock art using a combination of methodological approaches, see Bruneau and Bellezza 2013, pp. 6–9; also see Tang Huisheng and Gao Zhiwei 2004; Suolang Wangdui 1994, pp. 33, 34; Chayet 1994, pp. 55, 56. To maintain continuity throughout this work, the chronological terminology presented here is also applied to Ladakh and northern Pakistan to denote the same time periods.

List of rock art sites with stepped shrines

Upper Tibet in the wider region. Map by Brian Sebastian and J. V. Bellezza.

Map of rock art sites in Upper Tibet with stepped shrines. Note that modern international boundaries are not designated. Map by Brian Sebastian and J. V. Bellezza.

Site No. Site Name County Type
1 Brag gdong East Ru thog Petro
2 Ser mdzod rdo ring Ru thog Petro
3 Rwa ’brog ’phrang Ru thog Petro
4 Sngo sog Ru thog Petro
5 Rno bo g.yang rdo Ru thog Petro
6 Nag khung rdo ring Ru thog Petro
7 Chu mkhar gyam sgrub phug Ru thog Picto
8 Mchod rten sbug sna kha Ru thog Petro
9 Mtha’ kham pa ri Ru thog Petro
10 Ri mo gdong Ru thog Petro
11 Glog phug mkhar Ru thog Petro
12 Gyam rag East Ru thog Petro
13 Gna’ ba lung Ru thog Petro
14 Skabs ren spungs ri Ru thog Petro
15 Rdzong chung Ru thog Petro
16 Sgog ra Ru thog Petro
17 Ru thog rdzong Ru thog Petro
18 Brag gyam Sgar Petro
19 Rdu ru can Rtsa mda’ Petro
20 Gser sgam Rtsa mda’ Petro
21 Ri rgyal Sger rtse Petro
22 Brag khung mdzes po Nyi ma Picto
23 Dar chung Nyi ma Petro
24 Bshag bsangs Nyi ma Petro
25 Rdo ’khor phug pa Shan rtsa Picto
26 Rta ra dmar lding Shan rtsa Picto
27 Chu ro Shan rtsa Picto
28 Skyid sgrom sgo gru bzhi Shan rtsa Picto
29 Sgar gsol brag phug Shan rtsa Picto
30 Slop dpon phug Shan rtsa Picto
31 Rta mchog ngang pa do Dpal mgon Picto
32 Ra ma do Dpal mgon Picto
33 Lug do Dpal mgon Picto
34 Stong shong phug Dpal mgon Picto
35 Se mo do South Dpal mgon Picto
36 Bkra shis do chung ’Dam gzhung Picto
37 Bkra shis do chen ’Dam gzhung Picto

Part I – The origins of stepped shrines in Upper Tibet

Elementary stepped shrines

Elementary stepped shrines in Upper Tibetan rock art seem to be related to contemporary ritual monuments used in the placation and regulation of elemental, territorial, protective and ancestral deities throughout Tibet. These types of constructions are variously known as rten mkhar, gsas mkhar, lha mkhar, lcog mkhar, lha rten, lha gtsug, lha tho, btsan khang, klu khang, and klu ’bum, etc. They come in many forms and sizes from heaps of rocks to elaborate chapels. These structures are a ubiquitous presence in Tibet, an indispensable part of the ritual life of households, villages and monasteries.

It is not known how ancient Upper Tibetans called the elementary stepped shrines they constructed or carved and painted on rock surfaces. Stepped shrines for the worship and control of elemental deities comparable in form to ancient rock art and built examples are still used in western Tibet.* The deities to which they are dedicated include yul-lha and gzhi bdag (territorial spirits), btsan (red protective spirits of the rocks), gnyan (protective and ancestral spirits), klu (water spirits), and sman (female spirits of various kinds).† The Tibetan textual tradition indicates that such shrines and the ritual practices associated with them are of considerable antiquity.

For examples of stepped shrine constructions still in use, see Bellezza 2008 pp. 144, 145, 147; 2014b, pp. 521, 522.

The most comprehensive survey of the Tibetan pantheon of elemental and protective deities remains Nebesky-Wojkowitz 1956.

Pl. 1. ’Brog pa women making incense offerings in a brazier before a rten mkhar shrine for the celebrated mountain god Gnyan chen thang lha, ’Dam gzhung. Photograph taken in 1997.

Pl. 2. One of two large stepped shrines called Gu ru ’bum pa in Sdod yod, Spu rang/Pu hreng (southwestern Tibet). They are constructed from cobbles laid in dry-stone random courses. According to local legend, these shrines were magically built by the adept Gu ru rin po che. Their form is very much in keeping with stepped shrines in the rock art of Upper Tibet. On this monument see, Bellezza 2014b, pp. 521–523; 2008, p. 145.

The oldest Tibetan textual reference to come to light mentioning shrines is a divination manuscript, Pt 1052. This scroll from the Dunhuang collection of the Bibliothèque nationale de France is composed in the Old Tibetan language and probably dates to the late Imperial period. Called lha-mkhar (castle of the lha) and gsas mkhar (castle of the gsas), these shrines are named after two important and closely related classes of deities, the lha and gsas.* These spirits have a celestial, ancestral and tutelary identity in Old Tibetan literature of the Early Historic period. At least in more recent times a variety of different kinds of deities are enshrined in these types of constructions. The unnamed divinities that dwelt in the Pt 1052 examples were harnessed to insure a uniformly dependable prognosis in divination.

Gsas-mkhar along with incense braziers (bsangs-khang) were erected near homes, and deer antlers, spears and juniper boughs, etc. were stuck on top of the shrine. They contain a receptacle (rten-mkhar) for the god of males (pho-lha), warrior god (dgra-lha) or territorial god (yul-lha). The gsas-mkhar of powerful families typically acted as the main shrine of the yul-lha of a locale. Inside the rten-mkhar is the inner sanctum (nang-rten) of the gsas-mkhar, which encases cloth, seeds, precious objects, armor, weapons, musical instruments, animal teeth and claws, etc. as supports and offerings to the enshrined deity. See Tucci 1966, pp. 188, 189.

The description of the traits and functions of the lha mkhar and gsas mkhar in Pt 1052 match those attached to such shrines today. The physical form of the ancient constructions in the text is unknown, but their names and description indicate that they were house-like or castle-like, a characterization still commonly articulated in Tibet. These metaphors express the tabernacular qualities of the structures. Like those noted in Pt 1052, many active shrines were raised on mountain slopes, passes and summits, acting as abodes for a host of elemental, ancestral and territorial spirits. The popular practice of erecting horns of wild ungulates (e.g., wild yak, blue sheep, argali sheep, deer, etc.) on contemporary shrines is also recorded in Pt 1052. Moreover, there is a corresponding function of protecting the countryside in both the text and present-day tradition.

The concerned Pt 1052 passage is illegible in places adding to its abstruseness.* It relates that on a mountain of the lha (lha rǐ), a majestic mountain (lhun po), the lha mkhar is a palace (brang), which seems to have the cylindrical shape of a milk pail (zo zo). This is followed by another line of six syllables that appears to indicate that this structure is for summoning (spran = sbran) spirit servants. The next line of the text is fragmentary, but refers to the protector (srungs ma) connected to the lha mkhar. On the sunny side of a mountain (dags rǐ) and the obscured side (rmun-po) is the gsas mkhar with horns on top (ru dangs rtse). It is described as the mansion (rmang brang) on a snow mountain. The protector of the land (sa srung) connected to this shrine wields thunder (’brug) and lightning (glog). The inner sanctum (thugs snying) of the lha mkhar and gsas mkkhar is the support (rten) for the stable basis (gzhung brtan) of this divination (mo), which presages in any situation (ji la btab) a reliable sign (bzang ngo).†

The first occurrence of lha as well as lhun, khab, dangs (= dang), and rten are designated uncertain readings, while da and bzhu are incomplete syllables in the Romanized Tibetan transcript of the text available to me for consultation. See: http://otdo.aa.tufs.ac.jp/archives.cgi?p=Pt_1052

See lns. 103–108. This passage exhibits a parallelism in terms of vocabulary and composition, consisting of built-in redundancies that lend it lyrical qualities. This parallelism is regularly seen in Old Tibetan non-Buddhist mytho-ritual texts. Some other divinatory procedures expounded in Pt 1052 conclude with a “reliable outcome” (bzang rab bo). The characterization of the probably small shrine constructions as a ‘palace’ and ‘mansion’ is metaphorical in nature, relaying that they are attractive to the deities of propitiation and suitable for their occupation. In the Old Tibetan Chronicle (lns. 271, 272), the lha ri is beseeched to protect the g.yang (good fortune potential of individuals, clans and communities and that of their surroundings, properties and interests): lha rǐ ni g.yang skyong shig /. For a discussion of the narrative setting of this line, see Walter 2010, pp. 231, 271 (n. 29). The prime function of good fortune bestowal attributed to the lha rǐ in the Old Tibetan Chronicle was retained in the Tibetan cult of divine mountains that has come down to the present day.

An account of the origins of a major class of stepped shrines called lha-rten or rten-mkhar is found in a ritual text for the Bon tutelary deity Ge khod,* who is thought to have been the chief god of the Zhang zhung kingdom.† This text of the Ge khod ritual cycle, Ge khod lha la rten mkhar gzugs pa, is recorded as being composed in the distant past in a cave on the mountain of Ru thog (Ge khod mkhar lung).‡ A castle active in the Protohistoric period is situated in an eponymous locale at the foot of this mountain.§ Despite containing what appears to be archaic lore, the Ge khod text is a Lamaist work composed well after 1000 CE. The main component (rta/gta’) of the lha rten shrine is its central axis, which pierces five graduated platforms representing the five elements (space, air, fire, water and earth). This shrine or tabernacle is erected for a large spectrum of indigenous deities of the binary sky and earth known collectively as lha-srin.

This article uses the proper noun ‘Bon’ to refer to the G.yung drung (Yungdrung) Bon religion, a Lamaist faith not unlike Tibetan Buddhism. Where appropriate, the common noun ‘bon’ is employed as a label for older non-Buddhist Tibetan religious customs and traditions. On questions related to the application of Bon/bon in an archaeological nomenclature for the Western Tibetan Plateau (includes Byang thang, Stod, Spiti, Ladakh (La dwags), Baltistan (Sbal te), Dolpo (Dol po), Mustang (Smos thang, etc.), see Bellezza in press-a.

For a detailed description of this account see Bellezza 2008, pp. 252–258.

For the incense offering ritual and origins myth included in this text, see Bellezza 2005, pp. 446–448. For an invocatory ritual carried out at a gsas mkhar shrine, see ibid., pp. 334–340.

On this castle, Ge khod mkhar lung, a large all-stone corbelled complex of ruined buildings overlooking an erstwhile agricultural valley, see Bellezza 2014a, pp. 138–144.

According to the origins myth (cho rabs) appended to the Ge khod ritual text, the lha rten was established to enhance the political and military standing of the kingdom of Zhang zhung. The myth centers around King Khri men of Zhang zhung, the holder of a crown with bird horns of iron (lcags kyi bya ru).* He is said to have practiced esoteric bomb-making rites known as the red this and to have raised a lha rten for the worship of Ge khod and his circle of 360 subsidiary deities, in order to defeat an invading army led by King Khri to rgyal ba of Stag gzigs (generally understood to be a country northwest of Tibet).

According to Martin (1991: 134–136) and Vitali (1996: 162) the bya ru crown of Zhang zhung was probably adopted from a Sassanian royal symbol around the 5th century CE. However, more recently obtained textual and archaeological evidence I have assembled indicates that the Sassanians came on the scene too late to have influenced the adoption of the bya ru motif in Tibet. The oldest textual references to bird horns (byu ru and khyung ru) occurs in an archaic funerary ritual context in Old Tibetan literature, often as part of an equestrian psychopomp headdress (Bellezza 2008: 506, 507, 509, 522; Bellezza 2013, pp. 69 (n. 90), 207, 230–232). Although these manuscripts from Dunhuang were composed in the Early Historic period, the origins of the ritual practices they contain are attributed explicitly to a much earlier era. In the archaeological record, the use of headdresses on sacrificial horses can be traced back to the Scytho-Siberians of the Iron Age (Bellezza 2008: 544–557; Bellezza 2016a, figs. 29–31). Moreover, there are figures of griffins with crests and horn-like ears among horse trappings used in burial rites of the Scytho-Siberian. On these figures, see Cheremisin 2009. Horned eagles form a class of subjects in Upper Tibetan and Spitian rock art of Iron Age, Protohistoric and Early Historic antiquity. For this rock art, see Bellezza in press-a; 2017d, pp. 26–30; 2015a; 2013d; 2012a; 2008, pp. 172 (fig. 303), 175 (fig. 310); 2002, pp. 134, 140, 217 (fig. XI-18c), 221 (fig. (XI-26c), 234 (figs. XI-4e, XI-5e). On early horned anthropomorphic rock art in Upper Tibet, see Bellezza 2014d. These findings encourage us to see the bya ru as a deeply embedded Tibetan cultural emblem. On the significance of the bya ru in headgear and chortens of the Bon tradition, see Norbu 2009, p. 132; Bellezza 2008, pp. 442–447.

The mythological and religious constitution of Bon texts purporting to describe the origins of stepped shrines, like those associated with Ge khod and the important group of harm-reduction rituals called mkha’ klong gsang mdos, obscures their historicity. Yet, the tales being told are interlaced with ancient lore. How this might accord with the actual symbolic and ritual functions of early stepped shrines remains difficult to ascertain. Nevertheless, the pre-Buddhist antiquity of these structures is confirmed by rock art in Upper Tibet, some of which can be assigned to the Protohistoric period. Stepped shrines however are not represented in earlier rock art of the region. Rock art sites with content attributed to the Late Bronze Age and Iron Age (such as Ri rgyal and Sgog ra) are devoid of these subjects. If there are ritual structures or tabernacles in Iron Age rock art of Upper Tibet they are very uncommon and of a different form (see infra, Part II, specimen 9h).

Pl.3. The remains of a group of ancient shrines situated on the erstwhile island of Do dril bu, east side of Bkra ri gnam mtsho/Kre de’i mtsho, Phyogs bcu (Mtsho chen county). Bkra ri gnam mtsho is one of the largest lakes on the Byang thang. Note the overlapping upper courses of masonry and small niches in the two larger specimens. These shrines must have once contained holy objects. Stripped to the bare rock core and bereft of their upper sections, the original height and configuration of these constructions is in question. They are found in association with what was an extensive settlement of all-stone corbelled residences and are best dated to a period predating the 9th century CE. On Do dril bu, see Bellezza 2001, pp. 237–243; 2014c.

Pl 4. A ruined shrine complex on the island of Mtsho gling gi do, Da rog mtsho, ’Brong pa tsho dgu (’Brong pa county). This island in the Yundrung Bon and local oral tradition is identified with ancient bon po practitioners, the most famous of which was Snang bzher lod po (8th century CE). The shrine complex appears to predate the 9th century CE. Note the niche in the shrine on the far left and the plinth on which the specimen on the far right sits. In the center is an adobe construction. Also note the all-stone corbelled residential structures in the background. On Mtsho gling gi do, see Bellezza 2014b, pp. 462–469; 1999, pp. 67, 79–81.

Pl. 5. A ruined stepped shrine or chorten situated in the mouth of a shallow cave on the north shore of Da rog mtsho, Mtsho chen county. On the rectangular base are the vestiges of three rounded vases (bum pa). As other specimens in the vicinity indicate, the stone substructure was enrobed in mud plaster painted with red ochre and white (probably calcium based) pigments. The sheltered location helps to explain the relatively intact condition of the structure. Such monuments at Da rog mtsho are attributed by local residents to the ancient bon po. For rock art analogues of this shrine, see infra, Part II, Group XI. Given their peculiar design and bon associations, these shrines were probably built prior to the 11th or 12th century CE.

Just as important to an archaeological investigation of stepped shrines as rock art are the actual remains of small cubic monuments found in much of Upper Tibet.* Some of these stone and mud constructions accompany all-stone corbelled fortresses and other types of large residences, presumably as ritual or ceremonial extensions of the facilities. These types of structures are relatively common, probably going some way to explaining the prominent place that facsimiles hold in the rock art of Upper Tibet.

For this archaeological evidence, see Bellezza 2002, pp. 123, 124; 2008, pp. 141–145; 2014a, passim; 2014b, pp. 521–523; 2014c. Among my findings is what appears to be a stone shrine or tabernacle installed inside a partially intact corbelled building at the Ge khod mkhar lung citadel. Radiocarbon dating of a small round of wood recovered from this stronghold yielded a calibrated date of ca. 150 BCE to 100 CE, indicating that the site was active in that period. See Bellezza 2008, p. 142 (fig. 230); 2014a, pp. 138–140.

Pl. 6. A remarkably well-preserved shrine made of specially hardened clay (and stone?) at Glang chen brag khung, Smad pa (Shan rtsa county).

In a highly remote red limestone formation called Glang chen brag khung (Elephant Grotto) there is a circular cave around 12 meters in diameter with four separate entrances. This southern-aspect cave is said to be the heart of a self-formed (rang byon) elephant. Its trunk is equated with a spur at the end of the formation. A natural archway around 4 m in height near the cave is likened to the mouth of the elephant. Resting on a pile of rocks in the center of the cave are the remains of a primitive shrine draped with prayer flags. The original extent of this ritual structure could not be determined. Currently it is 47 cm high and 57 cm wide. On its surface are traces of red ochre and a white pigment that once covered the construction, mimicking the bichrome treatment of some stepped shrine and chorten rock art in the eastern Byang thang. The shrine consists of a prominent hemispherical vase surmounting a square base with an overhanging upper tier. The hemispherical vase appears to be composed of neat courses of small bricks, the interstices between them skillfully filled and inset to produce a seamless structure. The element on the top of the vase is not recognizable.

The cave of Glang chen brag khung is maintained as a cult site by the monks of Gzims phug monastery and other local residents. However, no historical lore seems to be connected to the shrine. Rock art in the vicinity of the cave and monastery has a non-Buddhist orientation, such as the counterclockwise swastika painted in red ochre at the base of the archway.* The presence of this rock art as well as archaic cave shelters and the unusual nature of the religious construction inside Glang chen brag khung suggest that it too was made by non-Buddhists or bon po. If this is indeed true, the shrine predates the late 11th century CE. Buddhists consolidated their control of the region with the foundation of Gzims phug monastery in 1095 CE (Bellezza 2014a: 446, 447).

This ancient cultural icon is deeply embedded in the figurative traditions of the region. In the Iron Age and Protohistoric periods, swastikas in rock art face indiscriminately in both directions (although counterclockwise varieties predominate). Beginning in the Early Historic period the orientation of swastikas assumed sectarian connotations. Nevertheless, there are exceptions to the usual directions of swastikas in Upper Tibetan rock art. Even in the contemporary period, there are instances of anticlockwise swastikas depicted in Upper Tibetan folk art made by Buddhists. On swastikas in the rock art of Upper Tibet, see Bellezza 2016c.

Pl. 7. A stepped stone and adobe structure inside Lower East VII cave, Gnas kun bzang, Dang chung, Nag tshang. It was painted with red and white mineral pigments. This structure was installed near the rear of the cave and only its front face is stepped. Also see Bellezza 2008, p. 142; 2001, p. 113.

In addition to rock art representations and actual elementary stepped shrines, there are replicas in stone and metal.* Tiny stepped shrines made of a copper alloy and worn or installed as talismans have come to my attention.† These diminutive copies reproduce key features of rock art variants treated in this work.‡ Like much stepped shrine rock art in Upper Tibet, some stone and metal models can be assigned broadly to the Protohistoric period or Early Historic period.

In the 1990s, the late Shang Nyima, a Tibetan art connoisseur, showed me a remarkable elementary stepped shrine made of a ferrous (meteoric?) metal roughly 50 cm in height. It appeared to be of considerable age. The current whereabouts of this object are unknown to me.

There is also a common class of copper alloy (thog lcags) amulets consisting of a ram’s head surmounted by a chorten, some of which represent early forms of the monument. These amulets date from ca. 700 to 1500 CE. For examples, see Bellezza 1998, pp. 56 (fig. 42), 60 (fig. 58); John 2006, p. 130; Lin Tung-kuang 2003, p. 90.

On Tibetan copper alloy talismans and stepped shrine rock art, see Bellezza 2004, image sets, 15, 16; 2001, p. 344 (fig. 10.50).

Pl. 8. A small stepped shrine made of stone, which is hollow inside (presumably, it once served as a depository for sacred substances). This object has four main structural components: square base, five graduated tiers, small rounded vase and pinched finial or crown. It is being held by the individual who unearthed it from his village in southwestern Tibet. This photograph was taken in 2001. The object appears to be of much age, and has the same form as depicted in stepped shrines of Group X from the region’s rock art (see infra).

Pl. 9a. Stepped shrine (2.6 cm in height) of the thog lcags class of Tibetan talismans. Lhasa marketplace. Its face is engraved with an X and above it is a curvilinear design resembling water. There are two holes on the sides of the object for stringing to a cord. This object is comparable to rock art in Group X.

Pl. 9b. A small three-dimensional copper alloy Tibetan talisman (1.9 cm in height). Private collection. With its broad tiered body and bulbous top, this object is analogous to Upper Tibetan rock art shrines shown below in Groups II and III.

Pl. 9c. A copper alloy facsimile of a shrine with a banded body simulating different levels and a tricuspidate finial (4.4 cm in height). Lhasa marketplace. This thog lcags is comparable to rock art specimens in Group VIII.

Pl. 9d. A copper alloy chorten surmounting a ram’s head (4.7 cm in height). Private collection. An attachment loop for stringing the object was molded on the rear. The three-pronged finial of the stepped shrine is reminiscent of rock art variants in Group XIII. The finial on the talisman is not very pronounced possibly due to practical design matters concerning vulnerability to breakage.

Pl. 10. The upper plate of a golden death mask discovered in Gu ge (Rtsa mda’ county) by Li Linhui. Photo courtesy of Li Linhui. Note the three stepped structures each with three tiers and teardrop-shaped upper stage. There is a wild ungulate at the base of each stepped form. This mask appears to be related to the gser zhal (golden visage) of the archaic and Bon funerary traditions. This ritual instrument functioned as a receptacle for the consciousness of the deceased during evocation rites.

The dating of a plate metal object ornamented with what appear to be stepped shrines is on firmer ground than other portable objects described above, all of which are devoid of an archaeological context. A two-piece gilt funerary mask was discovered in a tomb at the Chu thags/Chu ’thag burial site, Gu ge (Rtsa mda’ county), by a Chinese archaeologist named Li Linhui, in 2009.* Like the stepped shrine made of stone (pl. 8), this funerary mask is linked to far western Tibet, the same region in which many of the rock art varieties are located.† In addition to a lower plate consisting of a repousse anthropomorphic face, a rectangular upper plate is ornamented with repousse birds, wild ungulates, trees and three tiered figures with bulbous tops (pl. 10). The latter subject looks remarkably like stepped shrines in the rock art of Upper Tibet (see especially infra, groups III and VII). Indications from archaeologists working at Chu thags, as well as my own comparative analysis of golden masks found in western Tibet and neighboring Himalayan regions, suggest that the Chu thags specimen is around 2000 years old. It belongs to an earlier phase in the use of such objects in the funerary rites of western Tibet and neighboring regions. This mask confirms that stepped shrines (or something identical in appearance) do indeed predate the Buddhist era, and were used ritually in regions that were part of the Zhang zhung kingdom no later than the early 7th century CE, just as the Bon textual tradition maintains.

On this discovery, see Yu Fei 2015; Chen 2016; Chinese Archaeology Writer 2015. The principal archaeologist working at Chu-thags, Tong Tao, has informed me (in personal communication) that a book about recently excavated tombs in Gu ge is under preparation.

For a discussion of this funerary mask, see Bellezza 2013a, pp. 157–159; 2013b; 2013c. On the use of golden visages (gser zhal) in the Bon funerary tradition, see Bellezza 2008, pp. 396, 412, 422.


Bon and Buddhism have assembled doctrinally coherent but highly idealized bodies of literature pertaining to the Tibetan chorten. There are widely divergent historical perspectives in both religions pertaining to the development of this monument. Their unique sectarian makeup is epitomized in the different personalities and geographic origins each religion espouses. Bon literature claims that chortens were established in Tibet in the remote past by their founder, Gshen rab mi bo of ’Ol mo lung ring and other primal saints. For their part, Buddhists trace the chorten to the ancient stūpa of India. On the surface of it, there is little room to reconcile these two traditions with one another. However, as we shall see there is actually more historical correspondence between them than the sectarianized views espoused  by the two religions admit.

The long biography of Gshen rab, Dri med gzi brjid (composed ca. 1400 CE), describes four archetypal chortens of Bon in the mythical land of ’Ol mo lung ring, which are oriented in the intermediate directions around a central mountain of nine superimposed swastikas, G.yung drung dgu brtsegs (Denwood: 1978: 179, 180). In support of the conventional Bon historical perspective, the text Nyi sgron (14th century CE), quoted in the Legs bshad mdzod (1920s), states that chortens as well as Bon literature and temples (gsas khang) thrived in Zhang zhung before they appeared in Central Tibet (Karmay 1972: 22).* Kun ’bum (late 12th or 13th century CE) avers that when Gshen rab descended to Zhang zhung he constructed 1008 chortens to subdue negative chthonic forces of the four cardinal directions, as well as building temples and castles (Bellezza 2001: 21, 22).† The Bon historical text Bsgrags pa rin chen gling grags (11th century CE) is even bolder in its assertion, holding that a crystal chorten with bird horns was brought by Gshen rab to an Indian king and established as an object of worship in India (Martin 1991: 121).

The belief that chortens existed in the time of the Zhang Zhung kingdom has filtered down to the oral tradition. For example, at the base of an archaic citadel called Gyang pa’i gtsug rdzong, at Dang ra g.yu mtsho, there is a ruined structure local sources claim was a chorten founded at the same time as the citadel. See Bellezza 2001, p. 96.

A sketch of the traditional historical development of the Bon chorten is furnished by ’Gyur med rab rgyas (2001: 78–80). It can be paraphrased as follows: It was Ston pa gshen rab who was the architect and disseminator of the chorten of great efficacy. In the world of humans, on the peak of Ri rab, the gsas mkhar (shrine) Lha rtse dgung nam (Peak of the Divinities in Space) was established, and eventually others such as the ’Gro ba rab ’dren mchod rten (Superlative Guide of Living Beings Chorten) with the kho ma ru rings (long horns of the wild yak), G.yung drung bkod legs [mchod rten] (Swastika Chorten of Good Auspices) and Gsas mkhar sgyogs pa rtse dgu. Gshen rab mi bo devised the Lo paṇ shel gyi mchod rten (Crystal Chorten of Translators and Scholars) to purify the defilements of the body, speech and mind (lus ngag yid gsum) of Gto bu dod de.* Its three heads/crowns (dbu) represent the Lha gshen srid gsum [Gshen lha ’od dkar, primal deity of light; Ston pa gshen rab, bringer of salvation to humanity; and Srid pa sangs po ’bum khri, the appointer of existence]. These are the kinds of chortens found in the Bon tradition of Zhang zhung. The methods of construction and beautiful forms of these chortens became the tradition in Tibet some 3900 years ago. Terms sometimes used interchangeably for chortens founded by Ston pa gshen rab include gsas khang, gsas mkhar and gzhal yas, among others (Denwood 1978: 178).

According to Gzer mig, Gto bu dod de was a king that Gshen rab freed from heaven. The chorten built for him was decorated with animals of the sky, earth and intermediate realm. See Kværne 1997, pp. 9, 10.

Another mythic account set in very early times concerning the chorten is found in a text describing the origins of the Khyung clan entitled, Dbra dkar khyung po’i gdung (Bellezza 2008: 289 [n. 275]). According to this text, the three progenitor brothers of the Khyung clan founded 108 temples and crystal chortens for worship. It is written in the Ti se’i dkar chag (mid-19th century) that a Zhang zhung sage named Ye shes rgyal ba taught the Bon doctrine to a huge assembly beside a chorten named Ghan dha chen po at the foot of Mount Pos ri ngad ldan (Bellezza 2001: 60; Vitali 2008: 406). The Bon historical texts Bsgrags pa gling grags and Mdzod sgra ’grel (12th century CE) assert that at each of the 37 assembly centers of the ancient bon po (’du gnas) a chorten, as well monuments such as tombs (bang so), pillars (rdo ring) and border markers (sa rtags), were established.* The claim that there were Buddhist-like chortens long before Buddhism took root in Tibet is also made in a Bon canonical text describing the architectural character of the palace temples (pho brang gzhal yas) of Zhang zhung, which includes a chorten in the heart of its design (Bellezza 2008: 295).

For references to the monuments of the 37 ’du gnas, see Uebach 1999, p. 269; Bellezza 2001, p. 85; 2008, pp. 291, 292. According to Uebach (ibid., 271), the formation of the Bon tradition of the 37 assembly places can be probably attributed to the 8th or 9th century CE but also preserve lore of prehistoric origins. Though still not comprehensively surveyed, archaeological evidence I have assembled supports Uebach’s assessment that the origins of these sites should be sought in the distance past. For example, Mang mkhar (near Lha rtse) boasts all-stone corbelled residential structures, the only site reported in Central Tibet with archaic architecture characteristic of Upper Tibet. On the archaeology of the Mang mkhar lcags ’phrang ’du gnas, see Bellezza 2010a. Other ’du gnas sites with early residential traces include Gnam mtsho do (now called Bkra shis do) and Rta phu dron lhas (at base of Byang la bod la in the Mus valley).

Pl. 11. Modern Bon chortens of the G.yung drung bkod legs mchod rten complex, ’Om bu, Dang ra g.yu mtsho. Photograph taken in 1994.

As we shall see, forked and pronged finials crowning stepped shrines are common motifs in Upper Tibetan rock art. These may well correspond with the bird horns of the textual tradition. The existence of horned raptors in the rock art of Upper Tibet supports the presumption that similarly shaped finials on early stepped shrine represent a horned motif. Furthermore, the occurrence of bya ru and khyung ru in archaic funerary ritual literature of the Early Historic period may indicate that chortens of that era carried an analogous motif. However, there is no foolproof method to confirm this, as the two disparate bodies of evidence (rock art and textual) do not lend themselves to a stringent cross-referencing exercise.

Turning to the Buddhist chorten, Tibetan literature has India as its historical source, the cultural wellspring of that religion. As pertains to the design and construction of chortens, Pema Dorjee (2001) concludes that Tibetan variants conform with plans and proportions prescribed in Sanskrit literature, but with architectonic adaptations peculiar to Tibet.* Pema Dorjee (ibid.) finds that norms standardizing proportions in construction were established in the 13th and 14th centuries CE,† based on codification of designs by prominent personalities such as Bu ston, Stag tshang lo tsa ba and Sprul sku ’phreng kha ba, etc.‡

Likewise, the indigenization of the stupa is seen in China, Korea, Thailand, Burma, Sri Lanka, Japan, etc. For a photographic survey of stupas, including many different Tibetan ones (pp. 242–300), see Cook and Yeshe De Project 1997. For another extensive photographic survey of Tibetan chortens, see Dkon mchog bstan ’dzin and ’Phrin las rgyal mtshan 2007.

For a discussion on Buddhist chortens in western Tibet and their derivation from Indian prototypes through the codification of Sanskrit texts, see Tucci 1932, pp. 15–21.

For a study of handbooks on chorten symbolism, design and construction and their religious use and benefits, see Cayton 1996. She includes translations of relevant materials in Sde srid sangs rgyas rgya mtsho’s Baidur dkar po and Baidurya g.ya sel, Rong tha blo bzang dam chos rgya mtsho’s Thig gi lag len du ma gsol bar bshad bzo rig, and Bu ston rin chen grub’s Byang chub chen po’i mchod rten gyi tsad and Mchod rten la mchod pa byas pa’i phan yon.

That the Buddhist tradition of chorten design in Upper Tibet was not fully standardized until well after the second diffusion is reflected in archaeological and architectural evidence. Many ruined and a few rebuilt chortens dating to ca. 1000–1250 CE abound in western Tibet (especially in Gu ge). With their wide terraced bases, squat vases and broad spires, these chortens are precursors to conventional types of the later period. This monumental evidence accords well with chorten-making handbooks, corroborating that the prevailing Tibetan Buddhist view concerning chortens being faithful copies of Indian prototypes was not fully articulated until the 13th century CE.*

In western Tibet, Buddhist chortens (including eight major types) adhering closely to standardized forms are depicted in the murals of Lha khang dmar po (ca. 14th century CE) in Rtsa pa rang, as documented by Giuseppe Tucci in 1933. See Tucci 1936: 150, 151, figs. CXXXV, CXXXVI. For a color photograph of these eight types of Buddhist chortens, see ’Phrin las mthar phyin 2001, p. 152. Although these chortens are well developed conventional forms, they retain the pyramidal spire (and prominent flowing banners) of the earlier architectural tradition.

Votive plaques made of clay (tsha tsha) also collected in Gu ge bear images of Indian chortens as well as inscriptions in the proto- Śarādā script.* Some of these tsha tsha presumably date to the 11th century CE and the second diffusion of Buddhism (bstan pa phyi dar) in western Tibet. That such types of chortens and companion inscriptions are not found in western Tibetan or Spitian rock art is an indication that the old tradition of painting and carving on natural stone surfaces did not thrive after the Buddhist renaissance at the turn of the second millennium CE.

Many of these plaques are illustrated in Tucci 1932. For discussion on early tsha tsha bearing the impression of chortens, also see Pema Dorjee, pp. 110–114.

It appears that in the late 10th century CE rock art figuration came to be considered old-fashioned, as artistic energies and religious sentiments were diverted into the creation of frescoes and sculptures in the Buddhist temples and grottos of western Tibet and Spiti. The absence of evidence for Indian canonical forms of chortens in these regions prior to the second diffusion of Buddhism also suggests that this religion did not have a particularly vigorous grip on the region.* On the other hand, there is scant evidence for the systematized Bon religion having gained a place in western Tibet or the western Himalaya (cf. Tucci 1932: 67–70). In any case, as this religion postdates the 10th century CE, there was little historical scope for its chorten iconography to be expressed in Upper Tibetan rock art, a moribund tradition by the time of the second diffusion.

Petech (1980: 85) maintains that after the collapse of the Tibetan monarchy in ca. 842 CE, Zhang zhung (western Tibet) was left to its own devices. This was certainly true in a cultural and religious sense as much as it is a political statement. The first biography of Lo tsa ba Rin chen bzang po and the edicts of Lha bla ma ye shes ’od describe the existence of a well-established archaic religious culture often referred to as bon prior to the 11th century CE. See Karmay 2014, pp. 3–17; Bellezza 2015b. An interesting tale of the subjugation of the old religious order involving the cutting down of an ancient juniper tree on the north shore of Mtsho ma pham and the establishment of Kha char monastery in southwestern Tibet is found in the text Nges don ’brug sgra. See Vitali 1996, pp. 553, 554. For a study of the religious complexion of western Tibet and other regions of Upper Tibet in the Early Historic period and Vestigial period based on rock inscriptions, see Bellezza forthcoming.

Pl. 12. A highly eroded stone depicting the Rigs gsum mgon po (the bodhisattvas Spyan ras gzigs, ’Jam dpal dbyangs and Phyag na rdo rje) chortens, Spiti. On stylistic grounds, these carvings appear to date to the time of the second diffusion of Buddhism, ca. 11th or 12th century CE.

Pl. 13. Buddhist chorten at Ma nams, Gu ge, 11th century CE. Chortens of this period in western Tibet still retained a few archaic features associated with chortens in the Upper Tibetan rock art of the Early Historic period.

Pl. 14. Modern Rigs gsum mgon po chortens painted in colors characteristic of the three Bodhisattvas, Za rang, Gu ge. These chortens were designed and built in accordance with standard iconometric principles. Rigs gsum mgon po chortens are a popular sacred accompaniment to houses in far western Tibet, particularly in agricultural villages of Ru thog.

The archaeological, architectural and art historical evidence considered evinces a historical relationship between elementary shrines, non-Buddhist chortens and Buddhist chortens in Upper Tibet. It can be inferred that inputs for the design of Buddhist variants were, in part, derived from pre-existing religious monuments of the region (other cultural influences will be discussed in Part III). These historical links are supported by other aspects of the Tibetan literary tradition.

In his collected works, Dge ’dun chos ’phel asserts that Buddhist stepped chortens owe part of their architectural character to the receptacles of dmu deities (dmu mkhar) of the bon tradition (after Bellezza 2014a: 448 [n. 1]).* One of the oldest textual references to the influence of native architectural traditions on Buddhist chortens occurs in a Tibetan consecration text composed sometime before 1039 CE by the Indian saint Atiśa, who notes bird horns in his explication of chorten symbolism (Martin 1991: 122, 123). According to an Upper Tibetan oral tradition, set later in the 11th century CE, two ruined chortens located near the sacred hot springs at Stag lung, Nag tshang (Shan rtsa county) were constructed by the Bon po Khro tshang ’brug lha and the Buddhist Pha dam pa sangs rgyas working as a team (Bellezza 2001: 133).†

Dge ’dun chos ’phel (ibid.) also acknowledges a historical connection between the Buddhist chorten and the lha rten shrine. While he does not specify how precisely they are related, ’Gyur med rab rgyas (2001: 78) suggests that chortens are historically linked to ubiquitous rten mkhar, ’bum pa (tiered monument for localized and personalized deities), rdo ’bum (heaps of stones or more elaborate structures used for ritual purposes), rdo mchod (stones from sacred or pristine places used as offerings), and mtho cog (= mtho lcog; shrines on rooftops and other high places for localized and personalized deities). ’Phrin las rgyal mtshan (2007: 4) holds that chortens were influenced by the bon tradition of establishing supports for the worship and placation of btsan, yul lha and gzhi bdag, a tradition that remains widespread and unbroken to the present day.

More coercive forms of interplay between the makers of chortens of the two religions are also recorded. Legs bshad mdzod reports that during the persecution of Bon in the late 8th century CE, some of its chortens were destroyed or repainted and given Buddhist names (Karmay 1972: 85). This kind of re-appropriation does not appear to have been uncommon. For instance, a Bon chorten maintained at Bya do monastery (Dpal mgon County) until 1959 is said to have been emptied of its original contents and re-consecrated with Buddhist sacred articles (Bellezza 2001: 59 [n. 30]). A similar process of re-appropriation may possibly have occurred with the Gser-gdung bcu-gsum, 13 chortens at the base of the south face of Ti-se. On this site, see Bellezza 2013i.

A more revealing account of the debt owed by the Buddhist chorten to earlier bon traditions comes from the Bka’ chems ka khol ma (mid-11th century CE; see Sørensen and Hazod: 2005: 155–158). This text relates that the A ya bon po (sacrificial priests) Co tshug and Sra tshug transformed themselves into a falcon that slew the klu bdud demon of a lake, from which a roaring sound came (giving rise to the name of the nearby Buddhist temple: Khra ’brug). The lake was infilled by the bon po and King Srong btsan sgam po erected Mchod rten rtse lnga ba on the site.* According to the Khra ’brug gnas bshad, Mchod rten dbu lnga (sic) was built by the king to atone for the killing of the klu, becoming the first chorten erected in Tibet (ibid., 93–95).

According to Rgyal po’i bka’ thang yig, during the funeral of King Srong bstan sgam po (ca. 651 CE), a chorten was carved in stone (rdo la mchod rten bskos) at the tomb site (Haarh 1969: 352, 353).

Given the prevailing archaic religious complexion of western Tibet and Spiti before the late 10th century CE,* it is not surprising that there are few local oral traditions ascribing the construction of Buddhist chortens to the Early Historic period.† Chortens in the rock art of Upper Tibet datable to the Early Historic period are of a decidedly non-Buddhist or syncretic character, often with horn-like finials and counterclockwise swastikas and other motifs in close proximity, which subsequently became associated with the Bon religion (there is little evidence for the existence of a monolithic Buddhist-like religion calling itself bon prior to the late 10th century CE). ‡ Many of the rock art chortens of the Early Historic period, therefore, are attributable to local cults still practicing a body of archaic religious traditions. These cults, or more correctly, their ritual and mythological aspects, were in some instances known as bon, as various Old Tibetan manuscripts of the Early Historic period indicate. Nevertheless, the relatively elaborate chortens of the Early Historic period strongly suggest that certain religious groups at that time were imbibing Buddhist artistic customs. This implies that cults of a syncretic composition were taking root in Upper Tibet at that time. There may even have been Upper Tibetans who regarded themselves as full-fledged Buddhists in the Early Historic period. Buddhist temples in the region of that time must have attracted local interest and patronage. Be that as it may, epigraphic and monumental evidence for a Buddhist presence in western Tibet before the 10th century CE is quite minimal.§

The historical picture in Central Tibet and Khams was different. King Srong btsan sgam po (reigned ca. 629–650 CE) is supposed to have built Mchod rten dkar po on top of Dmar po ri, in Lhasa, (’Gyur med rab rgyas 2001: 81). This king is also said to have established a chorten of five points (rtse lnga) to destroy a savage klu with five serpent heads in the same period as the founding of the temples of the mtha’ ’dul (suppressing the limbs) and yang ’dul (suppressing the extremities) of the supine demoness (srin mo; ibid.). For a review of a few traditional accounts of chorten building in Central Tibet in the Imperial period, see Seegers 2015, pp. 55, 58, 59. In a mythic account of the establishment of the first chorten in Tibet, it is stated that the translator Li thi se presented a golden chorten fallen from the sky to the 28th Tibetan king, Lha tho ri gnyan btsan, when he was 60 years of age (’Gyur med rab rgyas 2001: 79). According to a local oral tradition, two chorten in Skye gum do (Khams) were founded by Princess Wencheng on her way to Lhasa to marry King Srong btsan sgam po (Seegers 2015). For a couple of Tibetan references to chortens in the Dunhuang manuscripts, see Stein 2010, pp. 48, 51.

There is the pair of large stepped shrines in Pu hreng known as Gu ru ’bum pa attributed to Gu ru rin po che, but these are simple stepped structures.

In a different cultural context, the stupa and swastika were fundamental geometric symbols upon which the Buddhist tradition of settlement planning was organized in the Kathmandu valley. See Pant and Funo 2007.

Archaeological evidence for Buddhism in western Tibet during the Early Historic period appears to extend to a few all-stone corbelled temples constructed in a ‘transitional style’: Ri’u dgon pa, Spyil bu dgon pa and Gzims phug. At Spyil bu dgon pa there are stone slabs inscribed with primitive ma ṇi mantras assignable to the Early Historic period. See Bellezza 2014a, pp. 311–322, 340–344; Bellezza 2015c. In the textual arena, there are references to famous Buddhist temples in western Tibet of the Imperial-period such as Dpal rgyas and Pra dum/Skra-bdun of the yang ’dul tradition. On these monuments, see Vitali 1996, pp. 267, 276. However, Buddhist rock inscriptions predating the 10th century CE are uncommon in Upper Tibet and these are largely confined to a scattering of mantras. In contrast to Ladakh and northern Pakistan, there are very few chorten dedications or references to administrative or military personnel in Upper Tibetan epigraphy dating to the Early Historic period. This evidence indicates that until the second diffusion of Buddhism, non-Buddhist cultural traditions held sway in Upper Tibet. On the ancient rock inscriptions of Upper Tibet, see Bellezza forthcoming.


Next Month: The panoply of Upper Tibet stepped shrines!

Previous Post

Next Post