Download the PDF version of this article

Facsimiles of the Earliest Ritual Architecture on the Plateau
A Comprehensive Survey of Stepped Shrines in the Rock Art of Upper Tibet
by John Vincent Bellezza

Text and images © the author except where otherwise noted

Published May 2020

(click on the small image for full screen image with captions.)


This article presents findings from a comprehensive survey of stepped shrines in the rock art of Upper Tibet. This region of some 600,000 km² is traditionally known as the Byang thang and Stod. Stepped shrines represented in this rock art are of two major types: elementary and complex tiered structures. The latter architectural type is commonly called chorten (mchod rten) in Tibetan.[2]

Both chortens and simpler stepped shrines in Upper Tibetan rock art depict monuments with religious, 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 religious tradition also characterized the historical scene in antiquity. As is well known, Tibetan cultural and religious traditions underwent fundamental developments in the last millennium, altering or effacing earlier intellectual and material patterns to some degree or another.

The Tibetan chorten, especially that of the Buddhists, is conventionally seen as the equivalent of the Indian stūpa (hereinafter: stupa).[3] The celebrated functions of the Tibetan chorten can be summarized as follows:

  1. Model of the Buddha, particularly his mind (thugs)
  2. Monument commemorating life events of the Buddha and deeds of other great Buddhist teachers
  3. 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[4]
  8. Receptacle for sacred substances

The most salient design feature of stepped shrines in the rock art of Upper Tibet is the stacking of one rectangle or square upon another to create tiered arrays. These strata are typically graduated with each successive quadrate reduced in size.[5] In more intricate examples, there are two or more sets of layered rectangles and squares interspersed or flanked by 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.[6]

 The stepped shrine rock art sites of Upper Tibet

Stepped shrines were documented by the author on twenty different expeditions to Upper Tibet over the course of eighteen years (1995–2013).[7] Rock art shrines were recorded at 37 sites, spanning the region from the southeastern corner of Gnam mtsho to the western portion of Ru thog near the Indian frontier.[8] Four-fifths of these sites are clustered in just three areas of Upper Tibet: Gnam mtsho, Shan rtsa and Ru thog. Save for one exception, pictographic stepped shrines are only situated at sites in the eastern Byang thang. A total of 255 chortens were documented in Upper Tibetan rock art (several uncategorized chortens notwithstanding).[9]

Upper Tibet in the wider region

Map of rock art sites in Upper Tibet with stepped shrines

Site Number

Site Name




Brag gdong East

Ru thog



Ser mdzod rdo ring

Ru thog



Rwa ’brog ’phrang

Ru thog



Sngo sog

Ru thog



Rno bo g.yang rdo

Ru thog



Nag khung rdo ring

Ru thog



Chu mkhar gyam sgrub phug

Ru thog



Mchod rten sbug sna kha

Ru thog



Mtha’ kham pa ri

Ru thog



Ris mo gdong

Ru thog



Glog phug mkhar

Ru thog



Gyam rag East

Ru thog



Gna’ ba lung

Ru thog



Skabs ren spungs ri

Ru thog



Rdzong chung

Ru thog



Sgog ra

Ru thog



Ru thog rdzong

Ru thog



Brag gyam




Rdu ru can

Rtsa mda’



Gser sgam

Rtsa mda’



Ri rgyal

Sger rtse



Brag khung mdzes po

Nyi ma



Dar chung

Nyi ma



Bshag bsangs

Nyi ma



Rdo ’khor phug pa

Shan rtsa



Rta ra dmar lding

Shan rtsa



Chu ro

Shan rtsa



Skyid sgrom sgo gru bzhi

Shan rtsa



Sgar gsol brag phug

Shan rtsa



Slop dpon phug

Shan rtsa



Rta mchog ngang pa rdo

Dpal mgon



Ra ma do

Dpal mgon



Lug do

Dpal mgon



Stong shong phug

Dpal mgon



Se mo do South

Dpal mgon



Bkra shis do chung

’Dam gzhung



Bkra shis do chen

’Dam gzhung


A chronology of stepped shrine rock art

The styles of stepped shrines in Upper Tibetan rock art cannot be used in isolation to ascertain the age of individual examples. That is mainly because elementary forms were used alongside more complex varieties in Upper Tibet, rendering analysis based on seriation unreliable. Moreover, dates for the introduction of simple stepped shrines and chortens in the region’s rock art have not been determined with any surety.

Generally speaking, chortens were introduced in Tibet with the advent of Buddhism and the cosmopolitan cultural influences that accompanied it. However, as in Indus Kohistan and probably Ladakh as well, it cannot be ruled out that chorten rock art reached Upper Tibet before Indian Buddhism took root there in the 7th and 8th centuries CE. Elementary stepped shrines not only appear in rock art but also as actual stone ruins in Upper Tibet. Significant numbers of these structures appear to predate the 7th century CE but precise dates for them have not yet been formulated. 

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. Utilizing various kinds of data, I have devised a relative chronology for Upper Tibetan rock art, serving as a provisional means for dating rock art in broad terms (subject to verification whenever appropriate technologies and methods become available). The major criteria employed in the chronological analysis of rock art, including stepped shrines, can be outlined as follows:
  1. Stylistic and thematic categorization of motifs, subjects and scenes
  2. Appraisal of the general characteristics of the contents of sites
  3. Analysis of complementary textual and ethnographic materials
  4. Cross-cultural comparisons with rock art in other regions
  5. Application of collateral archaeological data
  6. Gauging environmental conditions depicted in rock art
  7. Assessment of the techniques of carving and painting
  8. Examination of the degree of erosion and re-patination of carvings and the degree of browning and ablation of pigments.
  9. Determination of the placement of superimposed images (palimpsests)

Relying on a chronological analysis of rock art based on the above criteria, the following chronological regime for the stepped shrine rock art of Upper Tibet (applicable to certain other regions on the Tibetan Plateau as well) has been developed:

Protohistoric Period (ca. 100 BCE to 600 CE)
Early Historic period (ca. 600–1000 CE)
Vestigial period (ca. 1000–1300 CE)[10]

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

 Categorization criteria for stepped shrine rock art

The 232 stepped shrines (petroglyphs and pictographs) forming the basis of this article are divided into twenty different groups according to appearance. These groups are organized beginning with the simplest forms and progressing to the most intricate, charting the stylistic variability of these ritual monuments in rock art. There is a continuum in the designs of stepped shrines, making it difficult in certain cases to distinguish a chorten from other kinds of tabernacles. Indeed, both types of monuments are closely related culturally and historically. The field survey carried out indicates that the Tibetan chorten is as much an architectural product of pre-existing indigenous monuments as it is the Indian stupa.
Stepped shrines in each of the 20 categories are not necessarily affiliated with one another chronologically or functionally. Some simple stepped shrines seem to postdate more complex chortens and some are actually chortens rendered in a rudimentary manner. Conversely, certain elaborate stepped shrines appear to be rten mkhar or other kinds of non-chorten depictions.

Pl. 11. This quintessential Bon chorten possesses a pyramidal spire (’khor lo), bya ru bya gri (horns of the bird, sword of the bird) finial (tog) and lotus petal base (sa gdan). These traits as well as flowing banners (dar thag; not shown) are vestigial in nature, and can be traced back to chorten rock art in Upper Tibet and other regions of the western fringe of the Tibetan Plateau.[11]

The expanded standard terminology for various parts of chortens in the Bon and Buddhist religions (used here somewhat interchangeably) is as follows:
Base or plinth (sa ’dzin)
Graduated tiers or platforms (bang rim)
Square central section (pho brang)
Base of vase (bum gdan)
Vase (bum pa)
Base of spire (bre)
Spire (’khor lo)
Horizontal extension above spire (char khebs)
Finial (tog)
Banners (dar thag)

Typological illustrations and data on the stepped shrines of Upper Tibetan rock art

In this work one example from each 20 category of stepped shrines is illustrated to furnish a comprehensive review of this rock art genre in Upper Tibet. A total of 232 stepped shrines belong to the typology that follows, which consists of 20 different groups. Typological drawings of all of them have been prepared.[12]

Group I: elementary two-tiered shrines (three specimens)

This is the most elementary category of stepped shrines in the rock art of Upper Tibet. Its specimens were made utilizing different pecking techniques and vary considerably in terms of wear and re-patination.

Fig. 1a. Petroglyph, Brag gdong East (Ru thog), 27 cm.

Group II: elementary three-tiered shrines (ten specimens)

This category of stepped shrines is diverse stylistically and in the techniques of production. Both petroglyphs and pictographs are represented. Some among them may be incomplete renditions of shrines. 

Fig. 2c. Petroglyph, Brag gdong East (Ru thog), 9 cm.

Group III: elementary four-tiered shrines (11 specimens)

With one exception, the specimens in this category consist of three graduated tiers topped by a rounded extension or vase. On the basis of physical condition, some of these are among the oldest stepped shrines in the rock art of Upper Tibet.

Fig. 3d. Petroglyph, Rdu ru can (Rtsa mda’), 6 cm.

Group IV: elementary multi-tiered shrines (23 specimens)

Some of the shrines in this category are very similar to Group III, only somewhat more complex in form. Others depict a larger series of tiers or platforms, simulating towering structures. What appear to be older (Protohistoric period) examples of stepped shrines in Upper Tibet are found in this group and include specimens such as 4c.[13]

Fig. 4c. Petroglyph, Rdzong chung (Ru thog), 14 cm.

Group V: elementary segmented shrines (nine specimens)

This category of stepped shrines is characterized by carved and painted vertical lines that create segmented layers. Other than this distinguishing trait, this is a diverse group stylistically and technically.

Fig. 5a. Petroglyph, Rwa ’brog ’phrang (Ru thog), 26 cm. / Note: Ma ṇi mantras were partly carved on top of the stepped shrine.

Group VI: elementary shrines with spatulate and crescent-shaped finials (four specimens)

This category of stepped shrines is distinguished by its unique style finials. Overall site characteristics, as well as stylistic, technical, physical features, suggest that specimen 6c dates to the Protohistoric period.

Fig. 6c. Petroglyph, Rdu ru can (Rtsa mda’), 15 cm.

Group VII: elementary shrines with multi-foliate finial (two specimens)

This category of stepped shrines with their finials are unique in Upper Tibet. They are attributable to the Protohistoric period, and have a more advanced patina then examples of Early-Historic-period rock art and inscriptions at the same site.[14]

Fig. 7a. Petroglyph, Rdu ru can (Rtsa mda’), 31 cm. / Note: Carved over a chariot that is more heavily re-patinated.

Group VIII: elementary shrines with tricuspidate finial (four specimens)

This unusual category of what appear to be pictographic variations of stepped shrines may straddle the Protohistoric and Early Historic periods.[15] The prominent three-pronged finials are particularly noteworthy. Specimen 8a can be assigned to the Early Historic period.

Fig. 8a. Pictograph (red ochre), Bkra shis do chung (’Dam gzhung), 14 cm. / Note: Accompanied by five-pointed star and counterclockwise swastika painted by the same hand. See Bellezza 2000b; 2002, pp. 127, 202.

Group IX: idiosyncratic elementary shrines (12 specimens)

This category contains miscellaneous figures, some of which cannot be positively identified as shrines.

Fig. 9d. Pictograph (red ochre), Bkra shis do chung (’Dam gzhung), 26 cm. / Note: On this specimen, see Bellezza 2002, pp. 130, 206.

Group X: shrines with small bulbous upper section (22 specimens)

This fairly complex category of stepped shrines appears to straddle the Protohistoric and Early Historic periods. These examples are mostly distinguished by small rounded vases and some have simple masts. Most stepped shrines in this group are the product of non-Buddhist religious traditions.

Fig. 10d. Pictograph (red ochre), Rdo ’khor phug (Shan rtsa).

Group XI: twin shrines sharing common base (three specimens)

This category is comprised of the depiction of twin tiered structures sharing a common base. Most probably all three examples are depictions of non-Buddhist religious monuments.

Fig. 11a. Petroglyph, Brag gdong East (Ru thog), 32 cm. / Note: On this stepped shrine and adjoining specimens, see Bellezza 2000b.

Group XII: simple style chortens (seven specimens)

This category is comprised of rudimentary depictions of chortens, all of which (except for one) were carved at the site of Brag gyam. This site boasts the most complex rock art chortens found in western Tibet (see Group XIX). Most of the chortens in this category have a lower section comprised of five graduated tiers surmounted by a small rounded vase, but do not have spires or finials. This group of chortens are assigned to the Early Historic period. Its religious orientation is not very clear; some examples may have been made by practitioners of syncretic traditions.  

Fig. 12f. Petroglyph, Brag gyam (Sgar), 29 cm.

Group XIII: chortens with forked finial (20 specimens)

This category is distinguished by stepped shrines with horn-like finials or crowns, a lower section of four or five graduated tiers and a small vase. The two spikes or branches of most forked finials curve gracefully and sometimes flare out at the ends. They resemble the horns of wild ungulates in some Upper Tibetan rock art. Stepped shrines in this group can be ascribed to the Protohistoric period and Early Historic period.

Fig. 13j. Petroglyph, Ser mdzod rdo ring (Ru thog). / Note: Carved on an archaic funerary pillar.

Group XIV: chortens with cross-piece spire (15 specimens)

This category is characterized by simply designed stepped shrines with a mast bisected by a series of short lines. It is otherwise a fairly varied group technically and stylistically. Two exceptions aside, the specimens in this group are all pictographs. They all appear to be chortens and date to the Early Historic period and possibly the Vestigial period. They are primarily or entirely non-Buddhist renditions, as their designs and surrounding rock art and inscriptions indicate. 

Fig. 14c. Pictograph (red ochre), Bkra shis do chen (’Dam gzhung), 26 cm.

Group XV: Giant pictographic shrines (five specimens)

The five pictographs in this category are between 300 cm and 500 cm in height. They are situated in a row on an east-facing rock face, on the south side of the limestone escarpment at Bkra shis do chung, Gnam mtsho.[16] They rise above a shrine called Sri gcod bum pa. They are all heavily eroded, precluding an assessment of fine stylistic details. These appear to be non-Buddhist portrayals of stepped shrines, and probably date to the Early Historic period or Vestigial period.

Fig. 15a. Pictograph (red ochre), Bkra shis do chung (’Dam gzhung). / Note: This is the most northerly specimen at the site.

Group XVI: flat-topped chortens (19 specimens)

This category is characterized by stepped shrines terminating at the base of what commonly underlies a spire. Many of these chortens are attributable to the Early Historic period and vestigial period. Most examples appear to be non-Buddhist or syncretic portrayals of religious monuments.

Fig. 16f. Petroglyph, Ser mdzod rdo ring (Ru thog). / Note: Carved on an archaic funerary pillar.

Group XVII: chortens with simple spires (12 specimens)

This diverse category of chortens is marked by elementary spires of various kinds. Many of these chortens were the workmanship of those adhering to non-Buddhist or syncretic traditions, and can be dated to the Early Historic period.

Fig. 17e. Petroglyph, Brag gdong East (Ru thog), 24 cm. / Note: The part of the rock surface with the right half of carving is now missing.

Group XVIII: intricate non-Buddhist chortens (23 specimens)

This category contains the most elaborate group of non-Buddhist or bon (in a generic sense) chortens found in Upper Tibet. Among them are many idiosyncratic examples. Most specimens are datable to the Early Historic period, a time in which non-Buddhist Upper Tibetans adopted the Buddhist chorten for their own ritualistic and cultic purposes. Some pictographs from the eastern Byang thang in this group are polychrome paintings.

Fig. 18u. Pictograph (red ochre), Bkra shis do chung (’Dam gzhung), 76 cm. / Note: In close proximity is an extensive red ochre inscription. Unfortunately, it is highly worn and no longer legible, although a few individual letters can be made out. On this chorten and the inscription, see Bellezza 2000a, pp. 40, 41; 2000b.

Group XIX: intricate chortens of Brag gyam (25 specimens)
All except two of the chortens in this category come from a single, site, Brag gyam (Rock Formation Shelter). A total of 68 specimens were documented here in 2004. It appears that the elaborate chortens of Brag gyam were produced over a relatively short span of time, coinciding with the Early Historic period (and perhaps extending into Vestigial period as well). Some of the elementary stepped shrines of the site may be even older.

There are no clear indications that any of the elaborate stepped shrines of Brag gyam were made as patently Buddhist monuments. Rather, their creation may have to do with syncretistic traditions, derived from indigenous religious customs combined with Buddhist artistic inspiration emanating from Ladakh.

The shrines of Brag gyam range in style from simple stepped structures to elaborate chortens. Some were crudely carved while others were more finely engraved. They have weathered to assume a polished effect on the limestone surface and in certain instances they have acquired a reddish discoloration. Three inscriptions are found at the site (see Bellezza 2008, p. 187 [n. 193]), which in terms of wear and weathering, closely match proximate chortens.

Fig. 19t. Petroglyph, Brag gyam (Sgar), 234 cm. / Note: The middle of the chorten was engraved over the carving of an almost life-sized anthropomorphic figure with outstretched arms. This figure appears to belong to the same time frame as the chorten and may possibly represent a religious subjugation theme like that in the sri gnon rite. The chorten appears to have one or two eyes in its vase.[17]

Group XX: Buddhist chortens (seven specimens)

This category of chortens contains Buddhist versions of the monument. They represent the terminus of the great rock art tradition in Upper Tibet, carvings and paintings that had at least a 2000-year-old legacy by that time. In the Vestigial period relatively few rock carvings and painting were still being produced in Upper Tibet. Specimen 20a appears to be a portrayal of a byang chub chorten, one of the most common types in Buddhist Tibet of the last millennium. It dates to the Vestigial period.

Fig. 20a. Pictograph (red ochre), Bkra shis do chen (’Dam gzhung), 29 cm / Note: Situated in sTag lung phug. On this chorten, see Bellezza 2002, pp. 129, 130.

 Cross-cultural influences in the stepped shrines of Upper Tibet

There are chortens in the rock art of Ladakh, Baltistan, Indus Kohistan, Gilgit, and even in Chitral and Wakhan with design traits recalling those in Upper Tibet, including forked finials (some with a pointed or rounded element in the middle), pyramidal and cigar-shaped spires, squat vases, and wide terraced bases. Accompanying Tibetan inscriptions brand some of these northwestern chortens Buddhist monuments. There are also more primitive types of stepped shrines in the large mountainous tract northwest of Upper Tibet, suggesting that non-Buddhist cultural features diffused throughout the wider region in the Protohistoric and Early Historic periods. 

Not surprisingly, the region with tiered shrines most closely matching types from Upper Tibet is Spiti, adjoining its western flank.[18] Traditionally, Spiti enjoyed excellent geographic access with Upper Tibet (Gu ge Chu mur ti region), including in winter months. Many of the more than 30 early stepped shrines I have documented in the rock art of Spiti are stylistically and technically analogous to Upper Tibetan Groups II, III, IV, X, and XVIII. Given parallels in style to Upper Tibetan variants and the lack of Tibetan inscriptions, most Spitian examples appear to have a non-Buddhist or possibly a syncretic religious identity.

A comprehensive survey and typology of stepped shrines from Ladakh has not yet been published.[19] Nonetheless, through a number of academic publications covering the topic as well as unpublished materials,[20] it is clear that the elementary stepped shrines and chortens of Ladakh can be both compared and contrasted with those from Upper Tibet. Some are accompanied by inscriptions in the Tibetan, Kharoṣthī, Brāhmī and proto-Śarādā scripts.[21] The rock art of Ladakh is almost entirely comprised of petroglyphs, and taken as a whole, constitutes a distinctive body of cultural materials. Many chortens occurring in Ladakh are stylistically remote from those in Upper Tibet.

As already mentioned, Brag gyam is the only site in Upper Tibet with an unambiguous chorten dedication inscription. On the other hand, there are many dozens of comparable examples from sundry sites in Ladakh. Despite minor differences, the sheer numbers in Ladakh strongly suggest that it was the center of gravity for these types of rock art chortens, which spread south slightly beyond the current border of western Tibet. Both the Brag gyam and Ladakh specimens are datable to the Early Historic period and Vestigial period, the chorten being a cosmopolitan religious monument of those times. The engine for their transmission appears to have been the Tibetan military conquest of Ladakh and the concomitant expansion of chorten rock art south and eastwards.

The filtering of rock art chortens south and eastwards into western Tibet in the Imperial period is also seen in a single bell-shaped specimen found at Ru thog rdzong. Like Brag gyam, its existence there documents a Buddhist presence, something hardly seen in Byang thang tracts to the east. Two such bell-shaped chortens have been documented in Ladakh,[22] while there are many examples in northern Pakistan.[23] In this case, the artistic and cultural springboard appears to be northern Pakistan, with Ladakh serving as an intermediary link in the diffusion of the bell-shaped chorten to Ru thog.[24]

Like Ladakh, there is a huge supply of rock art in northern regions of Pakistan.[25] This rock art includes a great many stupas as well as inscriptions in Kharoṣthī, Brāhmī and proto-Śarādā, Sogdian, Bactrian, Tibetan, Chinese, and Hebrew, as well as a rich variety of Buddhist iconographical images (Buddhas, bodhisattvas, etc.; Thewalt 1985: 779).[26] According to Thewalt (ibid.), the oldest depictions of stupas resemble those in Sancī and Bharhut in Central India as well as Gandhāran examples.

Of special interest to this study are unrefined stupas and other kinds of simple stepped shrines from northern Pakistan that share stylistic affinities with Upper Tibetan rock art. Jettmar (1982: 29) attributes the most roughly rendered stupas of Indus Kohistan to the possible existence of a folk religion that co-existed with Buddhism. Jettmar (1985) opines that some of these carvings may represent “Bon” sacred mountain symbols. By the use of the term Bon, Jettmar subscribes to the existence of widely distributed religious and cultural traditions at variance with Buddhism, but he does not consider what degree of institutional and ideological coherence these traditions may have had. Some irregularly executed stupas in northern Pakistan are found in conjunction with crude Brāhmī inscriptions, while others are connected to those in proto-Śarādā (Jettmar 1982: 18, 20). These stupas and inscriptions appear to be datable to the 5th to 8th centuries CE (Hauptmann 2007: cf. 34).[27]

The stupas of Indus Kohistan that most resemble Upper Tibetan varieties have highly particularized designs and are rudimentary in form and execution. These stupas are characterized by a minimum of graduated tiers, a small vase, and a three pronged finial often squared in form and resembling a trident. Although these kinds of stepped shrines recall those of Upper Tibet, they are more closely related to specimens in other areas of northern Pakistan and Ladakh. Most importantly, these so-called stupas appear to portray constructions belonging to peoples practicing non-Buddhist or hybridized religious traditions.

According to scholarly estimates, the earliest stepped shrines in northern Pakistan rock art belong to the first three centuries of the Common Era. On epigraphic and artistic grounds, the same can probably be said for those from Ladakh. Some elementary shrines in the rock art of Upper Tibet are attributable to the Protohistoric period and may be of comparable age. Moreover, a repoussé golden funerary mask discovered in a tomb in Chu ’thag, Gu ge with depictions of three stepped shrines dates to circa 100 BCE to 100 CE (an early phase of the Protohistoric period).

By virtue of the oldest elementary stepped shrines in the rock art of northern Pakistan, Ladakh and Upper Tibet seemingly of equal antiquity, it is not yet possible to ascertain which region may have been first to produce them. Nor is it known how stepped shrines spread in the greater region. Unless they were developed independently by each people who adopted them, an unlikely prospect, interactive processes were involved in conveying information (architectural, ritual, mythological, social, etc.) about them to various peoples and places. This realm of interactivity may have stretched from Gnam mtsho in the east to Gilgit and beyond in the west.

It can be concluded that one layer of cross-cultural communications in the greater region was distinctly non-Buddhist in makeup. This is evidenced in elementary stepped shrines found in Upper Tibet, Ladakh, Indus Kohistan, Gilgit and Swat. This corpus of cognate forms and styles in rock art serves to define a common artistic and ideological ground of significant depth. For some researchers this signifies bon. I am in agreement, provided this term is applied generically to denote customs and traditions that circulated throughout the wider region in contradistinction to Buddhism. However, for the pre-Imperial era there is little archaeological or art historical evidence for a singular religion extending from Gnam mtsho to Gilgit and Wakhan. To the contrary, it would appear that there were many languages and self-identified tribes and cultural formations in this gigantic territory, as in more recent times. It is hardly possible that they all possessed the exact same religion, as evidenced in the diverse archaeological records of the wider region. Rather, it appears that there existed a religious substratum or body of cultic materials that transcended the intellectual and geographic bounds of individual cultures, infiltrating and rearranging them to some extent. Hence, the spread of stepped shrines is best seen as a trans-cultural phenomenon.

This trans-cultural legacy of similar artistic and ideological elements is embodied not only in elementary stepped shrines, but in other types of archetypal rock art subjects of Upper Tibet, Ladakh and northern Pakistan such as swastikas, sunbursts, archers, and the hunting of large wild ungulates like ibexes, wild sheep and wild yaks.

The origins of elementary shrines in rock art remains to be determined. They may ultimately have been inspired by the Indian stupa or instead they may have sprung up indigenously. Or perhaps an entirely different source must be considered? In any case, these facsimiles of ritual constructions were well established in Upper Tibet, Spiti, Ladakh and northern Pakistan ideologically and artistically speaking; so much so that they may have been taken for granted as a natural or unremarked cultural prop of all peoples who embraced them.

The origins of stupa rock art seem more straightforward than those of the elementary stepped shrines. The largest and most sophisticated concentration of engraved stupas is in northern Pakistan.[28] This hothouse of stupa rock art surely had an impact on Ladakh, as analogous stupa forms and inscriptions in a variety of languages indicate.[29] It is from Ladakh that carved chorten art most probably entered largely non-Buddhist western Tibet and Spiti in the Imperial period.[30] On the other hand, the painted chortens of Gnam mtsho may be of Central Tibetan inspiration.

Pre-existing elementary stepped shrine rock art in Ladakh, Upper Tibet and Spiti had a potent influence on how chortens in those regions were conceived and articulated in the Imperial period. It may have been during this time when certain artistic elements of Upper Tibet rock art like forked and pronged finials found their way to the northwestern frontier of the Tibetan empire. The carriers of any such Upper Tibetan artistic traits were adherents of non-Buddhist (bon in a generic sense) or syncretic religious traditions, the troops, administrators and traders who reached northern Pakistan and Wakhan.


Anagarika Govinda, Lama. 1976. Pyscho-cosmic Symbolism of the Buddhist Stūpa. Emeryville: Dharma Publishing.

Beer, Robert. 1999. The Encyclopedia of Tibetan Symbols and Motifs. London: Serindia Publications.

Bellezza, John V. 2016. “A Review of a Recent Scientific Article: “Earliest tea as evidence for one branch of the Silk Road across the Tibetan Plateau””; “Art and Shelter: An ancient settlement in far western Tibet”, in Flight of the Khyung, April 2016. Tibet Archaeology:
Forthcoming. Drawn and Written in Stone: An Inventory of Stepped Shrines and Early Rock Inscriptions in Upper Tibet (ca. 100 BCE to 1400 CE). British Archaeological Reports International Series. Oxford: BAR Publishing.

_____2014a. Antiquities of Zhang Zhung: A Comprehensive Inventory of Pre-Buddhist Sites on the Tibetan Upland, Residential Monuments, vol. 1. Miscellaneous Series – 28. Sarnath: Central University of Tibetan Studies. Online version, 2011: Tibetan & Himalayan Library (

_____2014b. Antiquities of Zhang Zhung: A Comprehensive Inventory of Pre-Buddhist Sites on the Tibetan Upland, Ceremonial Monuments, vol. 2. Miscellaneous Series – 29. Sarnath: Central University of Tibetan Studies. Online version, 2011: Tibetan & Himalayan Library (

_____2014c. “Exquisitely Carved Chortens from Far Western Tibet”, in Flight of the Khyung, July 2014. Tibet Archaeology:

_____2014d. The Dawn of Tibet: The Ancient Civilization on the Roof of the World. Lanham: Rowman & Littlefield.

_____2013a. “Of a Different Form: Tabernacles of the ancients”, in Flight of the Khyung, February 2013. Tibet Archaeology:

_____2013b. “A unique collection of ancient ceremonial monuments both drawn and carved”, in Flight of the Khyung, August 2013. Tibet Archaeology:

_____2008. Zhang Zhung: Foundations of Civilization in Tibet. A Historical and Ethnoarchaeological Study of the Monuments, Rock Art, Texts and Oral Tradition of the Ancient Tibetan Upland. Philosophisch-Historische Klasse Denkschriften, vol. 368. Wien: Österreichischen Akademie der Wissenschaften. 

_____2002. Antiquities of Upper Tibet: An Inventory of Pre-Buddhist Archaeological Sites on the High Plateau, Delhi: Adroit.

_____2001. Antiquities of Northern Tibet: Archaeological Discoveries on the High Plateau, Delhi: Adroit.

_____2000a. “Bon Rock Paintings at gNam mtsho: Glimpses of the Ancient Religion of Northern Tibet”, in Rock Art Research, vol. 17 (no. 1), pp. 35–55. Melbourne.

_____2000b. “Images of Lost Civilization: The Ancient Rock Art of Upper Tibet”, in Asian Art Online Journal.

Bruneau, Laurianne. 2010. “Le Ladakh (état de Jammu et Cachemire, Inde) de l’Âge du Bronze à l’introduction du Bouddhisme: Une étude de l’art rupestre, vol. 3. Répertoire des pétroglyphes du Ladakh”. Ph.D. dissertation, Université Paris I-Panthéon-Sorbonne.

_____2007. “L’architecture bouddhique dans la vallée du Haut Indus (Pakistan): Un essai de typologie des représentations rupestres de stūpa”, in Arts Asiatiques, vol. 62, pp. 63–75.

Bruneau, Laurianne. and Vernier, Martin. 2015. MAFIL. Mission Archéologique Franco-Indienne au Ladakh Report: Fieldwork 2015.

Cayton, Lori J. 1998. “The Buddhist Stūpa: Its History, Dimensions and Symbolism”, doctoral dissertation, University of Wisconsin. Ann Arbor: UMI.

Combaz, Gisbert. 1933. L’évolution du stūpa en Asie. Bruxelles: L’institut Belge des Hautes Études Chinoises.

Davidson, Ronald, M. 2002. “Hidden Realms and Pure Abodes: Central Asian Buddhism as Frontier Religion in the Literature of India, Nepal, and Tibet”, in Journal of the Institute of Buddhist Studies, vol. 4 (no. 3), pp. 153–181.

De Marco, Giuseppe. 1987. “The Stūpa as a Funerary Monument: New Iconographical Evidence”, in East And West, vol. 37 (nos. 1–4), pp. 191–246. Rome: IsMEO.  

Denwood, Philip. 1980. “Temple and Rock Inscriptions at Alchi”, in The Cultural Heritage of Ladakh. Zanskar and the Cave Temples of Ladakh. (by D. L. Snellgrove and T. Skorupski), vol. 2, pp. 117–163. Warminster: Aris & Phillips.

_____1978. “Stupas of the Tibetan Bonpos”, in The Stupa (ed. A. Dallapiccola), pp. 175–186. Heidelberg.

Harvey, Peter. 1984. “The Symbolism of the Early Stupa”, in The Journal of the International Association of Buddhist Studies, vol. 7 (no. 2), pp. 67–93.

Hauptmann, Harald. 2007. “Pre-Islamic Heritage in the Northern Areas of Pakistan”, in Karakoram: Hidden Treasures in the Northern Areas of Pakistan (ed. S. Bianca), pp. 21–40. Turin: Umberto Allemandi & Co.

Internews Report. 2011. “Dam to Drown Gilgit-Baltistan’s Vandalised Rock Art”, in Maverick:
Accessed May 24, 2016:

Jettmar, Karl. 1985. “Non-Buddhist traditions in the petroglyphs of the Indus valley”, in South Asian Archaeology. in South Asian Archaeology. 1983. Papers from the Seventh International Conference of the Association of South Asian Archaeologists in Western Europe, Held in the Musées Royaux d'Art et d'Histoire, Brussels (eds. J. Schotsmans and M. Taddei), pp. 751–778. Naples: Istituto Universitario Orientale Dipartimento di Studi Asiatici, Series Minor XXIII.

_____1982.  Rock carvings and Inscriptions in the Northern Areas of Pakistan. Islamabad: Institute of Folk Heritage.

Kozicz, Gerald. 2014. “The Chorten (mChod rten) with the Secret Chamber near Nyarma”, in Art and Architecture in Ladakh: Cross-cultural Transmissions in the Himalayas and Karakoram (eds. E. Lo Bue and J. Bray), pp. 141– 158. Brill’s Tibetan Studies Library, vol. 35. Leiden: Brill.

Linrothe, Rob. 2003. “Rock Art of the Himalayas, Tibet and Central Asia”, in Himalayan Art Resources. Accessed May 3, 2016:

_____1999. “A Summer in the Field”, in Orientations, May 1999, pp. 57–67. Hong Kong.

Menyag Geshe Namdak Tsukphu. 1998. Mchod rten gyi dpe ris blo gsar mgul rgyan (Bon Stupa: Illustrations and Explanations). Solan: Tibeto Zhang-Zhung Research Centre.

Orofino, Giacomella. 1990. “A note on some Tibetan petroglyphs of the Ladakh area”, in East And West, vol. 40 (no. 1-4), pp. 173–200. Rome: IsMEO.

Pant, Sushila. 1976. The Origin and Development of Stūpa Architecture in India. Varanasi: Bharata Manisha.

’Phrin las mthar phyin (author of preface). 2001. Krung go’i bod ljongs mnga’ ris kyi gna’ bo’i ldebs ris bdams sgrig (A Selected Collection of Ngari Ancient Frescoes of Tibet, China). Lhasa: Bod ljongs mi dmangs dpe skrun khang.   

Ricca, Franco and Lo Bue, Erberto. 1993. The Great Stupa of Gyantse: A Complete Tibetan Pantheon of the Fifteenth Century. London: Serindia Publications.

Takeuchi, Tsuguhito 2012. “Old Tibetan Rock Inscriptions Near Alchi”, in Historical Developments in Tibetan Languages. Proceedings of the workshop B of the 17th Himalayan Languages (Kobe 6th-9th, September, 2011.   Symposium) (eds. T. Takeuchi and N. Hayashi). Journal of Research Institute, vol. 49, 2012. pp. 29–71. Kobe: Research Institute of Foreign Studies, Kobe University of Foreign Studies.

Tang Huisheng and Gao Zhiwei 2004. “Dating Analysis of Rock Art in the Qinghai-Tibetan Plateau”, in Rock Art Research, vol. 21 (no. 2), pp. 161–172. Melbourne.  

Thakur, Laxman, S. 2008. “Bonpos of the Western Himalayas”, in Journal of the British Association for South Asian Studies, vol. 24 (no. 1), pp. 27–36. London.

Thewalt, Volker. 2008. Stupas und verwandte Bauwerke in Felsbildern der Stationen Chilas I, Chilas II und Oshibat am Indus (Nordpakistan). Accessed online May 11, 2016:

 _____2000. Stupas, Tempel und verwandte Bauwerke aus drei Felsbildstationen bei Chilas am oberen Indus (Nordpakistan). Accessed online May 11, 2016:

_____1985. “Rockcarvings and Inscriptions along the Indus: The Buddhist Tradition”, in South Asian Archaeology. 1983. Papers from the Seventh International Conference of the Association of South Asian Archaeologists in Western Europe, Held in the Musées Royaux d'Art et d'Histoire, Brussels (eds. J. Schotsmans and M. Taddei), pp. 779–800. Naples: Istituto Universitario Orientale Dipartimento di Studi Asiatici, Series Minor XXIII.
Accessed online May 10, 2016:

Tucci, Giuseppe. 1932. Stupa: Art, Architectonics and Symbolism. Indo-Tibetica I (trans. U. M. Vesci). Roma. Reprint, 1988: Aditya Prakashan, Dehli. Sata Pitaka Series, vol. 347. 

Watt, Jeff. 2007. Bon: The Magic Word. The Indigenous Religion of Tibet (eds. S. G. Karmay and J. Watt). New York: Rubin Museum of Art.


1. The delivery of my lecture at the IATS XIV conference in Bergen, Norway (upon which this article is based) was made possible by a travel grant provided by the Lumbini International Research Institute (Lumbini). The writing of this article was enabled by a recurring grant from the Shelley & Donald Rubin Foundation (New York). My documentation of rock art in Tibet in more recent years was supported by grants conferred by the Unicorn Foundation (Atlanta), Tise Foundation (USA) and by Mr. Joseph Optiker (Berglen). Most of the black and white drawings of stepped shrines in this article were commissioned from the Tibetan artist Lingtsang Kalsang Dorjee and his atelier. Other drawings of stepped shrines in this work were kindly done by Rebecca C. Bellezza

2. 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.

3. 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 Combaz 1933; Pant 1976. Also see De Marco 1987; Thewalt 2000.

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

5. Stepped structures are of course known throughout the world as 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. 

6. The verticality of chortens is magnified by the central axis (srog shing) that runs along its entire height, its most fundamental element. According to an intercultural mythological perspective, the central axis 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).   

7. For earlier presentations of these findings, see Bellezza 2002, pp. 128, 129, 144, 145; 2014d, pp. 189–193; 2013b. 

8. I have surveyed a total of 76 sites in Upper Tibet with rock art, thus nearly half of the total number of sites have stepped shrines depicted. For a list (incomplete) of rock art sites 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.

9. There are also seven chortens, including six highly elaborate ones, dating to the Vestigial period, from the Rgyal la lding site (Gu ge) not included in this study. For these Buddhist chortens, see Bellezza 2014c. In 2015, a Chinese research team working under the direction of Zhang Jian Lin (Shaanxi Provincial Institute of Archaeology) discovered a small rock art site on the south bank of the Glang chen gtsang po (Sutlej river), a little southwest of Khyung lung, including a boulder with three elementary stepped shrines carved on it. Two of these specimens can be placed in Group II and the other one in Group IV of the stepped shrine typology I have devised.

10. On the dating of Tibetan rock art using a different set of methodological approaches, see Tang Huisheng and Gao Zhiwei 2004.

11. For an actual artistic rendering of a Bon chorten based on a similar schematic drawing, see Watt. 2007, p. 35 (fig. 20). This chorten forms the central theme of a thang ka from Dol po, dated to the 17th or 18th century CE (ibid.). For drawings of some of the 360 types of Bon chortens of the ‘Three Bodies of the Buddha’ (Sku gsum), intended for iconometric study and usage, see Menyag Geshe Namdak Tsukphu 1998. The preface of Tsukphu’s work (pp. x–xii) presents a heavily Buddhacized conception of these chortens. For a review of the form and functions of the Bon chorten, see Denwood 1978.

12. They will appear in Bellezza forthcoming, complete with extensive archaeological, cultural and historical comments. Additionally, there are 23 stepped shrines from the Brag gyam site not enumerated in this typology of 232 examples (they are largely repetitive).

13. For an artist’s conception of what these and other elementary shrines may have looked like as three-dimensional objects, see Bellezza 2014a, p. 24.

14. For this Early-Historic-period art and inscriptions, see Bellezza 2016.

15. For these four shrines, see Bellezza 2013a.

16. For these pictographs, see Bellezza 2001, pp. 209, 210, 339–341.

17. Two chortens in a mural painted inside a chamber of a chorten near Nyar ma, Ladakh, have an eye in the vase (Kozicz: 2014: 155). For an eye in the vase of an elaborately carved chorten in far western Tibet (Rgyal la lding, Gu ge), see Bellezza 2014e, figs. 19, 20.

18. On these depictions of non-Buddhist elementary shrines in Spiti, see Bellezza 2015a, figs. 47–70, Thakur 2008, pp. 30 (fig. 8), 31, 33 (fig. 15). Thakur (ibid.) characterizes these archaic representations as “bon po”. I have also previously used the decidedly generic term ‘bon’ to identify the religious orientation of analogous rock art chortens in Upper Tibet. See Bellezza 2002, pp. 128, 129.

19. This would be a huge undertaking. The art historian Rob Linrothe (1999: 60, 61) writes that he has seen, “hundreds – perhaps thousands – of petroglyphs of stupas all over Zangskar and Ladakh”. This contrasts with Upper Tibet (an area approximately ten times the size of Ladakh) where about 250 stepped shrines predating the 13th century have been documented in the rock art.

20. Thanks to the generosity and graciousness of my colleagues Laurianne Bruneau, Martin Vernier and Viraf Mehta, I have been supplied with hundreds of digital photographs of images of Ladakhi rock art. Among the images I have been given are several dozens of stepped shrines.

21. See Bruneau 2010, pl. 53; and numerous examples in Linrothe 2003; Takeuchi 2012; Orofino 1990; Denwood 1980.

22. These are located on the same stone surface at a site called Digar Kharpoche. See Bruneau and Vernier 2015, p. 18.

23. On bell-shaped chortens in northern Pakistan, see Bruneau’s (2007: 65) typological groupings, 1, 2, 5 and 7.

24. Denwood (2008: 8, 9) notes that Ru thog linked what he calls the Byang thang Corridor with the Shyok, Indus and Nubra valleys, which served as egress points into northern Pakistan. It is along such geographic lines that the bell-shaped chorten ultimately reached Ru thog rdzong. According to ’Phrin las mthar phyin (2001: 209), the citadel of Ru thog rdzong had three underground tunnels for water and hundreds of houses for commoners (mi ser), and was a trading hub for grains, salt, wool and goods moving between Mnga’ ris and Ladakh. In ancient times Mnga’ ris was on a kind of silk road (dar gos lam bu) (ibid.). On Ru thog rdzong, also see Bellezza 2001, pp. 102–104; 2014a, pp. 152, 153. On aspects of trade in western Tibet during the Protohistoric period, see Bellezza 2016.

25. Harald Hauptmann estimates that there are more than 50,000 carvings and 6000 inscriptions in northern Pakistan, the majority around Chilas. See Internews Report 2011.

26. Tibetan inscriptions are found in Shigar (Baltistan) and along the Gilgit river. See Hauptmann 2007, p. 35.

27. As regards the rock art stupas of Indus Kohistan, Thewalt (2008: 63, 64) notes that a precise chronology is not yet possible due to discrepancies in styles and the re-patination of the carvings. Based largely on epigraphic evidence, Thewalt (ibid., 63–65) divides the stupas into two separate phases: 1st and 2nd centuries CE, as seen in carvings at Chilas II and Chilas III; and 6th and 7th centuries CE, as represented in specimens from Chilas I and Thalpan I and other sites. Very few stupa carvings were made after the 10th or 11th century CE (ibid., 64), which is less true of Upper Tibet and Ladakh.

28. Matters concerning Buddhist sources for the stupa rock art of northern Pakistan is beyond the scope of this paper. Clearly major Buddhist centers were involved; viz., the Kushan empire, Gandhāra and later Kashmir. According to Davidson (2002 158), the Kaniṣka stupa in Peshawar, probably the most important and largest of its kind, was a religious focal point for Buddhists from Gandhāra, the Karakorum and the upper Indus valley, informing stupa figures in the rock art of northern Pakistan and stupa art farther afield.

29. Recently, a sample taken from the wooden axis pole of a ruined chorten at Tirisa (Nubra valley) was subject to radiocarbon analysis and yielded at calibrated date in the fifth and sixth centuries CE. See Bruneau and Vernier 2015, pp. 22, 23. Provided the axis of the Tirisa chorten is reflective of the date of establishment of this monument and its age is not biased by the ‘old wood effect’, it could confirm a pre-Imperial presence for the chorten in Ladakh.

30. I have found no epigraphic or artistic evidence indicating transfer of stupas from cis-Himalayan regions of Himachal Pradesh to Spiti and western Tibet, strengthening a northern vector of transmission.

Download the PDF version of this article