Śiva or Brahma? The Masque Court at the BN, Paris

© Koninklijke Brill NV, Leiden, 2013 DOI: 10.1163/15728536-13560303

Indo-Iranian Journal 56 (2013) 381–396 brill.com/iij

Śiva or Brahma?The “Masque Court” at the

Bibliothèque Nationale, Paris

Harry FalkFreie Universität [email protected]

AbstractAnumber of “masks” cast inmetal are known from theNorth-West, particularly Kashmir andHimachal Pradesh. They are not used as theatre masks, they lack openings for view or speech,and are found in temples fixed to awall.One, foundnear Peshawar inGandhara is conspicuousby its artistic beauty. It is inscribed on its lower rim; both script and style allow to date it in themiddle or late Gupta period. The inscription has been read and interpreted by G. Fussman,who assumed that some of his readings would need revision. A new reading is proposed anddiscussed here, as it admits of two rather divergent interpretations.

KeywordsMetal masks; Gandhara; Śaivism; Brahma; four-faced statues of gods; fourth face; beardedgods

Introduction

The hollow mask of a bearded face cast in metal found its way to Paris in 1834as a present to King Louis-Philippe, who forwarded it to the royal Cabinet desMédailles, now part of the Bibliothèque Nationale in Paris, where it rests to thisday. It was found near Peshawar, but if this was its original location we do notknow.Themask is 46cmhigh and consists of three registers, themiddle onebeingthe face with half-closed eyes; only the lower half of the pupil is outlined belowthe heavy lid.There is amoustache covering the upper lip; the short beard on thecheeks is expressed through wavy parallel lines.

In the upper register, the tightly combed hair is folded into a bun, held by thecoiled body of a snakewhose three heads stand right above this bun.The earlobesare prolonged from formerly wearing heavy ear-ornaments, the empty extendedearlobes touching the lower register. This consists of four concentric rims whichsupplant the neck. On the lowest of the four rims the letters of an inscription inGupta-Brāhmī can be seen.

http://brill.com/iij

mailto:[email protected]

382 Harry Falk / Indo-Iranian Journal 56 (2013) 381–396

Afirst short reportwas published by J. Prinsep,1 togetherwith a drawing of themask by General A. Court, lithographed by Prinsep2 himself. Although Courthad drawn the inscription (Fig. 1a) rather carefully, Prinsep involuntarily addedsome frills while copying with the result that the letters would “not form anyintelligent Sanskrit sentence”.

The full treatment it deserved was provided only by G. Fussman in four pub-lications.3 The first paper in the Revue de la Bibliothèque Nationale provides thebest photographs of the front, from the side, the inscription and from the back.The second paper covers the essence and many additional aspects extensively.Fussman’s texts from 1992 and 2003 provide short summaries in English andFrench. Errington in 1992 added some reflections on her own.

The piece is exceedingly beautiful. As Taddei4 remarked, “the direct sight ofit only allows us to appreciate its great quality to which no photographs are ableto do justice”. Despite its artistic qualities and singular features, it hardly findsmention in treatments of Śaivism. In fact, wemay ask if this deity is fully Śaiva atall, since no third eye is seen.The long earlobes are empty and clearly reminiscentof those of Śākyamuni, where they are testimony of his having given up theworldly riches he once enjoyed as a prince. The half-closed eyelids remind ofmanyBodhisattva statues fromGandhāra, as does themoustache.Thehair-bun issimilar to many hairdos of Bodhisattvas in Gandhāra, where, however, no snakereplaces the ribbon.

Metal masks as such are well-known, and never show a figure from the Bud-dhist pantheon. Similar masks come from the same area, i.e. the North-Westincluding Kashmir. Two have been published in 2003 by M. Ghose, one Śivafrom Himachal Pradesh by Larson,5 two more masks from Kashmir by Siud-mak,6 to name just the most recently published ones.

The Inscription

My purpose is mainly to reinspect the inscription on the piece, which was firstread by Fussman in 1990 and never subject to revision in the succeeding papers.No other scholar before him had cared or dared to look into thismost importantbut badly preserved legend.

1) Prinsep 1836: 484–485.2) Prinsep 1836: pl. XXVI.3) Fussman 1990, 1991, 1992, 2003.4) Taddei 1993: 348a.5) Larson 2004: 57.6) Siudmak 2005.

Harry Falk / Indo-Iranian Journal 56 (2013) 381–396 383

Figure 1. Inscribed lower rim of the metalmask from the Bibliothèque Nationale, Paris.


My inspection of the piece was highly facilitated when themask was at Lattes,near Montpellier in southern France, for the Conference De l’ Indus à l’Oxusin May 2003. The director, Michel Landes, allowed me to examine it. A firstresult was to see that the inscription consists of two sections. First, the wholetext was incised into the waxmodel. After casting, it must have become apparentthat the first two words were more or less illegible. Therefore eight letters at thebeginning of the line were cut into the metal with a sharp instrument on topof the traces of the original letters (Fig. 1c). The rest was kept as it was. There-cut letters are clear and deep, the original script, however, needs several tricksto disclose its secrets. Some letters are clear enough, but others need to be seenfrom all angles, or can at least be seen partly when lit from different angles. Thisway it soon became apparent that the published reading can be provided withalternatives.

A full study was granted at the Bibliothèque Nationale in April 2012 andsolved the reading of those characters which were still creating problems.

Fussman read and translated the legend7 as:

divodharma ya˙m cortarūna śrī haradevo graghegutparmatoka …

Ceci est le don pieux de Cortarūna. ⟨C’est⟩ le Seigneur Haradeva, ⟨aussi dit?⟩ Graghe-gutparmatoka,

or:8

This is a pious gift from Cortarūna; ⟨he is⟩ Śrī Haradeva, ⟨also called?⟩ Graghegutpar-matoka

He understood themask therefore to portray Śiva, who comes by various names,Hara being one of them.

My own reading has some Sanskritic terms for those inexplicable to Fussman:

divadharma ya˙m vartakekagra-sthira-divya-brahmegutpa-mastaka

˙h

Wehave to take into account that the inscription has been overwritten in its firstpart as said above.The letters in Fig. 1b showclearly legible lines in black, and cer-tain, but faint lines in grey; extremely shallow letter-grooves are shown in lightgrey. All forms have been taken from a number of close-up photographs, com-pared to pencil rubbings on paper done earlier. Whoever retraced the first eightletters, probably was not identical with the original scribe, maybe he was illiter-ate, as he just followed the faint traces of the original lines. He wrote the first va

7) Last in Fussman 2003: 355.8) Fussman 1992: 14.


as a loop with crossed ends and the second one as a loop with a serif. He retraceda ka as no writer probably would do, by writing one part of the crossbar whilegoing down and adding the right part later, instead of adding a horizontal stroketo a vertical one. The new ka is scratched slightly left of the old one so that wehave the remnant of the original by the side of its surrogate (Fig. 1d, e). The textwhich seems to constitute one long compoundneeds to be discussed part for partbefore we can venture a translation:

In divadharma ya˙m we meet with an introductory formula for a donation,

which is deyadharmo ’ya˙m in its earliest Sanskrit form and which was used origi-

nally inBuddhist contexts, but can also be found in the fourth century on objectsdonated to śaiva communities on Kashmir Smats.9 From the faded first writingthe retracer omitted the o-strokes above the ma. What is now a partial ya origi-nally was ya

˙m; traces of the vertical closing the ya on its right are still there, as is

the dot above for the nasal (Fig. 1d).Because of the retracer’s ignorance and because Fussman10 stuck to the good

rule that dissimilar forms should represent dissimilar letters, vartaka- was readas cortarū(+ na). The va here differs slightly from the one in divadharma, havinga sort of serif instead of ending in forked lines, but still it is certainly not a co.The ka does not show a continuous horizontal cross-bar, but has the right partof it attached separately, as said above. In Court’s eye-copy the right stroke iscombined with the trace of the original ka’s vertical to a hook pointing upwardson the right side. Possibly because of this strange hook Fussman assumed he wasdealing with a rū which can occur in similar shapes, from the late sixth centuryonwards.

The term vartaka- is found with many different meanings in the dictionar-ies, one of them is “a sort of brass or steel” according to Monier-Williams. ThePetersburg Dictionary has “damascierter Stahl” drawn from the Rājanigha

˙n˙tu,

which itself does not provide means to define the exact nature of the metal. Itschapter 13 lists all metals, speaking also of kā

˙msyāyovartaka

˙m kānta

˙m ki

˙t˙ta˙m

mu˙n˙da

˙m ca tīk

˙s˙naka

˙m, “Brass, copper, vartaka, iron, rust, iron (?), steel (?)”. The

Ānandakā˙n˙da11 lists in lines 2.1,9f. twelve kinds of loha, containing a similar list:

tīk˙s˙na

˙m ca mu

˙n˙dakam. bhujanga

˙m trapusa

˙m caiva rīti

˙h kā

˙msya

˙m ca vartakam.

The composition is dealt with in 2.7,19 kā˙msyārkarītilohāhijāta

˙m tad vartalo-

hakam, “varta-metal arises from bronze, (pure) copper, brass, iron (and) lead”.This can be compared to the metal analysis as given by Fussman,12 according to

9) Falk 2011: 22b.10) Fussman 1991: 152.11) Radhakrishna Sastri 1952.12) Fussman 1991: 144.


whom 18 to 19% of zinc was found, imported via brass, less than 2% each ofiron and lead, and around half a percent of silver and tin, the latter imported viabronze. The rest is copper.

The Ānandakā˙n˙da, written in South India probably after 1500 cites exten-

sively from the Rasendracū˙dāma

˙ni and the Rasaprakāśasudhākara, both with

origins in the 12th to 14th centuries. Considering the distance in time and spaceit is remarkable how close the analysis of the actual alloy components of themaskis compared to the description of the alloy many centuries later than its manu-facture.All components arementioned; unfortunately, the Sanskrit text does notprovide percentages.

There is another object from the same area where the mask was found andnot far removed in time which also tells us from which alloy it was made: a sac-rificial ewer of high quality from the Kashmir Smats, in the mountains northofMardan, reads:13 śrī-bhīmaya guhavasīnīya pa

˙nīavahīka

˙nama[

˙thām]āgalasya

rītīgha˙takasya dīvadharma nagara, “Pious donation of a pot (made) from bell-

metal for auspicious occasions, for śrī-Bhīmā residing in the cave, in themeeting-house of the water-carriers, (in) the town ⟨of Ahuramazda⟩.”The term rīti is partof the list of alloys cited above from the Ānandakā

˙n˙da. On the three ewers from

that site, twice we read devadharma and once dīvadharma, which can be com-pared to divadharma on our mask. Note that in two of the three cases of ewersno donor’s name is given.

-ekagrasthiradivya-: the sandhi is deficient and a stroke for a long vowel ismissing, as in so many other cases, so that vartakekagra- should be understoodas vartakaikāgra-. Fussman read our ka as na (cortarū+na), although it was seenin ka-shape already by Court in his eye-copy. The base-line would be positionedtoo high for a na, and in sidelight the full letter—although faint—can be madeout clearly, changing the supposed base-line of a na to the crossing line of a ka.Our gra (Fig. 1e) was taken by Fussman as śra with an untraceable -ī-vowel tomake up śrī. Since gra and śra are almost homomorphic this interpretation isunderstandable. Our sthi (Fig. 1e, f, h) is the first letter of hara in Fussman’sreading. In fact, it was this letter which in my first rubbing showed that a hais very unlikely. The letter shows a clear and plain full circle in its lower half anda sort of U-shape above it, which should be the right part of a sa or the body of a

˙sa. A

˙sa is unlikely since this would leave a much too large space between it and

the letter preceding it, namely gra. A sa, however, fills this space nicely with itsleft curl.

In principle, sva or stha could account for the full circle below; if we accepttraces as those of a dot inside the circle in the rubbing, only stha remains. The

13) Falk 2011: 22b no. c.


beginning of a half-circle leading into the curved part of the rim should be thatof a bent -i-mātrā. In combination with the lines preserved to its left, the lettercan only be sthi.

Fussman read -devo, where I see rather -divya-. The bend of the -i is nicelypreserved. The right side of the va’s body continues below in a pronouncedbend to the left (Fig. 1f, g) which returns to the right below, although almosterased. The only case of doubt is Fussman’s -o-mātrā, which should consist oftwo strokes slanting to the left and right above the va. On the right side the serifis so prolonged that it can be taken for such a stroke; on the left side nothingis seen beyond the serif itself. Therefore it is possible to read -divya- or -divyo, adifference which affects the syntax.brahmegutpa was read as graghegutpa before. The difference in assumed let-

ter shapes is slight, although the new reading makes more sense, once we under-stand its genesis. My bra instead of gra takes into account that the small bodyof the letter is definitely closed at its lower side (Fig. 1f, g) and that the alterna-tively considered ga in this script shows the usual outward curl at its lower leftleg as seen in ekagre and gutpa-. The next letter (Fig. 1g) has a clear -e-mātrāand shows three verticals, which Fussman combined to ghe. The right-most ver-tical is not visibly connected to the first two. Fussman obviously enlarged thebase-line to include the third vertical, while I see a connection in the upper androunded part, so that inmy view the vertical comes from above.The resulting let-ter can only be a ha. Fussman disregarded the faint traces below the ha, althoughCourt had seen them when he drew a curved vertical and a short horizontalline. These are two parts of a ma; the straight vertical on the right side is pre-served only in traces. Starting from the clear brawe can now read the first part asbrahme.

This could be a masculine noun in the locative, as brahma is found fromthe Śatapathabrāhma

˙na (4.1.4.10) onwards and is frequent in the Purā

˙nas. The

locative hardly makes sense, and it is tempting to read *brahmo, a nominativegoing with the other nominatives, but there certainly is no trace of a secondstroke to the right to build up the -o-mātrā. If the scribe added the stroke slantingin the wrong direction then a brahmā was intended. As a third possibility we canthink of a Prakritic nominative in -e. All such considerations would change thedepicted god from someone unmentioned to Brahma(n), amazing with regardto his features, but indicated by the palaeography.

The following gutpa seems to show a reversal of letters in a ligature and ismostlikely meant for gupta. Alternatively, we could consider gra for gu, but *gratpayields still less meaning. If it is correct to reconstruct an intended gupta thenboth gra in ekagra- and gu in gutpa- show a ga with a downwards bend for -uas well as -ra. Many scribes in the middle of the first millennium maintain theold difference by writing -ra with a downwards bend and an -u with an upwards


bend.Ourmodel of two similarly-looking shapes for both gra and gu is also foundin the Indore copper-plate inscription of Skandagupta, dated 466ad14 or theMajhgawam copper-plate of Hastin, dated 510ad.15mastaka

˙h was read as rmatoka. Although the differences look considerable

with regard to the transcription, the former reading can well be understood:a ma in Gupta times can have one or two bends on the left side. A rma witha single-bended ma looks like a double-bended ma. The only difference withcareful scribes is that the ra hook or bend protrudes above the top-line. Here(Fig. 1i), however, the top-bend is well inside the line. If we regard our scribe asexercising at least some care, then this is ma and not rma. What Fussman tookas the two wings of a medial -o is in fact a wavy line like a retrograde S lyingon its side (Fig. 1i), showing a bend downwards on the left and upwards on theright, something only sa provides. Similar forms are common in the fifth andsixth century, but in those cases the sa is written first inside the line and the taattached below, whereas here we have another case of a confusion with regard toligatures, with the ta written first and then the sa added, not below to producetsa as in the cases of tpa, but above. The resulting word mastaka at least providesa meaning suitable for a metal mask: “head”.

There are some dots in vertical succession after the ka (Fig. 1i), two of whichcan be taken for a visarga.

A real problem now arises about how to divide brahmegutpamastaka˙h: is this

to be understood as brahme = brahmā guptamastaka˙h or as brahmagupta-masta-

ka˙h? In the first case we had a Brahma(n) with a “hidden head”, in the second

“the head of Brahmagupta”.In this second case Brahmagupta will be hardly anything else than the donor.

The name as such is well-known. However, its place in the compound appearsstrange and in addition brahme- would be an inexplicable mistake for brahma-.This interpretation would provide a donor, without telling us who the god is.Śiva seems to be the most natural solution.

In the first case the solution is reverse: we had a god named, Brahma(n), butno donor. As seen above, a donation record from the Kashmir Smats also usesthe devadharmo ’yam introduction, provides information on the alloy used andleaves out the donor completely. I prefer this solution but, as Fussman did before,I think that further work may find arguments to reverse it.

14) Bhandarkar 1981: 308–312 with pl. XXX; Indoskript no. 77.15) Fleet 1888: 106–109; Indoskript no. 50.


Translation

The reading is, according to these considerations:

divadharma ya˙m vartakekagra-sthira-divya-brahme gutpa-mastaka

˙h

which can be translated as

Thispious donation is the hiddenhead ofBrahma, (made from) a copper-zinc alloy, fullyconcentrated, motionless (and) heavenly.

Palaeographical Remarks

Fussman dated the mask to 400 to 600ce, with a preference for the second halfof this time span. A check of all letters with IndoSkript leads to a peak around500ce, with all crucial forms already present in the span of 400 to 500, andso while confirming Fussman’s general impression, my preference goes towardthe earlier half. My learned friend compared the script to the documents foundin graffiti on rocks in the upper Indus valley. In contrast, IndoSkript shows aremarkable heap of hits in the true Gupta domains. The first crucial letter is thegu in gutpa. InAśokanBrāhmī an -u-mātrā is addedhorizontally to the base of ta,ga, ka, ra to produce tu, gu, ku and ru. If the stroke were added vertically it wouldbe difficult in many case to decide if the -u is there or not; with the horizontalstroke such an ambiguity can be avoided.This horizontal stroke is maintained insome areas, while it is changed to the vertical bend in others. On themask the guhas a bend downwards, which can be found in eastern UP and Madhya Pradeshfrom themiddle of the fifth century onwards, while it is absent in theWest at thesame time.

Language

While the text certainly aims at being Sanskrit, the scribe may have been influ-enced by a vernacular.He is remarkablyweak in ligatures and occasionally starts aligaturewith the letterwhichwould be usedfirst or exclusively in its Prakrit form.Hewrites vartakekagra, where vartakaikāgrawas required, as if ewashis choice ofai.Hewrites gutpa, as if influenced from gutta; andhewritesmata+s+ka, as if thePrakrit form matthaka had made him write the dental first. Another Prakritismcould be the -e-ending of a nominative in brahme. All cases have parallels on sealsand sealings from the North-West at the time when Kharo

˙s˙thī was being given

up for Brāhmī.16

16) Falk 2011: 14f.


The Religious Background

After having lost hold of a clearly śaiva Hara we have to ask who this god reallyis. He is bearded and sports a moustache unlike Śiva, he lacks the third eyeof Śiva, his earlobes are empty unlike Śiva’s. On the other hand his connec-tion with a snake on his head is reminiscent of Śiva, but already the Buddhaat Mathura was guarded by a many-hooded snake while in meditation. On theother hand we are told in clear terms who this fellow is: Brahma (if we read aPrakritic nominative brahme) or Brahman (if we suppose a mis-slanted -e meantfor -ā).

We are told that the figure depicted has three qualities, being “fully concen-trated, motionless and heavenly”. These terms are well-known from yogic con-texts. A passage added to the 13th book of the Mahābhārata (Mbh) (supp.13.11,235–236) says about the yogin:

nirudhya sarva-sa˙mkalpā

˙ms tato vai sthiratā

˙m vrajet

ekāgras tad vijānīyāt sarva˙m guhyatama

˙m param

suppressing all mind activities he will then turn to (bodily) motionlessness. Fully con-centrated he will recognize all this highest most concealed (entity).

Here we have two of our crucial terms, sthira and ekāgra. In Mbh 12.304,23–27we even get a mention of Brahman:

tathaivottaramā˙nasya ekāgramanasas tathā.23cd.

sthiratvād indriyā˙nā

˙m tu niścalatvāt tathaiva ca,

eva˙myuktasya tu muner lak

˙sa˙nāny upadhārayet.24.

sa yukta˙h paśyati brahma yat tat paramam avyayam,

mahatas tamaso madhye sthita˙m jvalanasa

˙mnibham.25.

etad dhi yoga˙m yogānā

˙m kim anyad yogalak

˙sa˙nam.27ab.

He should carry the marks of one who has the Highest on his mind and has his mindfully concentrated on account of the motionlessness of his sense-organs and of (hisown)motionlessness, (i.e. themarks) of amuniwho is united in thisway (with thedeity).Engaged he will see Brahman, that highest unchangeable (being), like fire placed in themiddle of a great darkness. 25.

Exactly this is the yoga of the yogins; what else could be indicative of yoga? 27ab

The Yogasūtra of Patañjali uses our terms as well. The sitting posture should bemotionless and comfortable (2,46 sthirasukham āsanam), ekāgratā with regardto themind (citta) is referred to in 3,10 and 11.However, Brahman is not found,the īśvara (2,45), a form of the puru

˙sa (1,23) is taking his place instead.

If we take the portrait as the one of Brahma(n) all icons fall in line: Brahman isoften bearded, Brahman has no third eye, Brahman is found in deep meditation.


The snake coiled around his head bun is unique, its three heads possibly just ameans to avoid the frugality of a single head.

The Hidden Head

If we look for habitually bearded gods, we find just Agni and Brahman and therelated B

˙rhaspati. There are other gods who have just one bearded head apart

from several without. An early example from Gandhara is seen in Srinivasan,17 astatue of Maheśa according to Srinivasan, with the three-eyed Śiva marking thecentre and three more heads in the round, one of them bearded.

The four-headed forms of Śiva are frequent and early and a bearded head in theback is known as well, as, e.g., on the multi-headed Maheśvara from Bāramūla,western Kashmir, dated to the sixth century by Paul.18 A fourth and beardedhead added to statues of Vi

˙s˙nu seems to be a development which took place

in Kashmir in the 7th century.19 More elaborate are the Vaiku˙n˙tha images from

Kashmir with the three heads of a lion, a man and a boar, representative of thethree gu

˙nas, while the back can show a head with the same headgear as ourmask,

showing a moustache and a snake.20It would make sense to call the fourth head in the back the “hidden one”, as

its existence is only realized by inspecting the back of the statue. There is onefour-faced statue of Brahman in theNationalMuseum,Karachi, slightly youngerthan the Masque Court (Fig. 2). Taddei saw that stylistically it stands very closeto our mask.21

Itwas found in a field atMīrpur-Khās, i.e. 42miles east ofHyderabad in Sindh,in “brass, or bronze”. The full body measures 97cm (3’ 2”). “It has four clean-shaven faces—those at the back and sides being smaller than the front face—andtheir hair is elaborately worked”. The right hand may have held a rosary, the leftone still shows traces of a water-jug.22 “The half-closed eyes look very much likethose of some of the Buddha images from the Mīrpūr-Khās stūpa”.23 Regardingdates, van Lohuizen-de Leeuw opted for the sixth century,24 according to Harleanother piece from the same place is “unlikely to date from before the middle

17) Srinivasan 2008: 131 figs 1 and 2.18) Paul 1986: 117.19) Pal 1973/74: 38.20) Banerji 1962: pl. XV figs 1 and 2.21) Taddei 1993: 348a.22) Van Lohuizen-de Leeuw 1979: 170.23) Cousens 1929: 10; cf. pls. XXVI and XXVII; Harle 1974: pl. 123.24) Van Lohuizen-de Leeuw 1979: 162.


Figure 2. Four-faced Brahman from Mīrpur-Khās, National Museum, Karachi.

of the fifth century A.D.”25 This concerns the Brahma fromMīrpur-Khās, whichlooksmore advanced and developed compared to theMasqueCourt, so that ourdating it in the fifth century is not at all impossible.

The back-face (Fig. 3) on the Mīrpur-Khās Brahman has not been publishedso far in a recognizable form; it shows a strange protrusion on its front, already inthe tiny four-sided view in Cousens 1929. Through the kind offices of Muham-mad Shah Bukhari, Curator (manuscripts) at the National Museum, Karachi, I

25) Harle 1974: 30.


Figure 3. The fourth face at the backside ofthe four-faced Brahman from Mīrpur-Khās.

received views from its present installation which show that this is a metal piecesoldered to the face for a fixation in the wall. The pictures show too that theback-face has another expression than theother three visible fromthe front; how-ever, the beard is missing. This shows to my mind that the fourth face can have abeard, but it need not do so, just as four-faced forms of Śiva can have a beardedback, optionally. Only for Vaiku

˙n˙tha images the bearded back seem to be almost

mandatory.With theMīrpur-Khās statue in bronzemay be compared one fromGandhara

in stone, published by Bopearachchi.26 Where the Mīrpur-Khās Brahman hasonly two arms, but a fourth face, the one “probablement de Rabatak” looksdecidedly younger in having only three heads, but four arms, hands and insignia.

Often it seems difficult to decide between Brahma and Śiva, as in the caseof the Berlin Brahma, one-headed and with four side-figures, whom the earlier

26) Bopearachchi 2008: 34 fig. 14.


literature regarded as Śiva-Lokeśvara.27 This piece showing Brahma has a dedica-tory inscription startingwithdeyadharme ya, as does ourmask. Pal28 remarks that“there need be nothing unusual in the inclusion of Brahmā in the Buddhist pan-theon”, while I would rather assume a vivid cross-contamination of ideas, formsand dedicatory phrases over all religious groups and their artists of the time inquestion.The late Upani

˙sads have many places where one god is just an aspect of

another one.The Śvetāśvatara-Upani˙sad (3,5) transformsRudra into “Śiva” once

he abstains from using his weapons against the devotee. In a higher position he isBrahman, and he is “hidden” in many forms in everything living, being also thesingle Puru

˙sa behind everything. People become immortal once they recognize

him as Īśa (3,6). As Puru˙sa he stands motionless (stabdha) in the sky like a tree

(3,9), comparable to our sthira Brahma. During this phase of Hinduism it doesnot seem advisable to pin down this highest being to just one name. Althoughbeing called brahma or brahman, he may still have had links to Śiva, as assumedby Fussman twenty years ago.

The legend presents three terms to characterize this fourth face, calling it ekā-gra, sthira and divya, of which at least ekāgra and sthira are found to describethe yogi who manages to become one with Brahman. The donor is kept unmen-tioned, if he did not appear on one of the other three heads. The piece is unique,as a piece of art, and it is so far unique as well with regard to itsmessage expressedin a legend.

Acknowledgements

Michel Landes at the Musée de Lattes and Marielle Pic at the BibliothèqueNationale made working on the original possible. I thank Oliver Hellwig forhelping me to understand the alchemist literature pertaining to the alloy; AbdulSamad Khan and Muhammad Shah Bukhari were most helpful for procuringphotographs of the Brahman statue from Mīrpur-Khās. Britta Schneider andIngo Strauch read earlier versions and made useful suggestions. Stefan Baumslooked after comprehensible English. To all of them I am very grateful.

References

Banerji, Adris (1962) Interesting images from south east Rajasthan. Lalit Kalā 11: 21–25.Bhandarkar,D.R. (1981) Inscriptions of the EarlyGuptaKings—rev. byB.Chhabra&G.S.Gai

(Corpus Inscriptorum Indicarum, 3). New Delhi (Archaeological Survey of India).

27) Pal 1973/74: 34.28) Pal 1973/74: 35.


Bopearachchi, Osmund (2008), Les premiers souverains Kouchans: Chronologie et iconogra-phie monétaire. Journal des Savants 2008: 3–56.

Cousens, Henry (1929) The Antiquities of Sind with historical outline (Archaeological Surveyof India, Imperial Series, 46). Calcutta (Government of India).

Errington, Elizabeth (1992) Inscribed mask of a naga-raja. Errington, Elizabeth & Joe Cribb,with Maggie Claringbull (eds.), Crossroads of Asia—Transformation in Image and Symbolin the Art of Ancient Afghanistan and Pakistan. Cambridge (Ancient India and Iran Trust):237–239.

Falk,Harry (2011)Commentary on the scribal aspects.AmanurRahman&HarryFalk,Seals,Sealings andTokens fromGandhāra (Monographien zur indischenArchäologie, Kunst undPhilologie, 21). Wiesbaden (Ludwig Reichert): 13–30.

Fleet, J.F. (1888) Inscriptions of the earlyGupta kings and their successors (Corpus InscriptionumIndicarum, 3). Calcutta.

Fussman, Gérard (1990) Un masque de Shiva provenant du Gandhara. Revue de la Biblio-thèque Nationale 37: 43–46.

Fussman, Gérard (1991) Le Masque Court, une effigie en laiton de Śiva au Gandhara. JournalAsiatique 279: 137–174.

Fussman, Gérard (1992) A Gandharan brass mask representing Śiva. Lalit Kalā 26: 13–14.Fussman, Gérard (2003) No. 318 “Le masque Court”. Osmund Bopearachchi & Christian

Landes, Christine Sachs en collaboration avec Nathalie Cayzac et Florence Millet, Cata-logue de l’ expositionDe l’ Indus à l’Oxus—Archéologie de l’Asie Centrale. Lattes, AssociationIMAGO, Musée de Lattes: 355–356; 369 pl. 318.

Ghose,Madhuvanti (2003)No. 319 “Masque d’homme”; no. 320 “Masque à faces multiples”.OsmundBopearachchi&ChristianLandes,Christine Sachs en collaboration avecNathalieCayzac et Florence Millet, Catalogue de l’ exposition De l’ Indus à l’Oxus—Archéologie del’Asie Centrale. Lattes, Association IMAGO, Musée de Lattes: 355–356; 369, nos. 319,320.

Harle, J.C. (1974)Gupta Sculpture—Indian Sculpture of the Fourth to the Sixth Centuries A.D.Oxford (Clarendon Press).

IndoSkript, an electronic Indian Palaeography, released 2006. Concept Harry Falk, workinggroups led by Harry Falk (Berlin) and Walter Slaje (Halle). http://userpage.fu-berlin.de/falk.

Larson, Gerald James (2004) The aesthetic ascetic: Abhinavagupta’s perspective on poetry,religion, and visual art. Mārg 56,2: 54–61.

Narahari Pandita (ed.) (1933) Rājanigha˙n˙tu. Calcutta (Siddheshvarayantra).

Pal, Pratapaditya (1973/74) A Brāhmanical triad from Kashmir and some related icons. Ar-chives of Asian Art 27: 33–45.

Paul, Pran Gopal (1986) Early Sculpture of Kashmir (before the middle of the eigth centuryA.D.). Dissertation Rijksuniversiteit te Leiden.

Prinsep, James (1836) Facsimiles of ancient inscriptions, lithographed. Journal of the AsiaticSociety of Bengal 5: 482–486.

Radhakrishna Sastri, S.V. (ed.) (1952) Ānandakā˙n˙da. Srirangam (Sri Vilasam Press).

Siudmak, J. (2005) Early Bronze Masks from Ancient Kashmir and the North-West Region.Ute Franke-Vogt & Hans-Joachim Weisshaar (eds.), South Asian Archaeology 2003—Pro-ceedings of the Seventeenth International Conference of the European Association of SouthAsian Archaeologist (7–11 July 2003, Bonn). Aachen (LindenSoft): 437–441.

http://userpage.fu-berlin.de/falk

http://userpage.fu-berlin.de/falk


Srinivasan, Doris Meth (2008) Hindu-Gottheiten in der Kunst von Gandhara. ChristianLuczanits & Michael Jansen (eds.), Das buddhistische Erbe Pakistans—Legenden, Klösterund Paradiese. Mainz (Philipp von Zabern): 130–134.

Taddei, Maurizio (1993) Review of Errington, Elizabeth & Joe Cribb (eds.), Crossroads ofAsia—Transformation in Image and Symbol in the Art of Ancient Afghanistan and Pakistan.Cambridge. East and West 43: 344–348.

van Lohuizen-de Leeuw, J.E. (1979)The pre-Muslim antiquities of Sind. J.E. van Lohuizen-deLeeuw (ed.), SouthAsianArchaeology 1975—Papers from the third InternationalConferenceof the Association of South Asian Archaeologists in Western Europe held in Paris. Leiden(Brill): 151–174.

Śiva or Brahma? The Masque Court at the BN, Paris

Documents