The Disputed Umā-Maheśvara in the Los Angeles County Museum of Art: A Case Study in Reattribution and Reinterpretation

Transcription

1 The Disputed Umā-Maheśvara in the Los Angeles County Museum of Art: A Case Study in Reattribution and Reinterpretation Stephen Markel Archives of Asian Art, Volume 58, 2008, pp (Article) Published by Duke University Press DOI: For additional information about this article Accessed 7 Jun :37 GMT

2 The Disputed Umā-Maheśvara in the Los Angeles County Museum of Art: A Case Study in Reattribution and Reinterpretation stephen markel Los Angeles County Museum of Art The genesis of this study is a March 2007 Los Angeles Times article 1 on an Indian stone sculpture of Umā- Maheśvara (Figs. 1, 1A) in the Los Angeles County Museum of Art (LACMA) that had been offered for sale after formal deaccession approval by the museum s Board of Trustees. 2 An by LACMA s former Senior Curator of Indian and Southeast Asian Art ( ), Pratapaditya Pal, protesting the deaccession, had engendered the newspaper article. The , which objected to the deaccession on the basis of the sculpture s art-historical importance, was sent to the museum s Administration and to the Los Angeles Times on 12 March. Subsequently, the proposed deaccession of the sculpture was cancelled to undertake an extensive scholarly investigation of its art-historical complexities. The following discussion presents this investigation s analysis and findings. The appropriateness of the proposed deaccession of Fig. 1. Umā-Maheśvara. Date disputed. Uttar Pradesh, India. Gray sandstone; h , w , d cm. Los Angeles County Museum of Art, From the Nasli and Alice Heeramaneck Collection, Museum Associates Purchase. M Photograph Museum Associates/LACMA. Fig. 1A. Back of Fig. 1. Photograph Museum Associates/LACMA.

3 88 ARCHIVES OF ASIAN ART the sculpture will not be addressed here; rather this study reviews the sculpture s history of attributions and presents evidence for revising Pal s dating and geographical attribution of the sculpture, which were central to his argument against its deaccession. A later date and a more precise geographical provenance are postulated, which are based on iconographic and stylistic comparisons with the closest relevant works of art and on recent scholarship on central Indian architectural sculpture. 3 Finally, drawing on a new understanding of the sculpture s most salient iconographic feature, the distinctive sitting position of Umā, this investigation presents a fresh interpretation of the iconological meaning of this unusual representation of Umā-Maheśvara. It is useful to begin by identifying the sculpture s iconography and also by placing it in a scholarly context, both of which help trace the sculpture s history of published attributions. The sculpture (accession number: M ; h cm) is gray 4 sandstone and depicts the Hindu god Śiva and his spouse Umā (a.k.a. Pārvatī). They are seated on their respective bull and lion mounts, with smaller figures below their feet that represent their two sons Gaṅeśa and Kumāra (riding his peacock mount), the devout, emaciated Bhṙṅgi, an unidentified male ascetic, and a female flywhisk bearer. This type of composition is known in Sanskrit texts as Umā-Maheśvara ( Umā and the Great Lord [Śiva]), but is identified in the museum s official records by its more familiar name, Shiva s Family. The sculpture entered LACMA s permanent collection in July 1972 as part of an exchange that included a number of works, which was predicated on the return of an Indian stone column, previously purchased in 1969, to the well-known New York dealer Nasli Heeramaneck. 5 Over the course of the following thirty-seven years, in various LACMA records, sponsored publications, and affiliated writings, Pal attributed the Umā-Maheśvara with only minor variation as follows: 1970: Uttar Pradesh, 6th 7th century : Uttar Pradesh, 6th 7th century : Uttar Pradesh, Kanauj (?), 7th century : Uttar Pradesh, Kanauj, 6th 7th century : Uttar Pradesh (?), circa : circa 600, place of origin unmentioned 11 Slightly varying attributions of the Umā-Maheśvara have been published by four other authors: Stella Kramrisch ( Markandi, Chanda District, Maharashtra (?), 6th 7th century ); 12 Thomas Donaldson ( Uttar Pradesh, 7th century ); 13 Alice Heeramaneck ( Central India (?), 6th 7th century ); 14 and Carlton Rochell ( Uttar Pradesh, circa 600 ). 15 All four authors suggested dates for the sculpture consistent with those of Pal. The authors geographical attribution, Uttar Pradesh, generally agreed with Pal s. The one disagreement came from Kramrisch, who proposed Maharashtra as the sculpture s place of origin. 16 The next step in this investigation is to examine the evidence and arguments for the aforementioned attribution and dating of the Umā-Maheśvara. Pal s earlier, slightly varying attributions in 1970, 1972, 1973, and 1974 can be aligned with his 1986 catalogue entry and 2007 to form a general consensus of attribution to Uttar Pradesh (?), circa In 2007 he further described the Umā-Maheśvara as the only monumental stone sculpture from the late Gupta period and part of the museum s small group of Gupta art. 18 Ascribing the sculpture to the celebrated Gupta period imbues it with great importance, but is problematic given South Asian art scholars current recognition of the frequent incongruity between the duration of a political dynasty and an affiliated artistic style. Strictly speaking, the Gupta dynasty ruled circa , with the last seventy-five years being a chaotic, fragmented period disrupted by the incursions of the Hūṅas. Artistically, however, the sublime figural and architectural modes of the Gupta period proper continued to inspire works across a wide expanse of India throughout the sixth century. Accordingly, the traditional artistic chronology is that The Gupta period as a whole may then be divided into an early Gupta period, extending, depending on the region, well into the fifth century, a Gupta period proper, and a late Gupta period beginning in the west perhaps as early as the second quarter of the fifth century but considerably later in the east. Works undertaken after the middle of the sixth century are then considered to belong to the post-gupta period. 19 The traditional conception of the post-gupta style includes works made across northern and central India after the Gupta period, from the mid-sixth century through the tenth century, 20 even though the multitude of works lumped together under the rubric is stylistically diverse and was produced under widely disparate patronage. Few Indian art historians still classify South Asian works of art by dynastic or political labels (including numerous labels referring to later, smaller kingdoms). 21 Instead, South Asian works of this period are commonly classified according to their specific regional traditions

4 STEPHEN MARKEL The Disputed Umā-Maheśvara in LACMA 89 Fig. 2. Umā-Maheśvara. Mid-8th c. Nand Chand, Panna District, Madhya Pradesh, India. Sandstone. Photograph: Courtesy of the American Institute of Indian Studies, and chronological manifestations. In these terms Pal s assignment of the Umā-Maheśvara to the late Gupta period is outdated. Moreover, even if one accepts, as Pal did, the traditional chronology in which late Gupta art comprises works produced between the mid-fifth century and the mid-sixth century, his dating of the sculpture to circa 600 places it in the traditional post-gupta period. This discrepancy calls into question one of the bases of Pal s argument to retain the sculpture. Evidence presented herein for reassigning the sculpture to the Deogarh region, circa , further undermines the notion of a Gupta origin, one of Pal s bases for the sculpture s importance. During this period Deogarh (ancient Daśārṅadeśa) was likely subject to the political authority of the Gurjara-Pratīhāra rulers of Kanyakubja. 22 In arguing against the deaccession of the sculpture, Pal further asserted that it is the earliest of its kind, or, as quoted in the Los Angeles Times in a much more restricted context, the earliest of its kind at LACMA. 23 This claim is problematic, because there exist several different iconographic forms of Śiva seated with Umā/ Pārvatī; 24 thus it depends upon which specific form(s) Pal meant by its kind. Nevertheless, despite this claim s inherent ambiguity, it is possible to assess both versions of the claim of

5 90 ARCHIVES OF ASIAN ART Fig. 3. Umā-Maheśvara. 8th c. Sankargarh, Satna District, Madhya Pradesh, India. Sandstone. Photograph: Courtesy of the American Institute of Indian Studies, earliest by focusing on the basic compositional elements of Śiva seated with Umā/Pārvatī. The sculpture of the Umā-Maheśvara is indeed the earliest representation in LACMA s collection of the two deities seated together, regardless of whether one accepts Pal s dating of circa 600 or follows the revised dating of circa proposed here. In the overall context of South Asian art, again regardless of which proposed date is accepted, the sculpture is certainly not the earliest instance of Śiva seated with Umā/Pārvatī. Several earlier examples can be cited: a fourth-century image at Mathura, Uttar Pradesh; 25 fifth-century images at Nachna-Kuthara, Panna District, Madhya Pradesh, 26 and Bhitargaon, Kanpur District, Uttar Pradesh; 27 and a mid-sixthcentury image at the great Śiva cave-temple at Elephanta (a.k.a. Gharapuri) near Mumbai. 28 Significantly, however, none of the fourth- or fifth-century images has the Holy Family tableau represented beneath the divine couple, which is a defining iconographic feature of the present subset of Umā-Maheśvara image. 29 At Elephanta the area beneath Śiva and Pārvatī playing dice on Mt. Kailasa 30 is damaged, but enough of the composition survives to suggest that it was originally a representation of Śiva s gaṅas (dwarf followers of Śiva) tugging at his bull mount. Comparisons with analogous scenes at contemporaneous sites, such as at Sondni, Mandasor District, Madhya Pradesh, and the Dhumar Lena Cave (No. 29) at Ellora, Maharashtra,

6 STEPHEN MARKEL The Disputed Umā-Maheśvara in LACMA 91 Fig. 4. Umā-Maheśvara. 8th c. Banpur, Lalitpur District, Uttar Pradesh, India. Sandstone. Photograph: Courtesy of the American Institute of Indian Studies, indicate that this was a standard iconography during the sixth century. 31 Thus, it is highly unlikely that a Holy Family tableau originally figured in the Elephanta representation. The earliest known representations of Umā-Maheśvara images with a Holy Family tableau depicted along the base are from Nand Chand, Panna District, Madhya Pradesh, dating from the mid-eighth century (Fig. 2); 32 Sankargarh, Satna District, Madhya Pradesh dating from the eighth century (Fig. 3); and Banpur, Lalitpur District, Uttar Pradesh (near Deogarh), dating from the eighth century (Fig. 4). Thus, assuming Pal s dating, the LACMA sculpture is indeed among the earliest extant examples in South Asian art of an Umā- Maheśvara image featuring a Holy Family tableau on its base. But in light of evidence presented below for revising its dating to circa , and thus, the sculpture must be recognized as belonging to the seminal eighth-century group of examples rather than preceding them by at least a century. Having discussed Pal s general claims about the Umā-Maheśvara s dynastic affiliation and primacy, it is now necessary to review the specific works of art that he has proposed as pertinent examples in support of his arguments. Each of his comparisons are examined in the order presented in the catalogue entry in his Indian Sculpture, Vol Then I shall suggest alternative comparative examples that I consider more germane to

7 92 ARCHIVES OF ASIAN ART Fig. 5. Detail of Fig. 1: Heads and upper torsos of the LACMA Śiva and Umā. Photograph Museum Associates/LACMA. the attribution of the Umā-Maheśvara. Pal s first comparison is stylistic; the other three comparisons are iconographic. 34 The significance of this crucial distinction is addressed in detail below. First, for the sake of clarity, Pal s argument is quoted here in its entirety: While Kramrisch is correct in commenting on the uniqueness of the image and dating it to the sixth seventh century, her suggested provenance Markandi in Maharashtra is highly unlikely. Apart from the fact that the Markandi temples are five centuries later, it is very difficult to accept this work as a stylistic precursor. Moreover, the sculpture generally is not rendered in the style of the monuments of Maharashtra. On the contrary, the elongated faces are somewhat reminiscent of a terra-cotta Siva head found at Ahichchhatra (V. S. Agrawala , pl. XLIV), while Uma may be compared with the similarly seated Mother Goddesses from Madhya Pradesh (Harle 1974, figs ). Parvati s coiffure with coiled bun at the back of the head is worn by female figures in the Gupta-period temple at Deogarh (Williams 1982, fig. 204), while the curious cylindrical ear ornament is more commonly found in figures

8 STEPHEN MARKEL The Disputed Umā-Maheśvara in LACMA 93 from Bihar. Thus, Uttar Pradesh rather than Maharashtra is a more likely source for this intriguing sculpture. 35 Pal finds that the elongated faces of the LACMA Śiva and Umā (Fig. 5) are somewhat reminiscent of a terra-cotta Śiva head from Ahicchatra [modern transliteration], Bareilly District, Uttar Pradesh, which is now in the National Museum, New Delhi (acc. no.: ; Fig. 6). The date of the Ahicchatra Śiva head is not specified in Agrawala s excavation report, but may be presumed to accord with material recovered from the site s Stratum III: A.D. 350 to In subsequent publications the head was more specifically attributed to the fifth century 37 and recently to the end of the 5th century beginning of the 6th century (or c ). 38 On the basis of these attributions, the Ahicchatra Śiva head should be understood as dating at least a century earlier than Pal s attribution of circa 600 for the LACMA Umā-Maheśvara. A stylistic comparison reveals significant differences between the Ahicchatra Śiva head and the head of the LACMA Śiva (and that of Umā, which for the purposes of this analysis can be regarded as stylistically identical to the Śiva head). Although all three heads do indeed have elongated faces, the Ahicchatra Śiva s face is more oval and narrows toward the chin more than those of the LACMA Śiva and Umā. Its profile is also more pointed, and the nose protrudes farther than on the LACMA faces. More significantly, moreover, the Ahicchatra Śiva differs substantially in the specific shape of its facial features from the LACMA faces. The Ahicchatra Śiva s forehead is triangular, coming to a pronounced point at the bridge of the nose, and his brows are incised; whereas the lower border of the LACMA foreheads is continued by the bas-relief brows, and together they form the shape of a compound bow at rest. The Ahicchatra Śiva s eyes are narrowly almond in shape, angled upward at the outer corners, and set in stylized deep sockets formed by the bottom of the brow, which melds smoothly into the side of the nose. In contrast, the LACMA eyes are rounder, more nearly horizontal and with half-closed lids, and the sockets of the eyes are naturalistically depicted, replicating a facial skeletal structure rather than being stylized into a continuous plane. The Ahicchatra Śiva has an overly full lower lip with uplifted corners forming a broad smile, and its chin is set back from the lips front edge. The LACMA faces have full lips with deep dimples at the corners, more nearly horizontal within the facial contours, and their chins extend well past the lips in unnatural block-shaped protrusions. The two Śiva heads differ radically even in the formal artistic organization of their ascetic hairstyles ( jāṫamukuṫa). The Ahicchatra Śiva has a simple tripartite arrangement: two thick braids of hair rise to a point that is clasped by an equally thick circular braid. The jāṫamukuṫa of the LACMA Śiva is extremely elaborate, with numerous thinner braids complexly interwoven. Prominent vertical braids are clasped at the top and bottom by horizontal braids. The circular braid on the top is much less emphasized, being at the back of the Fig. 6. Śiva Head. Ca Ahicchatra, Bareilly District Uttar Pradesh, India. Terra-cotta; h. 17 cm. National Museum, New Delhi, Courtesy National Museum, New Delhi. From V. S. Agrawala, The Terracottas of Ahichchhatrā, Ancient India, Bulletin of the Archaeological Survey of India, no. 4 (July 1947 January 1948), pl. XLIV, 113a.

9 94 ARCHIVES OF ASIAN ART Fig. 7. Śiva head. Mid-8th c. Kota?, Shivpuri District, Madhya Pradesh, India. Sandstone. Archaeological Museum, Gwalior. Photograph: Courtesy of the American Institute of Indian Studies, head and partially hidden behind the front and surmounting components. Although in South Asian art the arrangement and depiction of contemporaneous ascetic hairstyles vary widely, the radical conceptual difference between the jāṫamukuṫa of the Ahicchatra and LACMA Śiva heads is strongly suggestive of different geographical and/or chronological origins. Thus, upon careful examination, the terra-cotta Ahicchatra Śiva head must be recognized as only superficially similar to the LACMA Śiva head. More closely comparable with the LACMA Śiva head is a stone Śiva head dating from the mid-eighth century that is probably from Kota, Shivpuri District, Madhya Pradesh (Fig. 7). The Kota Śiva head, now in the Archaeological Museum, Gwalior, was published at least a decade ago. 39 Many of its features are remark-

10 STEPHEN MARKEL The Disputed Umā-Maheśvara in LACMA 95 Fig. 8. Kaumari. Early 5th c. Badoh- Pathari, Vidisha District, Madhya Pradesh, India. Sandstone. Photograph: Courtesy of the American Institute of Indian Studies, ably similar to those of the LACMA Śiva head. The Kota Śiva head s jāṫamukuṫa is arranged in almost exactly the same manner as that of the LACMA Śiva head. It has the same distinctive complex organization, featuring prominent vertical braids of hair clasped between overlapping horizontal braids at the top and bottom. Looped braids frame the sides of the face. Indeed, the chief difference between the two representations is their stylistic treatment. The braids of the Kota Śiva head are flatter, slightly thinner, and more uniform than those of the LACMA Śiva. Therefore they form a more mannered, rhythmic pattern than those of the LACMA Śiva head. Other points of similarity between the Kota Śiva head and the LACMA heads are the rectangular face, bow-shaped relief brows, distinct foreheads that do not merge into the plane of the nose, lips neither overfull nor smiling, half-closed eyes (more visible in the LACMA Umā), and triple neck lines. The two Śiva heads also have locks of hair curling gracefully on the shoulders. Thus, the extremely close similarity in the stylistic treatment of the facial features and the distinctive manner of arranging the jāṫamukuṫa strongly suggests that the Kota Śiva head and the LACMA heads share a general geo-chronological origin. There are, however, enough stylistic differences, such as the Kota Śiva head s flatter

11 96 ARCHIVES OF ASIAN ART Fig. 9. Mātṙkā. Early 5th c. Besnagar, Vidisha District, Madhya Pradesh, India. Sandstone. National Museum, New Delhi, Photograph: Courtesy of the American Institute of Indian Studies, surface plane and aforementioned emphasis on rhythmic patterns, to indicate that these pieces are probably not from the same exact site or artist s workshop. Additional stylistic comparisons will be presented below in order to propose a more precise place of origin. Returning to Pal s stylistic comparisons, he compares the unusual posture of the LACMA Umā, who sits with her legs pendent in the so-called European posture (pralambapadāsana), with the similarly seated Mother Goddesses from Madhya Pradesh, 40 and states that the posture of the goddess, sitting imperious with her legs the way they are positioned, goes out of fashion in India after the 6th century. 41 The two Mother Goddesses from Madhya Pradesh to which Pal refers

12 STEPHEN MARKEL The Disputed Umā-Maheśvara in LACMA 97 Fig. 10. Ambikā. Late 7th c. Gyaraspur, Vidisha District, Madhya Pradesh, India. Sandstone. Archaeological Museum, Gwalior. Photograph: Courtesy of the American Institute of Indian Studies, belong to two well-known groups of Seven Mother Goddesses (Saptamātṙkā). 42 Both sets of Mother Goddesses are from the Vidisha District of Madhya Pradesh and date from the early fifth century. One is a rock-cut Kaumari from Badoh-Pathari (Fig. 8); the second is an unidentified Mother Goddess from Besnagar, which is now in the National Museum, New Delhi (Fig. 9). It should also be noted that there are three rock-cut groups of Mother Goddesses in the nearby vicinity of Deogarh, Lalitpur District, at least two of which (from Rajghati and Naharghati) show the goddesses in the European posture. 43 The significance of these sets of

13 98 ARCHIVES OF ASIAN ART Fig. 11. Gajalakṡmī. 9th c. Deogarh, Lalitpur District, Uttar Pradesh, India. Sandstone. Site Museum, Deogarh. Photograph: Courtesy of the American Institute of Indian Studies,

14 STEPHEN MARKEL The Disputed Umā-Maheśvara in LACMA 99 Fig. 12. Figures on doorjamb. Ca Daśāvatāra temple, Deogarh, Lalitpur District, Uttar Pradesh, India. Sandstone. Photograph: Courtesy of the American Institute of Indian Studies, Mother Goddesses to the new iconological interpretation of the LACMA Uma-Mahesvara will be further addressed below. Pal s assertion that the European posture used for female deities goes out of fashion in India after the 6th century is controverted by at least two later examples that are, in my view, very close chronologically and geographically to the LACMA Umā-Maheśvara. They are an Ambikā from Gyaraspur, Vidisha District, Madhya Pradesh, dating from the late seventh century, now in the Archaeological Museum, Gwalior (Fig. 10), 44 and a Gajalakṡmī from Lalitpur District, Uttar Pradesh, dating from the ninth century, now in the Site Museum, Deogarh (Fig. 11). In addition, the European posture was used for images in southern India as early as the seventh century and continued for over a millennium. 45 It was frequently used for depicting in both stone and copper alloy images of Umā while seated with Śiva. 46 The kinship of Umā s European posture to those of the Vidisha District and Lalitpur District Mother Goddess images has significant implications for the reinterpretation of the LACMA Umā-Maheśvara. As Pal also noted, Parvati s coiffure with coiled bun at the back of the head is worn by female figures in the Gupta-period temple at Deogarh 47 (Fig. 12). But the hairstyles of the female attendants on the door-

15 100 ARCHIVES OF ASIAN ART Fig. 13. Caurī bearer. 8th c. Kanauj, Farrukhabad District, Uttar Pradesh, India. Sandstone. State Museum, Lucknow. Photograph: Courtesy of the American Institute of Indian Studies, jamb of the Daśāvatāra Temple at Deogarh, which dates circa , consist of multiple bands of coiled hair in front of a multisegmented bun, and are much more complex than the single-band bun of the LACMA Umā. Much closer to the hairstyle of the LACMA Umā is that of a Kanauj female flywhisk bearer from style similar to that of the LACMA Umā; the female flywhisk bearer is from Kanauj, Farrukhabad District, Uttar Pradesh, dating from the eighth century, and currently in the State Museum, Lucknow (Fig. 13). The Kanauj attendant s hairstyle consists of a single band of coiled hair in front of a nonsegmented bun, and is almost identical to the LACMA Umā s hairstyle. Thus, the close similarity in hairstyle between the Kanauj attendant and the LACMA Umā, and the former s eighth-century date, are additional evidence supporting an eighth-century date for the Umā-Maheśvara. Oddly, Pal supported his claim that the LACMA sculpture originated in Uttar Pradesh by citing the distinctive type of earring that the Umā wears in her left

16 STEPHEN MARKEL The Disputed Umā-Maheśvara in LACMA 101 Fig. 14. Hara-Gaurī. 9th c. Bhanpura, Mandasor District, Madhya Pradesh, India. Sandstone. Central Museum, Indore. Photograph: Courtesy of the American Institute of Indian Studies, ear. But he then goes on to state that Umā s curious cylindrical ear ornament is more commonly found in figures from Bihar. 48 Leaving aside the ornament s use in Bihar as implausible support for locating the LACMA sculpture in Uttar Pradesh, the ornament is typically represented in Bihar sculpture dating from the eighth to ninth century 49 rather than in that from the sixth to seventh century, which would correspond to Pal s proposed date for the Umā-Maheśvara. Moreover, the ornament is also found on at least one ninth-century stone sculpture from Madhya Pradesh: an image of Hara-Gaurī (Śiva and Umā) from Bhanpura in the Mandasor District, which is now in the Central Museum, Indore 50 (Fig. 14). Finally, the curious cylindrical ornament is actually a well-known type of ear stud (generically called a tārki), which could be made of various materials, including bone, bamboo, wood, and metal. Its use survives in tribal traditions extending over Madhya Pradesh, the new state of Jharkhand, and Bihar. 51 In sum, the comparative framework Pal used to validate his date of circa 600 for the LACMA Umā- Maheśvara breaks down under closer scrutiny and in the face of alternative comparisons. The analysis provided above demonstrates that the sculpture most likely dates from the eighth century. It was most plausibly made in southwestern Uttar Pradesh or in neighboring Madhya Pradesh, but iconographic comparisons alone cannot determine which region is more probable. Analysis of figural style, as hereinafter performed on the LACMA Umā-Maheśvara, is generally far more accurate than iconographic comparisons for establishing the

17 102 ARCHIVES OF ASIAN ART Fig. 15. Śiva Andhakāsuravadha (Śiva Slaying the Demon Andhaka). Ca Probably southern Uttar Pradesh or neighboring Madhya Pradesh, India. Red sandstone; h. 69.2, w. 41.6, d cm. The Brooklyn Museum of Art, Gift of the Ernest Erickson Foundation, Inc., correct geographical and chronological attribution of historical South Asian sculpture. That is because figural style typically develops in and is confined to a particular locale and temporal duration. Stylistic differences between adjacent regions during the same time period are often subtle, but they can be ascertained by careful examination. In contrast, iconographic features, such as a particular posture, are normally based on textual prescriptions. Thus, they frequently cross regional boundaries and persist beyond their period of origin. Regrettably, with the exception of the Ahicchatra Śiva head, Pal did not attempt to present stylistic comparisons relevant to the Umā-Maheśvara. Although an extensive review of the art-historical literature and image databases failed to discover an extant Hindu temple site with sculpture that stylistically matches the LACMA Umā-Maheśvara, at least one sculpture now in a museum collection exhibits several comparable stylistic features the representation of Śiva Andhakāsuravadha (Śiva Slaying the Demon Andhaka) in the Brooklyn Museum of Art (acc. no.: ) (Fig. 15). It has been attributed by Darielle Mason to probably southern Uttar Pradesh or neighboring Madhya Pradesh, circa Her perceptive stylistic comments include: Among other features, the rectangular face and closely set eyes, which are pointed at the outer corners, tie this piece to images from ancient Daśārṅadeśa, particularly those of the late eighth-century Jaina temple 12 at Deogarh. 53

18 STEPHEN MARKEL The Disputed Umā-Maheśvara in LACMA 103 Fig. 16. Detail of Figure 1: Head of LACMA Umā (three-quarter view). Photography: 6 Museum Associates/ LACMA. Besides the analogously rectangular faces of the LACMA and Brooklyn sculptures, the distinctive treatment of the Brooklyn Śiva Andhakāsuravadha s eyes, pointed at the outer corners, is highly significant for refining the attribution of the LACMA Umā-Maheśvara. Erosion has abraded the corners of the LACMA Śiva s eyes, but Umā s eyes (Fig. 16) have the same markedly pointed outer corners as those of the Brooklyn image. Additional stylistic features common to the LACMA and Brooklyn sculptures include the deep dimples at the corners of the mouth (mentioned above), Śiva s disproportionately oversized hands (note in particular the LACMA Śiva s huge hand resting on Umā s shoulder), and the idiosyncratic treatment of Śiva s chest, with the lower edge of the pectoral muscles defined by a sharp break in the plane of the torso (see Figs. 1, 5). Following Mason s stylistic link between the Brooklyn Śiva and sculptures of the late eighth-century Jaina Temple 12 at Deogarh, 54 let us compare the facial features of Temple 12 s main jina image (Fig. 17) with those of the LACMA Umā-Maheśvara. The comparison reveals the same rectangular face, half-closed eyes, and protruding chin. Thus, the Brooklyn Śiva Andhakāsuravadha and the Deogarh jina strongly suggest the Deogarh region of southwestern Uttar Pradesh as the most likely origin of the LACMA sculpture. One remaining stylistic feature and a final iconographic feature of the LACMA Umā provide additional

19 104 ARCHIVES OF ASIAN ART Fig. 17. Detail of head of main jina image. Late 8th c. Jaina Temple 12, Deogarh, Lalitpur District, Uttar Pradesh, India. Sandstone. From Klaus Bruhn, The Jina-Images of Deogarh (Leiden: E. J. Brill, 1969), fig. 8. corroboration of its revised date and general place of origin. Under Umā s breasts are three prominent parallel convex rolls that arch in the middle and extend across the front of her torso (see Fig. 5). Mason interprets them as ridges or rolls of flesh, but on Umā they could also be intended as a stylized depiction of her ribcage an indication of her ascetic nature and prowess. These distinctive markings are commonplace in sculpture from Madhya Pradesh and Uttar Pradesh, primarily of the eighth century and continuing in some areas into the ninth century. 55 A representative example of this common stylistic feature can be found on an Ambikā from Deogarh dating from the late eighth or the ninth century (Fig. 18). Another iconographic feature

20 STEPHEN MARKEL The Disputed Umā-Maheśvara in LACMA 105 Fig. 18. Ambikā. 9th c. Deogarh, Lalitpur District, Uttar Pradesh, India. Sandstone. Photograph: Courtesy of the American Institute of Indian Studies, common to this general region and period is present on both the Deogarh Ambikā and LACMA Umā: goddesses wear a pendant on a long chain that idiosyncratically curves off to one side instead of following the body s vertical axis (see Figs. 5, 10, 11, 13). Iconographic comparisons relevant to the LACMA Umā-Maheśvara suggest its general place and date of origin to be Uttar Pradesh or Madhya Pradesh, eighth century. By incorporating the more precise analysis possible with stylistic comparisons, the most probable geographic and chronological origin has been refined to the Deogarh region of southwestern Uttar Pradesh, circa This revised attribution of the LACMA Umā- Maheśvara now allows us to correct earlier misinterpretations of the so-called stern facial expressions of Śiva and Umā. It will be apparent from the descriptive comments quoted below that Heeramaneck, Kramrisch,

21 106 ARCHIVES OF ASIAN ART Fig. 19. Four of the seven Mother Goddesses. Ca Madhya Pradesh or Uttar Pradesh, India. Sandstone; h. 56, w. 84, d. 15 cm. Asian Art Museum of San Francisco, Gift of Dr. Stephen A. Sherwin and Merrill Randol Sherwin, F Asian Art Museum of San Francisco. Used by permission.

22 STEPHEN MARKEL The Disputed Umā-Maheśvara in LACMA 107 Fig. 20. Umāpati. 4th c. Bhita, Allahabad District, Uttar Pradesh, India. Terra-cotta. Indian Museum, Kolkata, A10380/NS1209. From Doris Meth Srinivasan, Many Heads, Arms, and Eyes: Origin, Meaning, and Form of Multiplicity in Indian Art (Leiden: E. J. Brill, 1997), pl

23 108 ARCHIVES OF ASIAN ART and Pal, went to great lengths in their attempts to explain why the figures countenances differed so significantly from what would be expected if the sculpture had actually been made in the sixth or seventh century. This representation of Siva and Parvati is of a rather stern and monumental character. That they are indeed deities is indicated by their haloes, but as a certain individualization appears in the faces it is probable that a royal couple are here represented as Siva and Parvati. 56 Stern and straight, the Great Lord (Mahesvara) and Umā/Pārvatī, his wife, confront the devotee.... Śiva s erect bearing and commanding physique show him in his majesty rather than in his grace. 57 Siva is sternly dignified and majestic, while Uma s facial expression and posture convey aloofness, if not disdain.... The unknown sculptor certainly did not represent Siva s spouse as a timid acquiescent female, as she is generally shown in such compositions. 58 The so-called individualization and stern facial expressions of the LACMA Umā-Maheśvara can now be more accurately understood as the distinctive countenance created by exaggerated facial features characteristic of much sculpture made in Uttar Pradesh and Madhya Pradesh during the eighth and ninth centuries. A representative example that clearly illustrates this stylistic feature is a relief carving of four of the seven Mother Goddesses from Madhya Pradesh or Uttar Pradesh, dating from circa , now in the Asian Art Museum of San Francisco (acc. no.: F ) (Fig. 19). 59 Finally, a new iconological interpretation of the LACMA Umā-Maheśvara can be presented, based on a fresh understanding of the sculpture s most salient iconographic feature. There are three keys to this interpretation. The first key is Umā s unusual European posture, about which Pal simply noted, During the Gupta period this posture generally was assigned to Mother Goddesses rather than to Umā. 60 The second key is a terra-cotta sculpture of Śiva seated with Umā from Bhita, Allahabad District, Uttar Pradesh, dating (by excavation) from the fourth century, 61 which is now in the Indian Museum, Kolkata (acc. no.: A10380/NS1209) (Fig. 20). The iconographic feature of this sculpture that makes it crucial for reinterpreting the LACMA Umā-Maheśvara is that the Bhita Umā is depicted in the same European posture. Srinivasan identifies the Bhita image as Umāpati and provides a vague translation of Umāpati as the divine couple. 62 Donaldson follows Srinivasan in identifying the image as Umāpati, but also does not perceive the significance of Umā s European posture. 63 The third key in the new interpretation is that Umāpati literally means husband of Umā. This specific terminology refers to an aspect of the deities that emphasizes Śiva as the primeval Father God and Umā as the great Mother Goddess. 64 Therefore, given that the LACMA Umā is represented in the same European posture, traditionally used for images of the Mother Goddesses, the sculpture was likely made in the Deogarh region of southwestern Uttar Pradesh. Also, given that Mother Goddesses are frequently shown seated in the European posture at several sites within this fairly small general area (Deogarh, Besnagar, Badoh-Pathari, and Bhita), it seems logical to interpret the LACMA Umā-Maheśvara as iconologically emphasizing Umā s role as the great Mother Goddess and that of her beloved husband Śiva as the primeval Father God. Moreover, in light of the small size of the geographical area in which Mother Goddesses and Umā were shown in the European posture, it is likely that this represents a regional iconographic tradition, one which continued through at least four centuries of artistic production. Since the eighth century in Central India is recognized as a dynamic period of Hindu iconographic development and artistic innovation, 65 it is equally likely that the LACMA Umā-Maheśvara is an inspired expression of this fertile, creative age. To conclude, the previously proposed date of circa 600 for the LACMA Umā-Maheśvara and the affiliation of the sculpture with the Gupta dynasty cannot be sustained. Stylistic comparisons presented here call for a revised date of circa for the sculpture and narrow its probable place of production to the Deogarh region of southwestern Uttar Pradesh. The revised date of the Umā-Maheśvara explains its distinctive facial expressions not as anomalies of circa 600, but as characteristic of eighth-century images. Finally, a consideration of the Umā-Maheśvara s most important iconographic feature reveals the deeper iconological meaning of the sculpture. Notes 1. Suzanne Muchnic, LACMA does an about-face on art sale, Los Angeles Times, Saturday, 17 March 2007, Calendar: E1, E20; and Museum reverses decision: LACMA does an about-face on sale of ancient sculpture, Los Angeles Times. Calendarlive.com. 17 March

24 STEPHEN MARKEL The Disputed Umā-Maheśvara in LACMA As The Harry and Yvonne Lenart Curator and Department Head of South and Southeast Asian Art at LACMA, I was principally responsible for initiating the proposed deaccession. The museum s official deaccession policy was strictly followed in the formal approval process necessary to authorize the deaccession. 3. I wish to thank a number of scholars for sharing their expertise and discussing this sculpture s attribution with me: Robert L. Brown (LACMA and University of California, Los Angeles [ UCLA]), Tushara Bindu Gude (LACMA and UCLA), Julie Romain (LACMA and UCLA), Walter M. Spink (Professor Emeritus, University of Michigan), John C. Huntington (Ohio State University), Michael D. Willis (British Museum), and especially Darielle Mason (Philadelphia Museum of Art) and Donald M. Stadtner (formerly University of Texas, Austin). Although any errors in this article are, of course, my sole responsibility, it is nevertheless significant to note that all the aforementioned scholars concur with the revised dating for the sculpture in question. 4. At present, the entire front surface of the sculpture has a coating of a greenish brown synthetic polymer, polyvinyl acetate (PVAc), which was brushed onto it prior to its acquisition by the museum. The uncoated surface of the pristinely cleaned stone visible on the back of the sculpture is gray (see Fig. 1A). 5. According to the museum s official records, the Umā-Maheśvara was part of a group of objects received in exchange in 1972 for a purportedly early Indian stone column (M ), which had been purchased from the Heeramanecks in March Prior to its accession, the Umā-Maheśvara had been on loan at LACMA since July 1970 (Loan number: L ). No earlier provenance information is available in LACMA s records. 6. Incoming loan record. 7. Accession record. 8. LACMA Bulletin (1973), p. 50, fig Art of Asia Recently Acquired by American Museums 1972, Archives of Asian Art, vol. 27 ( ), p. 99, fig Pratapaditya Pal, Indian Sculpture: A Catalogue of the Los Angeles County Museum of Art Collection. Vol. 1: Circa 500 B.C. A.D. 700 (Berkeley, Los Angeles, and London: Los Angeles County Museum of Art and University of California Press, 1986), pp , no. S Muchnic, LACMA does an about-face. Pratapaditya Pal, personal to LACMA administration and Los Angeles Times (12 March 2007). 12. Stella Kramrisch, Manifestations of Shiva (exh. cat.) (Philadelphia: Philadelphia Museum of Art, 1981), pp , no Thomas Eugene Donaldson, Śiva-Pārvatī and Allied Images, 2 vols. (New Delhi: D. K. Printworld, 2007), vol. 1, p. 469; vol. 2, p. 448, fig Alice Heeramaneck, Masterpieces of Indian Sculpture from the Former Collections of Nasli M. Heeramaneck (New York: By the Author, 1979), no Sacred and Sublime: Art from India and Southeast Asia (New York: Carlton Rochell Asian Art, 2007), no Kramrisch s attribution will not be discussed herein because I agree with Pal s arguments disproving a Maharashtran origin for the sculpture; Pal, Indian Sculpture, p Rochell follows Pal in this regard and neither Heeramaneck nor Donaldson provide any justification for their geographical attribution. 17. Muchnic, LACMA does an about-face. Pal, personal ; and Pal, Indian Sculpture, p Pal, Personal . Curiously, however, Pal did not include the Umā-Maheśvara in his major exhibition and catalogue on Gupta art. See Pratapaditya Pal, The Ideal Image: The Gupta Sculptural Tradition and Its Influence (exh. cat.) (New York: The Asia Society, 1978). 19. J. C. Harle, Gupta Sculpture, Indian Sculpture of the Fourth to Sixth Centuries A.D. (Oxford: Clarendon Press, 1974), p J. C. Harle, The Post-Gupta Style in Indian Temple Architecture and Sculpture, Journal of the Royal Asiatic Society of Arts, no. 5253/125 (1977), pp For example, see Pramod Chandra, The Study of Indian Temple Architecture, in Studies in Indian Temple Architecture: Papers Presented at a Seminar held in Varanasi, ed. Pramod Chandra (New Delhi: American Institute of Indian Studies, 1975), pp ; Joanna Gottfried Williams, The Art of Gupta India: Empire and Province (Princeton: Princeton University Press, 1983), p. 3; Michael Willis, Temples of Gopakṡetra: A Regional History of Architecture and Sculpture in Central India AD (London: British Museum Press, 1997), pp , Michael W. Meister and M. A. Dhaky, eds., Encyclopaedia of Indian Temple Architecture. Vol. 2, pt. 2: North India, Period of Early Maturity, c. A.D (Delhi: American Institute of Indian Studies, Oxford University Press, 1991), pp ; Vishakha N. Desai and Darielle Mason, eds., Gods, Guardians, and Lovers: Temple Sculptures from North India A.D (exh. cat.) (New York: The Asia Society Galleries, 1993), p Muchnic, LACMA does an about-face. Ascertaining the validity of the assertion is complicated because it is unknown whether the Los Angeles Times qualified rephrasing represents a later clarification by Pal or is the reporter s supposition. 24. Besides the Umā-Maheśvara type of image featuring Śiva seated with Umā/Pārvatī (typically with their sons Gaṅeśa and Kumāra, and the ascetic Bhṙṅgi shown in a tableau beneath them), related iconographic forms include Śiva and Pārvatī playing dice on Mt. Kailasa (one of the earliest forms), Rāvaṅānugraha (Śiva and Pārvatī seated together on Mt. Kailasa with the demon king Rāvaṅa

25 110 ARCHIVES OF ASIAN ART imprisoned beneath them), Vṙṡavāhana (Śiva and Pārvatī seated on Śiva s bull mount), and Umāsahita (Śiva and Umā standing together). For a recent thematic survey of Śiva and Pārvatī images, see Donaldson, Śiva-Pārvatī. 25. T. K. Biswas and Bhogendra Jha, Gupta Sculptures: Bharat Kala Bhavan (Varanasi: Banaras Hindu University, 1985), p. 72, no. 91, pl. 37, fig Donaldson, Śiva-Pārvatī, vol. 1, p. 296, vol. 2, p. 271, fig. 233; vol. 1, p. 367, vol. 2, p. 325, fig Muhammad Zaheer, The Temple of Bhītargāon (Delhi: Agam Kala Prakashan, 1981), p. 87, pl. 65. Based on the composition, the Bhitargaon scene probably depicts Śiva and Pārvatī playing dice. There are figures represented below the divine couple, but the terra-cotta surface is severely damaged. The only figures that can be identified are two of Śiva s dwarf hosts (gaṅas), one at each end of the panel. The one on the viewer s right is playing a mṙdaṅga drum. The right foot of an additional figure (probably a gaṅa) also remains, but given the small amount of space between the flanking gaṅas, it is highly unlikely that there would have been room for Gaṅeśa, Kumāra, and Bhṙṅgi. Iconographically, it would also be more logical for the missing figures to be one or two additional gaṅas. 28. For example, see Carmel Berkson, Wendy Doniger O Flaherty, and George Michell, Elephanta: The Cave of Shiva (Princeton: Princeton University Press, 1983), p. 34, pl See n For other possible interpretations of this scene s meaning and intended locale, see Charles Dillard Collins, The Iconography and Ritual of Śiva at Elephanta (Albany: State University of New York Press, 1988), pp , fig Joanna Gottfried Williams, The Sculpture of Mandasor, Archives of Asian Art, vol. 26 ( ), pp ; Donaldson, Śiva-Pārvatī, vol. 1, p. 152, vol. 2, p. 174, fig. 2:79; vol. 1, p. 153, vol. 2, p. 176, fig. 2: Donald M. Stadtner, Nand Chand and a Central Indian Regional Style, Artibus Asiae, vol. 43/1 2 (1981), pp , unillustrated. 33. Pal, Indian Sculpture, pp , no. S I am distinguishing herein between an iconographic comparison and a stylistic comparison. The former involves the presence of a particular object worn on the body and its structural form, or the presence of a particular posture or positional arrangement for a body part, while the latter refers to the manner in which a physical feature of the body is represented. 35. Pal, Indian Sculpture, p V. S. Agrawala, The Terracottas of Ahichchhatrā, Ancient India, Bulletin of the Archaeological Survey of India, no. 4 (July 1947 January 1948), pp. 132, S. P. Gupta, ed., Masterpieces from the National Museum Collection (New Delhi: National Museum, 1985), p. 62, no L Age d or de l Inde classique: L Empire des Gupta (exh. cat.) (Paris: Réunion des musées nationaux, 2007), p Willis, Temples of Gopakṡetra, p. 53, pl. 57. See also its unillustrated listing in S. R. Thakore, Catalogue of Sculptures in the Archaeological Museum, Gwalior, M.P. (Lashkar: Modern Printing Press, n.d.), room 10, no. 1 (b). 40. Pal, Indian Sculpture, p Muchnic, LACMA does an about-face. 42. Besides the Harle 1974 reference given by Pal, see also Devendrakumar Rajaram Patil, Saptamātṙkās or Seven Mothers from Besnagar, Proceedings volume of the twelfth session of the Indian History Congress (Cuttack: South Indian History Congress, 1949), pp ; R. C. Agrawala, Mātṙkā Reliefs in Early Indian Art, East and West, no. 21/1 2 (March June 1971), pp , 88 89, figs , 19 24; Joanna Gottfried Williams, The Art of Gupta India, p. 51, pl. 48; Michael W. Meister, Regional Variations in Mātṙkā Conventions, Artibus Asiae, vol. 47/3 4 (1986), p. 256, fig. 12; Katherine Anne Harper, Seven Hindu Goddesses of Spiritual Transformation: The Iconography of the Saptamatrikas (Lewiston, NY: Edwin Mellen Press, 1989), pp , 84 85, figs , 45 49; and Shivaji K. Panikkar, Saptamātṙkā Worship and Sculptures: An Iconological Interpretation of Conflicts and Resolutions in the Storied Brāhmanical Icons (New Delhi: D. K. Printworld, 1997), pp , pls Agrawala, Mātṙkā Reliefs, pp , unillustrated; Williams, Gupta India, p. 136, fig. 210; Pannikar, Saptamātṙkā Worship, pp , pls Willis, Temples of Gopakṡetra, p. 42, pl For example, see Susan L. Huntington and John C. Huntington, The Art of Ancient India: Buddhist, Hindu, Jain (New York: Weatherhill, 1985), p. 298, fig For example see Donaldson, Śiva-Pārvatī, vol. 2, pp. 447, 453, figs. 516, 527, Pal, Indian Sculpture, p Ibid. 49. Frederick M. Asher, The Art of Eastern India, (Minneapolis: University of Minnesota Press, 1980), p. 86, pls. 186, Michel Postel, Ear Ornaments of Ancient India, Project for Indian Cultural Studies II (Bombay: Franco- Indian Pharmaceuticals Ltd., 1989), p. 120, fig. V Waltraud Ganguly, Earring: Ornamental Identity and Beauty in India (Delhi: B. R. Publishing, 2007), p. 186; Postel, Ear Ornaments, pp , figs. A and A I am grateful to Darielle Mason for suggesting the Brooklyn Śiva Andhakāsuravadha as a comparison with the LACMA Umā-Maheśvara. Her research on the Brooklyn image is an instructive example of the recent scholarly evolution in the study of South Asian sculpture, for only six years earlier it had been attributed to Rajasthan, 9th 10th century in The Collector s Eye: The Ernest Erickson Collections at The Brooklyn Museum (exh. cat.)

26 STEPHEN MARKEL The Disputed Umā-Maheśvara in LACMA 111 (New York: The Brooklyn Museum, 1987), p. 155, no For an additional iconographic discussion of this sculpture and a color illustration of it, see Rob Linrothe and Jeff Watt, Demonic Divine: Himalayan Art and Beyond (exh. cat.) (New York: Rubin Museum of Art, and Chicago: Serindia Publications, 2004), pp. 244, 246, 297, no Desai and Mason, Gods, Guardians, and Lovers, p. 176, no See Klaus Bruhn, The Jina-Images of Deogarh (Leiden: E. J. Brill, 1969). 55. Desai and Mason, Gods, Guardians, and Lovers, pp. 123, Heeramaneck, Masterpieces, no. 51. Rochell (Sacred and Sublime) accepts and paraphrases Heeramaneck s royal couple interpretation. 57. Kramrisch, Manifestations of Shiva, p. 58, no Pal, Indian Sculpture, p Sights Unseen: Recent Acquisitions, in Treasures: The Members Magazine of the Asian Art Museum Chong-Moon Lee Center for Asian Art and Culture 9/12 (Fall 2006), p Pal, Indian Sculpture, p The Bhita image was unearthed in an excavated stratum corresponding to the early Gupta epoch (Sir John Marshall, Archaeological Survey of India, Annual Report, [Calcutta: Superintendent Government Printing, 1915], p. 34), hence the fourth-century dating used herein. Later scholars apparently did not note Marshall s reference and generally attributed it to the Gupta period (Doris Meth Srinivasan, Many Heads, Arms, and Eyes: Origin, Meaning, and Form of Multiplicity in Indian Art [Leiden: E. J. Brill, 1997], p. 266, pl ) or to the fifth century (Donaldson, Śiva-Pārvatī, vol. 1, pp. 296, 469; vol. 2, p. 447, fig. 515). 62. Srinivasan briefly mentions the Bhita sculpture in a discussion of Mathura images of Śiva and Umā standing together, but she does not stress the importance and defining characteristic of Umā being represented in the European posture; see Srinivasan, Many Hands, Arms, and Eyes, p Donaldson does, however, provide a more detailed classification system for images of Śiva and Pārvatī. Under Donaldson s system, the Mathura image of Śiva and Umā standing together that is discussed by Srinivasan would more specifically be termed an Umāsahita ([Śiva] together with Umā; Donaldson, Śiva-Pārvatī, vol. 1, pp. 296, Jitendra Nath Banerjea, The Development of Hindu Iconography (3rd ed., New Delhi: Munshiram Manoharlal, 1974), p Umāpati as a name of Śiva is used in the Skanda Purāṅa (VII ) to emphasize the union of Umā with the body of Śiva (A.B.L. Awasthi, Brahmanical Art and Iconography: Studies in Skanda Purāṅa, Part IV [ Lucknow: Kailash Prakashan, 1976], p. 171; Ganesh Vasudeo Tagare, trans., The Skanda-Purāṅa, Part XX, in Ancient Indian Tradition and Mythology, vol. 68, ed. G. P. Bhatt [ Delhi: Motilal Banarsidass, 2003], p. 628). The term is also found in the Mahābhārata and other texts, chiefly in the context of names of Śiva at various sacred locations; S. Sörensen, Index to the Names in the Mahābhārata, reprint, 1904, [ Delhi: Motilal Banarsidass, 1963], p. 209; T. A. Gopinatha Rao, Elements of Hindu Iconography, 2 vols. in four; reprint, , [ Varanasi: Indological Book House, 1971], vol. 2/1, p. 85; Jagdish Narain Tiwari, Goddess Cults in Ancient India, with Special Reference to the First Seven Centuries A.D. [ Delhi: Sundeep Prakashan, 1985], p. 87, n. 193). 65. Krishna Deva, Extensions of Gupta Art: Art and Architecture of the Pratīhāra Age, in Seminar on Indian Art History 1962, ed. Moti Chandra (New Delhi: Lalit Kala Akademi, 1962), App. B, p. 103.