The Dhanesar Kherā Buddha in the British Museum and the ‘Politische Strukturen’ of the Gupta Kingdom in India

This article examines a bronze sculpture of the Buddha in the British Museum and offers a revised reading of its dedicatory inscription. A close examination of the paleography, combined with a study of the sculpture’s style, shows that the image belongs in the early years of the sixth century. This provides a date for the ruler Harirāja named in the record and allows us to disassociate him from the Harigupta known from early Gupta coins. In addition, the re-dating of Harirāja provides new material for understanding the historical and cultural complexities of the late Gupta kingdom.

In 1895 Vincent Smith and William Hoey published three bronze figures of the Buddha from the village of Icchawār in Bāndā district, Uttar Pradesh ( Figure 1). 1 The images were recovered from Dhanesar Kherā, an area just west of the village marked by ruins. The name comes from Dhanasir Dāī, a deity who is venerated there. He is described by Smith and Hoey as seated on a cushion with one leg drawn up, wearing a beard and a cap. A sword hangs at his side. 2 Not much more is known about Dhanasir Dāī, but the name suggests he is a local god connected with the protection of crops. The word kherā, variously written and pronounced, is widely used in this part of India to designate a place or location, and sometimes also a village. Dhanesar Kherā is, anyway, the site of an old settlement and an ancient Buddhist establishment, as shown by the Buddha images found there.
Of the three Buddhas collected by Hoey at Icchawār, one found its way to the National Museum in Bangkok, as noted by Sheila E. Hoey Middleton elsewhere in the pages of this journal. 3 The others were acquired by the British Museum and the Nelson-Atkins Museum of Art in Kansas City. The art historical position of the images has been discussed by a number of writers: that in Kansas City shows links to Gandhāra, while that in the British Museum shows debts to the school of Sārnāth. 4 The stylistic relationshipsthe main concern of old-school connoisseurshas meant that little attention has been paid to the inscription on the base of the British Museum image. The purpose of this article is to publish the inscription and explore its implications for the history of the fifth century and, in particular, the nature of political relationships in the Gupta kingdom. This will bring us back to questions of style because the regional connections betrayed by the styleboth palaeographic and artistichelp situate the object historically and culturally.
The British Museum image shows the Buddha seated in padmāsana with his hands in the teaching gesture (Figures 2, 3). As can be seen from the illustrations given here, the halo and back-plate of the image disappeared between its first publication in 1895 and its arrival in the British Museum in 1969. The back-plate of one of the other images has shared the same fate. The British Museum image shows traces of gilding in many places, and the face signs of rubbing from veneration. This has flattened the nose and lips slightly. There is a small gash on the proper right side of the uṣṇīṣa and similar cuts on the side of the proper right arm. When and how these happened is impossible to say, but the cuts appear to be ancient.
The sculpture is riveted to the pedestal and on the top moulding of the pedestal is a dedicatory inscription ( Figure 4). This inscription was published twice by D. C. Sircar based on the hand copy in Smith and Hoey. 5 Studying the original inscription and using Harry Falk's epigraphic tool Indoskript, I am able to offer a few corrections and to verify the letters that Sircar found uncertain due to the flawed nature of the old hand copy. Here, thanks are due to Dániel Balogh who took time to offer comments on the reading, and to Sam van Schaik for general guidance on Buddhist history. Smith and Hoey expressed the view that the Buddha was 'almost destitute of merit as a work of art, and is an ordinary Indian production of the conventional pattern'. 6 Considering that the 'merit' of the work is stated openly in the inscription to be 'for the benefit of all sentient beings, headed by parents', we need not be too concerned if the sculpture failed to fulfil the pre-conceived notions held by Smith and Hoey about what makes for a 'work of art' and the 'merit' such works might hold. Preoccupations of this order continue in the west and are unimportant. As for the Buddha's conventionality, the figure is certainly conventional to the extent that all Buddha figures reach back to an ideal type, as do images of the Christian saints. Within this framework, however, the British Museum's Buddha exhibits unique and interesting features both in terms of its inscription and its style.
Firstly, the inscription records that the image is a deyadharma, that is, an object that should be given (deya) because it has appropriate qualities (dharma). I have explored the meaning and historical use of this term elsewhere. 7 What is notable, as observed by Schopen, is that the word pada has been inserted into the phrase describing 'supreme knowledge' or anuttarajñāna. 8 This is a well-known idea in Buddhist thought and has been discussed, for example, by David Ruegg and Rita Langer (who offers a useful counterbalance from the Sri Lankan side). 9 However, the wording anuttarapadajñāna is indeed unusual, and only appears, as far as I am aware, in the Varṇārhavarṇastotra of Mātṛceṭa. The relevant portion of this text reads as follows: 10 anuttarapadajñāya sarvāpadapahāriṇe | apadāyānupādāya dvipadāgrāya te namaḥ || 55 Translated also into Tibetan and Chinese, this was rendered by Jens-Uwe Hartmann as: 'Dem, der die höchste Stätte kennt, der alles Unheil beseitigt, der ohne Spur ist, ohne Anhaften, dem Ersten unter den Menschen, dir (sei) Verehrung!' 11 Since the time of Winternitz, the author Mātṛceṭa is generally thought to have lived in the Kuṣāṇa period at the time of King Kaṇiṣka, thus in the second century CE. 12 His popularity in the later Gupta period is shown by Dignāga (c. 480-540 CE), who composed one verse to be placed before each of the one hundred and fifty verses of Mātṛceṭa's Śatapañcāśatka. 13 As Winternitz noted, the Chinese traveller Yijingwho travelled in India in the mid-seventh centurysays that Mātṛceṭa was a very famous poet whose hymns to the Buddha were sung far and wide: Throughout India everyone who becomes a monk is taught Mātṛceṭa's two hymns as soon as he can recite the five and ten precepts (śīla). This course is adopted by both the Mahāyāna and Hīnayāna schools. 14 On the basis of this evidence, the texts of Mātṛceṭa and the term anuttarapadajñāna cannot be taken to have specifically Mahāyāna connotations. This means that we cannot describe the British Museum image as somehow 'Mahāyāna' based on its inscription. What we can say is that the image was madeor at least inscribedin a milieu where the vocabulary of Mātṛceṭa was known. We may even speculate, quite reasonably I think, that Mātṛceṭa's works were studied at the monastery where the bronze images were found.

Harirāja and Harigupta
What we should like to know is the date of the British Museum image given the connections outlined in the preceding paragraphs. The main evidence for the date is the name Śrī Harirāja given in the inscription. This king belonged to the Gupta lineage and was married to queen Mahādevī. There are several historical possibilities. The first is that Harirāja is the same person as the King Harirāja mentioned in the Vārāṇasī copper-plate charter. This charter registers a donation by him and his wife Anantamāhadevī. 15 This seems promising, but a connection can be discounted because the record clearly states that Harirāja belonged to the Śūra lineage. 16 Moreover, the style of the writing, with square boxheaded letters, points southward to the Śarabhapurīya or Vākāṭaka kingdoms. So this plate, although belonging to the fifth century, was found outside of its first geographical context. This means we have another Harirāja from somewhere in the Deccan and he is not the Harirāja mentioned in the Buddha image inscription.
The other historical possibilities for Harirāja have been summarised and commented on by P. L. Gupta. 17 The most likely is that Harirāja is the same person as the Harigupta named in copper coins from Ahicchatra in Uttar Pradesh. P. L. Gupta has rejected the proposal that these are late Gupta coins, minted during the declining years of the dynasty. He is of the opinion that the coins cannot be placed at any distance from Kumāragupta I and Candragupta II, with whose issues they were found. As a consequence, if we identify Harirāja with Harigupta, then he must have flourished in the early fifth century. Given that the Harigupta coins bear some similarity to those of Rāmagupta, it is possible that Hāriraja's dates could be pushed back to the late fourth century. 18

Palaeographic considerations
A late fourth-or early fifth-century date for Harirāja could be supported by the palaeography of the inscription, provided the inscription offers letters that are chronologically indicative. The normal approach with palaeographic dating is to select letters from other inscriptions and describe the similarities and differences in detail. Palaeographic arguments of this kind, based on the style of the letters, are difficult to follow, especially when debates about a particular record become extended and heated. This whole manner of working and reporting has, however, been made redundant by Harry Falk's Indoskript. With this tool we are able to compare the letters in hundreds of records quickly, and instead of depending on subjective descriptions we can refer directly to the actual shapes. The relevant records can be referred to by their descriptive name and by the Indoskript number. If needed, the letters can be checked against the original publications. The result is a more comprehensive and reliable assessment of the palaeographic style. With the Dhanesar Kherā inscription before us, we can select some letters using this approach (Figure 3).
This letter is normally written like it is in modern script, even in Samudragupta's inscriptions. Here, however, the vowel is reduced to a simple horizontal stroke, a somewhat rare treatment. The earliest parallel is the Cāṅgu Nārāyaṇa Inscription of Manadadeva (Indoskript no. 158, c. 464 CE). It also appears in the Katni Plates of the Uccakalpa ruler Jayanātha (Indoskript no. 660, Gupta year 182, 501-02 CE) and the Bodhgayā Inscription of Mahānāman (Indoskript no. 15, c. 588 CE).

Line 1: the letter 'gu' in gupta
The first appearance of this letter, with the 'u' drawn as it is in modern writing, is found in the Sārnāth Image Inscription (Indoscript no. 802, of unsettled date but placed by Hargreaves at the end of the fifth century). Thereafter it appears with some (but not absolute) consistency from the early sixth century, as in the

Line 2: the letter 'jñā' in jñāna
This presents the same problem as the simple letter 'ja' mentioned above, the closest parallel being in the Horiuzi Palm Leaf Manuscript. The way this is written appears to be unique, probably because the space available was not sufficient.
What can we conclude chronologically from these comparisons? It is impossible to be precise because of the fragmentary nature of the data. However, when a particular letter shape is documented in, for example, an inscription of c. 500 CE, we can safely conclude it was in general use at that time and earlier instances may be found eventually. On the other hand, a number of shapes from c. 400 CE are well documented, so the appearance of a new type of letter does mean something historically. To put the matter in a slightly different way, while individual comparisons are not conclusive, the observations as a whole are not lacking in historical force. The observations made above thus suggest that the British Museum inscription belongs somewhere in the sixth century, probably the first half. Broadly speaking, the most useful parallel can be found in the Katni charter, the first leaf of which is illustrated here in Figure 5. 19 It should be noted that although this is called the Katni copper-plate, it was found Uchahara, the ancient Uchchkalpa, in Satna district, Madhya Pradesh. This charter is, anyway, dated Gupta year 182 (501-02 CE). The date and comparison means that the Harirāja in the British Museum inscription cannot be the Harigupta of the copper coins from Ahicchatra: they are at least a century apart. With that much secure, we can safely set aside the suggestion that the Harirāja of the British Museum inscription might be Govindagupta, one of the sons of Candragupta II. 20 This has some implications in 5 theoretical terms. Historians have long disliked freefloating rulers and attempted to writer neater and simpler histories. This was noted many years ago by Henige in an important article that attempted to challenge practices in history writing. 21 When we get away from this tendency to conflate sources, we arrive at a more complex picturepotentially unresolved at many pointsbut a picture, nonetheless, that will accommodate new data and allow new avenues of interpretation.
Art style, Gupta dates, and the 'politische Strukturen' To round off my analysis, I turn to the image itself in order to determine if its style conforms to the date indicated by the inscription. This is a matter of some interest because there is always a possibility that the piece is composite, i.e. that the Buddha was riveted to the plinth at a time later than the inscription. The question can be addressed through an analysis of the style. The diaphanous treatment of the Buddha's robe without folds of any kindhas long been recognised as characteristic of the 'Sārnāth School'. However, so little work has been done on this school that not much can be said about the geographical extent of the idiom in South Asia, especially before the sixth century. In other words, it is difficult to say if the British Museum sculpture was made locally at Icchāwar or brought from elsewhere. Chronology is a related question. This is not the place for a detailed history of the Sārnāth School, but we can, at least, make a few relevant points based on J. C. Harle's landmark study of Gupta art. This established a history of the style, and his conclusions are generally accepted. 22 In the late fourth century, to begin, sculpture showed many debts to the late Kuṣāṇa idiom, but by the early fifth century a vigorous new style had emerged. This new style is best exemplified by the monumental rock reliefs at Udayagiri in central India. Parallels to the Udayagiri material are found elsewhere in north India. In the eastern plains, notable examples are the Govardhana from Vārāṇāsī (now BKB collection) and the Buddha from Sārnāth (now National Museum of India). With the mid-fifth century a mature idiom appears, well-documented by dated sculptures from Mathurā and Mankuwār ( Figure 6). The date of the Mankuwār Buddha is Gupta year 109 rather than 129, as pointed out by D. C. Sircar some years ago. 23 This is readily confirmed by an examination of photographs and rubbings of the inscription.

Translation
Om. Reverence to the Buddhas! This image of the Lord who thoroughly attained perfect knowledge (and) whose tenets are consistent, 24 has been installed by the Bhikṣu Buddhamitra. Year 100 (and) 09 in the reign of mahārāja śrī Kumāragupta; the month Jyeṣṭha, day 10 8, with the object of averting all unhappiness.
Before continuing, it is worth commenting on the details of the date given in this record, and on Gupta dates in general. The month Jyeṣṭha is given with the day indicated by the abbreviation 'di' for divasa (i.e. day). The number eighteen refers to the day in the solar month. There are generally between twenty-nine and thirty-two days in a solar month and they are counted from one to twenty-nine, thirty, thirty-one, or thirty-two, without reference to a tithi or lunar day. 25 As for the year itself, D. C. Sircar, summarising scholarly opinion, noted that the Gupta era started on Caitra śudi 1 and that the months of the Gupta calendar are pūrṇimānta. This means 'ending with the full moon', i.e. that calculations begin with the new moon. 26 The Parivrājaka plates from central India show that the pūrṇimānta system was used for Gupta years, while the Sārnāth Buddha inscription of Gupta year 157 shows the years were expired. The problem and its historiography have been reviewed with skill by P. L. Gupta. 27 J. F. Fleet's view that the Gupta calendar began on or about 9 March 319 is therefore correct because the new moon fell on that day or the night before. The full moon before that occurred on 21 February 319. 28 So after a great deal of fuss and spilled ink, the old master is shown to be right once again! In terms of determining the year of Gupta inscriptions, this means we can take the Gupta year and add 319, provided the month comes before Pauṣa, the winter month that falls in December or January. After that we need to add 320. For the Mankuwār inscription, this means the image was dedicated in the summer of 428 CE. As noted by Harle, the British Museum's standing Buddha from Sārnāth belongs with the Mankuwār Buddha in the middle part of the fifth century (Figure 7).
Gupta art in its mature phaseas it is often termed is documented by the famous images in the Sārnāth Museum with the dates equivalent to 473 and 476 CE. The first of these is on the pedestal of a Buddha image and was dedicated when Kumāragupta was 'protecting the earth' in Gupta year 154 in the month of Jyeṣṭha (i.e. 473 CE in about the month on May). 29 The second inscription is dated 157 in the month of Vaiśākha when the asterism of Mūla was visible. 30 This corresponds to 476 CE in about the month of April. It should be noted that the suggestions about the dates made by John Rosenfield in his landmark article on the Sārnāth School in 1963 need to be updated in the light of the developments in our understanding of the Gupta calendar outlined in the preceding paragraph. 31 The Sārnāth sculptures have been published many times and do not need to be illustrated here. 32 The style is characterised by extreme refinement and abstract serenity. Moved by aesthetic emotion with a dash of religious fervour, art historians have vied with each other in a long series of descriptive rhapsodies that have become increasingly exaggerated and not a little absurd with the passage of time. Sooner or later a detached attitude about this important style will have to be arrived at given its continental impacts in the sixth century and beyond. How the style was developing after the last quarter of the fifth century is shown by a standing Buddha from Sārnāth, also now in the British Museum ( Figure 8). This is closely related to the dated images of 473 and 476, but is more attenuated, with the sway of the body exaggerated and the forms beginning to show slight signs of distortion. Especially notable are the ways in which the folds of the robe on the side of the figure are replicated in a more mechanical way compared to earlier examples. This is a harbinger of the sixth century, when ballooning distortions become plainly evident. This is usefully exemplified in one of the bronzes from Buddham in Andhra Pradesh ( Figure 9). 33 This breakdown of the Sārnāth canon -I believe it is accurate to describe the process in this way has parallels in different parts of India, as evidenced by several stone sculptures at Sānchī. A more settled style appears in the seventh century ( Figure 10). The example illustrated here shows how the distortions of the sixth century were brought under control and how the forms were filled with voluminous energy, with the fall of the robe and the hands and feet given a mellifluous outline.
While much needs to be done to refine this chronological outline and add the necessary understanding of regional variation, what has been presented here is enough to date the Dhanesvar Kherā image. The treatment of the figure presupposes the Sārnāth idiom, as shown especially by the chest and the robe covering it. The narrowness of the torso conforms to what we would expect in the late fifth century, but the slightly larger head and hands, and the conspicuous end of the robe in Buddha's left hand, point to a date in the early part of the sixth century. The flattened uṣṇīṣa also suggests that the figure is moving away from the conventions of the late fifth century. This latter feature may reflect a regional rather than chronological difference.
Given that elements of the palaeography point southward to the Deccan, the Dhanesvar Kherā bronze may, in fact, be from that area rather than Sārnāth. Such a connection would hardly be surprising given that the other Buddha from the sitethe one now in Kansas Citybetrays connections to Gandhāra, and was evidently imported from there. We can, therefore, easily accommodate the idea that the British Museum bronze came from the south in a Buddhist world with well developed inter-regional and international networks. In terms of the date, anyway, the Buddha figure belongs with the pedestal and its inscription, so Harirāja can be fixed to the early fifth century.
The foregoing points prompt a few theoretical observations about the history and organisation of the Gupta kingdom. The most recent book to address this issue was published under the title Politische Strukturen im Guptareich. 34 The book returns to an assessment of the primary sources and attempts to draw conclusions on that basis. This is laudable but, as Oskar von Hinüber has noted, the scope of sources has expanded considerably since the 1970s, when P. L. Gupta wrote his fundamental and still valuable survey. The sources have especially expanded in terms of Buddhist textual material preserved in Tibetan. 35 In addition, von Hinüber shows through close analysis of several examples that almost all Gupta period inscriptions need re-examination to establish a better reading and determine a more precise (and sometime more problematic) understanding of the contents. What he is pointing to is the fact that Virkus's book undertakes almost nothing new in terms of critical reassessment. Von Hinüber is rightly impatient with 'the brilliance of shiny theories'. To put the matter another way, we will be left with nothing but rusty old theories if we simply re-cycle established reading. This is central to the historical project because inscriptions are key to the structure of the Gupta kingdom and its history. I suppose it is especially telling that the Harirāja of the inscription taken up in this article is not mentioned by Virkus. What the inscription demonstrates is that, even as the Gupta Empire disintegrated and was overtaken by the Maukharis, Gupta princes were able to hold on to their ancestral lands. 36 This shows, I hope, the importance at looking at records long known, and of thinking about them again using the new methods and tools available.