All posts by drgethinrees

On Counting Coins

On 12 December Robert Bracey led a discussion organised by the Beyond Boundaries project on questions around Feudalism and money supply in sixth century India. This discussion included presentations by Gethin Rees and Daniel Balogh from the team and was one half of a piece of research work undertaken in the summer.

The other half of that work was a survey of coinage in the sixth century which was shared with the research team in advance. That covered the tri-metallic coins of the Alchons in Gandhara and the Punjab which seems to have ended in the mid-late sixth century; the small 1g local issues of Kashmir Smast, the transition from gold dinars to silver drammas in Sindh/Multan early in the century; the copper ‘Toramana’ coinage of Kashmir; and the debased gold dinar DNNVY series recently attributed to the Punjab; the silver ‘central provinces’ types issued by the Alchons, Maukharis, and then Harsha, from Kanauj; the gold dinars of the late Guptas and the ‘later’ Guptas; the gold dinars initially imitating Kushan and the Gupta types in Samatata which probably began in the late fifth century; and the silver coins of the Candras, Pyu, and Mon, issued along the coast east from Bengal whose inception different numismatists attribute anywhere from the first to the eighth century; then the silver dramma coins of Western India which include posthumous issues of Kumaragupta; Krishnaraja; the Maitrakas; and Nadol; and the copper and lead issues which seem to have accompanied them; the silver ‘Indo-Sasanian’ types of Rajasthan which seem to begin in the early sixth century and in the seventh and eighth dominate north India; the few attributed issues of the Aulikaras; and the Vakatakas; the widespread copper coins of the Vishnukundins; and the recently discovered small coins of the Chalukyas of Badami; the Pallavas; local anonymous issues of South India; and the Roman nummi and their imitations in Sri Lanka. And if that sounds like a lot then it will be understandable why the talk itself focused on a different aspect.


Silver Drachm of the Alchon Javukha, (3.53g, 21mm, BM 1894,0506.206)


Early Pratapa type coin attributed to Kanauj  (BM 1928,1018.12 7.26g, 21mm)


Pyu coin (BM 1983,0118.10)


Posthumous issue of Kumara Gupta (OR.9497 2.09g, 12mm)

How do you count those coins? A great deal of post-Gupta historical study is dominated by a theory known as ‘feudalism’. Propounded in the greatest detail by R S Sharma, he argues that after the Gupta Empire there were profound social changes. Specifically, for coins, that market exchanges (in the technical economic sense rather than the mundane physical location sense) diminished dramatically in favour of exchanges tied to land and social/political obligations. As a result there was markedly less need for money – and thus (the two are often conflated) coins.

So, assuming that you accept that coins represent the bulk of ancient money (that money of account, deposit banking, fungible loans, and commodity monies, are either modest or non-existent), how do you count the number of coins in circulation to test this? Do you count in absolute values or rates per year, the number of types, the number of hoards, coins in collections, or from archaeological sites, or the number of dies that were used to make them? And do you count the number of coins, the value they represent, or how frequently they were used?




Large gold dinar of Vainyagupta (9.37g, 21mm, BM MAR.1051) from NE India and a small copper/debased silver coin from W India, perhaps Nadol, (0.86g, 9mm, BM 1908,0702.45), what relative value in money do you assign these?

The talk tried to explore these different questions and also what sort of data you might be able to extract from the survey in order to answer the questions. One of the difficulties in this kind of research is trying to marry the detail which is where all of the useful analysis takes place with the broad sweep which is where all of the interesting questions are.



Charts showing counts of coins in collections and references to coins in inscriptions, earliest on the right.

Daniel Balogh illustrated this by talking about a fifth century inscription which has been read as involving a transaction using gold bars, but which has in fact been misread and probably does not mention money at all. The question of how you ensure large amounts of data are robust enough to answer big questions was an important part of the discussion.

Gethin Rees also spoke at the end about modelling that has been used to identify the degree of market integration in the Roman economy, and thus about how you can measure economic activity independently of coins.

This was very much research in progress, warts and all, and a somewhat experimental attempt to engage people from different disciplines with the details necessary to answer a larger problem. The presentation is now available on You-Tube with the comments section opened up to allow further discussion of the topics this link:

Data Management in Asian Humanities and Social Sciences

Our November event consisted of a two day workshop titled Data Management in Asian Humanities and Social Sciences. Speakers were invited from around Europe and North America to discuss the use of data in their projects and research.

Presenters were of three types:

  • Research assistants from the Beyond Boundaries project
  • Researchers working on collaborative projects based at SOAS and elsewhere
  • Experienced digital specialists managing web resources over the long-term

A playlist of presentation videos is available here:

The research assistants from Beyond Boundaries each outlined their research and how they use digital technologies before discussing plans for archiving their data. The data produced by the project is diverse but includes images, xml, spreadsheets, text and GIS files. is the primary repository for these data and the project is currently depositing in these communities, amongst others:

The presentations offered an opportunity for the research assistants to gain valuable feedback on the methods and processes they use for data management.

Speakers were invited from a variety of collaborative projects researching topics as diverse as multilingualism in modern Africa to the history of yoga. Although dealing with quite different subject matter, the projects encountered similar issues in the management of their data such as finding tools for organising and collaborating or ensuring that the archival formats used would have the greatest longevity. It was concluded that the difficulties in managing data effectively are often under-estimated and that forums for sharing experiences of working practices between projects like this workshop are valuable yet rare.

Finally, presentations by researchers working on the gSung-rten Database, the Endangered Languages Archive, Buddhist Digital Resource Center and the Chinese Text Project amongst others offered an insight into how data is managed at scale and over the long term. Understanding how the technical challenges facing these resources have been met was fascinating and helped stimulate discussion about the potential for the reuse and aggregation of data produced by Beyond Boundaries and other projects.


Language in Early Burma

By Marc Miyake, Beyond Boundaries Research Assistant

This month the seminar series took a break from its usual Indological focus and turned to Burma at the eastern edge of the Indic world.


Photograph by Tetsuya Kitahata, distributed under a CC BY-SA 4.0 license.

It is an unspoken assumption in the modern world that there is a strong correlation between countries and languages. Of course there are well-known cases in the Western world such as Canada and Belgium (one country with two major languages) or Germany and Austria (one language in two countries), but awareness of linguistic diversity outside the Western world is low. The average person assumes Burmese is the only language in Burma. A more informed person might have heard of the states of Burma and be aware of the major languages of Burma: e.g., Mon in Mon State, Shan in Shan State, etc. Some might even know that those languages are sometimes not even related to Burmese. Mon is related to Khmer and Shan to Thai. But even those languages are but a fraction of the 118 languages spoken in Burma today. And there may have been even more languages spoken there in the past.

The best known of those lost languages was Pyu, spoken in the first millennium CE in what is now Upper Burma and surviving only in the form of inscriptions. The name Pyu is Burmese; the Pyu autonym is not attested in the extant Pyu corpus, though Chinese historical records tell us they called themselves something like 突羅朱 dwot la tsyu in Middle Chinese transcription. Gordon Luce equated that transcription with Old Mon tircul ‘Pyu’. But no similar name has ever been found in Pyu itself.

The seminar began with an introduction by Nathan Hill on the decipherment of Pyu by Arlo Griffiths of EFEO, Julian K. Wheatley, and myself, and on his own project, the reconstruction of Proto-Burmish, the ancestor of one branch of the vast Trans-Himalayan (a.k.a. Sino-Tibetan) family.

Of course, the most famous Burmish language is Burmese, which has a largely unexplored internal diversity. Patrick McCormick gave an overview of that diversity in his talk on Burmese dialectology. Trained as both a historian and as a linguist, McCormick is not only interested in the phonetic mechanics of languages but what languages can tell us about their speakers’ histories. McCormick is not confined by traditional family tree models of language. For instance, although the Karen languages and their neighbor Burmese are in distant branches of the Trans-Himalayan family, McCormick sees the possibility of Karen influence in the Tavoyan dialect which has Karen-like traits such as the presence of glottalized consonants and the attenuation of nasalized vowels.

Hideo Sawada introduced the northern branch of the Burmish languages which are spoken in southern China as well as Burma. Some Northern Burmish languages retain a set of glottalized consonants lost in Burmese, a Southern Burmish language: e.g., they have both glottalized p’- ts’- t’- k’- and aspirated pʰ- tsʰ- tʰ- kʰ- corresponding to Burmese aspirated pʰ- sʰ- tʰ- kʰ-. They also retain final consonants lost in Burmese. Although these languages have considerable historical value, they have small numbers of speakers, one (Hpon) is nearly extinct, and some remain undescribed: e.g., Maingtha, spoken by a Shan tribe despite the fact that the Shan language is not Burmish or even Trans-Himalayan. Cases like Maingtha underscore the potential disconnect between ethnicity and language.

Tilman Frasch brought the seminar full circle back to Pyu, which he hypothesizes was the source of Old Burmese terminology for agriculture, land use, and irrigation. Frasch posited a scenario in which Burmese speakers arrived in Burma which was not terra rasa: the Pyu had already been there for centuries. The newcomers learned agricultural technology and the words needed for it from the locals. There is no doubt that Pyu is a substrate of Burmese: i.e., a language that influenced a dominant intruding language which, in this case, ultimately replaced it. There is also no doubt that Burmese has borrowings from Pyu: e.g., Written Burmese ပြည် praññ ‘country, royal city’ from Pyu pringh. (One Burmese pronunciation of the spelling praññ is pyi as in the name of the Burmese capital Nay Pyi Taw.) The first Burmese speakers in Burma did not come from an urban civilization like that of the Pyu, so it would be natural for them to borrow a word for ‘city’ – a word that is the name of Pyay, a city near the ruins of the Pyu city now known as Śrī Kṣetra. (Pyay is another modern Burmese pronunciation of Written Burmese praññ.) It would similarly also be natural for early Burmese speakers to adopt local words for agriculture.

The trouble is that no specifically agricultural terms have been identified in the Pyu corpus so far. The inscriptions contain only about 1,700 distinct syllables which may or may not be monosyllabic words. It is thus certain that only a fraction of the Pyu lexicon has been preserved in stone and metal. Only a fraction of that fraction is currently understood. There are syllables in Pyu that sound like Old Burmese agricultural words, but it is not yet clear whether they are coincidental soundalikes or the sources of those words. Frasch’s hypothesis needs to be tested by seeing if Pyu texts make sense if the soundalikes are provisionally glossed as if they were agricultural terms.

The Bactrian Archives as a Source for the History and Historical Geography of Afghanistan

This month it was a pleasure to welcome Nicholas Sims-Williams to discuss the difficulties and benefits of the archives of Bactrian documents for the study of Afghanistan in the first millennium AD. These archives surfaced together in the 1990s and consist of legal documents and letters. Properly ordered, these have the potential to shed light on the administrative centres in Bactria which produced them. They also offer an insight into Bactrian language itself, hitherto only dimly understood from a limited number of coins and inscriptions. However, a central problems is converting the dates in the documents into AD equivalents and dating those documents which lack a date. Sims-Williams took us through some of the thorniest issues and the most heated of the debates surrounding the chronology of the Bactrian archives. He pointed out some ways forward by using internal references (to the Sasanian Peroz, the Hephthalites, or Arab tax documents) that helped set parameters for the dating. He then showed how the dated documents could give clues as to the period of the undated ones. Finally, he outlined some of the ways in which these documents contribute to the history and historical geography of the area between the Oxus and Kabul, such as modes of water-use, (co-)ownership of land, inheritance, as well as providing the first mention of the Afghan people in the second half of the fifth century. These Bactrian archives offer unparalleled access to the ancient centres of Rob and  Kadagstān as different ruling dynasties gain paramount authority in the region, the Kushanshahs, the Sasanians, the Hepthalites, Western Turks, and finally the Arabs. The last stages of the archives contain both Bactrian documents and Arabic tax records relating to the same local family.

Michael Willis then followed up this talk by showing some of the Bactrian artefacts housed in the British Museum. He contextualised the fine artwork on seals and coins in order to bring up questions surrounding worship, divinity and kingship, tying the works to later Vaishnavite wall carvings and inscriptions from Udaigiri and other sites.

The Sanskritisation of Gandharan culture

This month we welcomed Ingo Strauch to speak about the Sanskritisation of central and northern India before and during the period of the Gupta empire. The focus of his talk was the greater Gandharan area, stretching from today’s Swat Valley in the north to what is now Islamabad in the south and including Bamiyan, Gilgit and Taxila in between. This area shared a common culture linked by language and script during the first millennium CE.


Original image by Asia Society. Uploaded by , published on 23 June 2015 under the following license: Creative Commons: Attribution-NonCommercial-ShareAlike. No changes made.

By the sixth century, what Pollock names the “Sanskrit cosmopolis” held sway over central India, but it affected different areas to various extents and in divergent ways. Different groups also held contrasting attitudes to Sanskrit, and these positons changed gradually over time. At the start of the millennium, Buddhists were generally hostile to the language, and Middle Indic dominated as their language, whereas Sanskrit was largely confined to ritual texts associated with Vedic literature, as propagated by the Brahmanical schools. The connected questions of why and how Buddhists adopted Sanskrit can now be investigated more thoroughly, at least in the northwest, with the benefit of important manuscript discoveries from 1994 onwards.

Strauch led us through the examples of Gāndhārī language to show that, on the level of the written word at least, biscriptualism was linked to bilingualism. Aśoka’s north-western edicts were written in Kharoṣṭhī (contrasting with his other, Brāhmī inscriptions). The evidence of the Bamiyan and Bajaur manuscript discoveries, among others, points towards the gradual adoption of Brāhmī in “Greater Gandhara,” and other evidence suggests the late-Kuṣāna period hegemony of this new script—to the extent that, even when Kharośṭhī was used, it began to reflect the language written in Brāhmī. Strauch argued for a gradual Sanskritisation (rather than a sudden conversion), in which for some time the Sanskrit of the Sarvāstivādins sat side by side with the Buddhist Sanskrit of the Mahāsāṃghikas (even in the same manuscript, for example the Bajaur Prātimokṣa sūtra). He also stressed the varieties of Sanskritisation taking place in different ways on coins, sculptures and reliquaries. Overall, though, it seems clear that the shift from Kharoṣṭhī to Brāhmī did not cause Sanskritisation, but rather that the latter was more suited to writing the complex ligatures necessary to reflect Sanskrit sounds. This pragmatic cause led to the loss of Kharoṣṭhī in “Greater Gandhara” by the end of the Gupta period, excepting for pockets of resistance in small outlying areas.

Strauch connected Sanskritisation itself to the rise of Hindus in Gandharan areas and in South Asia more generally. These groups used Sanskrit not only as a lingua franca and for administration, but also as a prestige religious language, and thus succeeded in convincing the Buddhists of “Greater Gandhara” of the necessity of accepting it as part of the spectrum of languages reflected in the Bamiyan and other manuscripts. Strauch cited Bronkhorst’s speculations concerning the reasons why the Buddhists eventually took up Sanskrit—the status of the language in mediating the important priest-patron relationship apparently led the Buddhists to use the same means to defend their own interests at the royal courts.

Michael Willis then showed some examples of similar shifts in language for the British Museum Collection, including sealings, coins and artifacts discovered at Sunet, near modern-day Ludhiana. Many of these resources will be put up on the research group’s Zenodo pages as part of the project’s important task of data collection and dissemination.

Finally, Robert Bracey discussed the coin-moulds at Sangdhol, a petty kingdom in the same area flourishing on the edge of the Gupta empire. These curious remains appear to correlate to no extant coins, problematizing Sangdhol’s description in other secondary literature as a mint site. Bracey speculated that the moulds, sometimes intended to cast multiple “coins” at once, may have fulfilled some very site-specific, folk ritual function and cast not metal but a material that has not survived in the ground.



Aspects of Royal Ideology in the Buddhist Vinayas

At our May workshop, Vincent Tournier discussed his research into the Mahāvastu and its part in the changing depiction of the royal, spiritual and buddhological lineage of Śākyamuni. The fruits of this work have recently been published as La formation du Mahāvastu et la mise en place des conceptions relatives à la carrière du bodhisattva (Paris:  École française d’Extrême-Orient, 2017). He showed how in the first millenium CE, what Peter Skilling called the “epochal career” of the historical Buddha was increasingly contextualised within the framework of other buddhas, and his family tree more and more intertwined with the very beginnings of kingship in India. This depiction of the Buddha then affected his status as an enlightened authority figure and that of his spiritual heirs, his disciples, within an increasingly crowded religious marketplace of first millennium South Asia.


Tournier first outlined the findings of his philological interrogation of the Mahāvastu, probably closed as a work in the sixth century CE and written like the Vinaya in the “Buddhist Hybrid Sanskrit” of the first-millennium Mahāsāṇgikas. Other works describe the Mahāvastu as a Vinaya text, and though it does not describe rules to be followed by the Buddhist community Tournier argues that this should cause us to broaden our conception of the Vinaya itself. The work itself attempts to line up a number of elements within the Buddha’s biography with corresponding Vinaya doctrinal categories. Yet the Mahāvastu consists of many strata, including two different prologues, dating to various points in the development of the Buddhist literary tradition. This leads to a rich but sometimes confusing mixture of themes and portrayals of the Buddha and his pre-incarnations, many of which Tournier has expertly separated and then investigated for the light they shed on the Buddha’s purported royal heritage.

The idea of seven buddhas of the past, present and future, found in much earlier Buddhist works such as the Mahāvadana, is structured as a lineage in the Mahāvastu. There, thematic elements and poetic imagery are repeated for each buddha, unifying them into a chronologically ordered set. Furthermore, the portayals of an earlier buddha pronouncing a prophecy (vyākaraṇa) for his immediate successor increasingly resemble a king’s appointment of his royal heir, for instance in Śākyamuni’s description of Maitreya’s future life as a dharmarāja. Tournier contextualised this trend with recourse to the Lalitavistara lifestory, the Avataṃsaka sūtra and sculptures found at the Borobudur temple complex. Finally, he showed how the narrative of the royal descent of the Buddha, found in the Mahāvastu, grew out of an independent aeteliologies concerning the Ikṣvāku family lineage and the cosmogonic beginnings of people and their first king, Mahāsammata. The Mahāvastu combines these two tales into an account of Śākyamuni lineage stretching from the First King down through the Ikṣvāku royal line. Tournier highlighted the similar strategies of appropriating the stories of past kings used in both creating a Buddhological lineage and also unifying the Buddha’s royal line. He suggested this perhaps reflects a Kushan and/or Gupta milieu, in which it was necessary for the Buddhists to argue that they did not merely consist of a mixture of social classes but were instead pure sons of a Buddha prophecied aeons ago, who was himself heir to an old and exalted lineage.


Lewis Doney’s Fieldtrip to Thailand

Between February the 19th and March the 16th, 2017, I travelled to Thailand to investigate the region’s long history of religious kingship, Buddhism and its attendant material culture.

I spent the first ten days in Bangkok, an area rich in temples, monasteries, monuments and lived religious practice. Examples include Wat Phra Kaew, Wat Pho, the Grand Palace and the Erawan Shrine.


Scene from the Rāmāyaṇa mural at the Grand Palace (Phra Borom Maha Ratcha Wang), Bangkok.

The time of my visit coincided with the sad period of mourning for the beloved King of Thailand. This gave me the opportunity to witness first hand both the organised Buddhist rituals of his commemoration among the civil service employees who still dress in black or white as a sign of respect, and also the preparations for the departed king’s cremation due to take place in December 2017.

During this time, I also visited the National Museum, a treasure-trove of artefacts from around Thailand dating from the prehistoric age over 6000 years ago to the Bangkok period. Of especial interest is the time of the Dvaravati city-states (6th–11th centuries CE), which borrowed Mon culture and non-Mahāyāna Buddhism from its neighbours. Numismatic and stone inscriptional evidence suggests that these were based around Nakhon Pathom, west of Bangkok. Like the earlier Guptas, this civilization maintained numerous overland trade routes—including east to what is today Cambodia, west to Burma, and north to Chiang Mai and Laos. These existed in addition to the maritime networks existing from the eighth century under the protection of the empire of Srivijaya, who controlled trade between the South China Sea and the Indian Ocean. Evidence of Dvaravati Buddhist artwork remains, and seems to show that the localised forms of originally Indic design had been interpreted through Mon and other cultures to form a unique local flavour. The Dvaravati city-states lasted until the westward expansion of the Khmer empire (9th–15th centuries) into the region during the eleventh century.


Relief depicting female musicians (Dvaravati style, 8th–9th) from Khu Bua, Ratchaburi province.

While in Bangkok, I had the chance to visit Mahidol University and to present my latest research and the ongoing work of the Beyond Boundaries project to staff and students of the Buddhist Studies international programme within the Faculty of Social Sciences and Humanities. I was also lucky enough to participate in a reading group hosted by Mattia Salvini and Harunaga Isaacson on Jñānaśrīmitra’s (fl. 975-1025) Sākārasiddhiśāstra and an opening talk by Shoryu Katsura on the Prajñāpāramitopadeśa of Ratnākaraśānti (c. 1000).

Further afield, I was able to document the archaeological sites of Ayuthaya, Sukhothai and Satchanalai to the north in Thailand. Between the thirteenth and fifteenth centuries, Sukhothai was a regional Khmer administrative centre and so lay on one of the main roads connecting the capital at Angkor along which temples in the Khmer style—Brahman and Buddhist—acted as a visible symbol of imperial power. After the Tai expelled the Khmer in the thirteenth century, King Sri Indraditya and his son Ramkhamhaeng made of Sukhothai a large regional power with dependant kings in the four directions. Satchanalai was the northernmost, and second most important of the cities in this kingdom. Lastly, Ayuthaya was the centre of power and basis for booming maritime trade during the eponymous period, from the mid-fourteenth century onwards. The site combines Khmer and local traits, since the Ayuthaya rulers conquered the Khmer in the 1430s and were connected in marriage to the closer kingdoms of Suphanburi and Lopburi and other vassal states. Unlike Sukhotahai, Ayuthaya’s rulers were given the epithet devarājā rather than dhammarājā. The region also maintained ties with China during this period.


Standing Buddha (h. 12m) at Wat Mahathat, Sukhothai archaeological site.

Finally, pursuing my research into Buddhist material culture, I was able to visit several important sites of Thai and Chinese temple bells, photograph them and record their dimensions. This will help to give an account of them during the panel titled “Bell Inscriptions across the Buddhist World” at the International Association of Buddhist Studies conference to be held in Toronto in August 2017.


Shrine and bell & drum tower at Wat Chana Songkhram, Bangkok.

Indian Philosophy Seminar

It was a pleasure to welcome to our workshop this month Prof. Larry McCrea of Cornell University, and Prof. Isabelle Ratié from the Université Sorbonne Nouvelle for a seminar on Indian philosophy. Larry McCrea talked to us about late-first millennium discussions of the opening phrases of Indian philosophical works. At stake in these debates was what constitutes the rational human agent who will decide whether or not to read such works based on certain criteria. The creation of gradually larger schema of criteria can help inter alia date certain commentaries ascribed to this period.

At first, within the Brahmanical commentarial tradition of the philosopher Kumārilabhaṭṭa (c. 650 AD), a tripartite set of criteria for judging the value of texts by their opening statement (ādivākya) was expressed, consisting of the text/topic (śāstra) itself; its purpose (prayojana); and the connection (sambandha) between the topic and its purpose. These three parts should motivate the reader to engage with the text at the very outset. These lively discussions suggest a mileu in which several authors and their works competed for readership within a busy philosophical and literary marketplace.

The eighth-century Buddhist philosophers took up this three-fold and unpacked its implicit consequences. Arcaṭa, for instance, in his Hetubinduṭīkā commentary on Dharmakīrti’s work criticized the claim that such criteria are sufficient for gaining even uncertain knowledge of the truth or falsehood of any work’s opening claim. Later, Vinītadeva’s commentary on Dharmakīrti’s Nyāyabindu added a fourth criterion, the purpose of the purpose (prayojanasyaprayojana): to learn epistemology in order to accomplish all human aims. Another commentator on the same text, Dharmottara (c. 750s), describes the fourth criterion instead as the purpose of the topic (abhidheyaprayojana), which he claims implies the first three criteria.

Finally, Jayantabhaṭṭa’s (c. 900) Nyāyamañjarī coins the term “opening statement” itself (with recourse to the above philosophers). However, Jayantabhaṭṭa broadens the scope of the opening statement, which he claims validates Nyāya epistemology / logic in order to ultimately validate the Vedas, out to apply to our criteria for deciding to follow any path. He argues, against Kumārilabhaṭṭa, that in the every-day world of choices, we can trust out “hunches” or “suspicions” (saṃśaya) without recourse to the intrinsic validity or otherwise of some claims. However, when approaching Vedic ritual the stakes are higher and the attendant sacrifices and responsibilities greater. Thus, McCrea suggests, the problem for Jayantabhaṭṭa’s philosophy is how one argues from the quotidian to the transcendent level of epistemology.

Isabelle Ratié presented her latest discoveries regarding the works of the grammarian-philosopher, Bhartṛhari (450–510). Among the works attributed to Bhartṛhari is the now lost work known as the Śabdadhātusamīkṣā or perhaps more properly the Ṣaḍdhātusamīkṣā, that is known only from later references and quotations. Bhartṛhari influenced both later Brahmanical and Buddhist philosophical traditions, but the content of this work, and so part of the content of his oevre, has long been in doubt.

It was assumed, given Bhartṛhari’s supposed concern with speech (śabda) in other works, for example the Vākyapadīya and Mahābhāṣyadīpaka commentary on Patañjali’s famous work, that this “investigation” (samīkṣā) also focused on language. However, Ratié has found evidence in quotations and criticism of his position in works by Somānanda (c. 900–950) and Utpaladeva (c. 925–975), who appear to be intimately familiar with this work, that it may instead have investigated the six elements (ṣaḍ-dhātu), most likely the five material elements of earth, fire, water, air and ether and the immaterial element of consciousness (cetanā) or self (ātman).

The surviving fragments of the Bhartṛhari’s text suggest that he attempted to unify these six elements as arising from a unitary consciousness or self in a similar way to the later Advaita Vedānta school. The existence of such an important antecedent to the Advaita Vedānta position of the eighth-century Ādi Śaṅkara has important consequences for the study of that school. The eighth-century Buddhist philosopher Śāntarakṣita and his disciple Kamalaśīla may also have been aware of it. Ratié notes, though, that the attribution of the Ṣaḍdhātusamīkṣā to Bhartṛhari is still not a certainty and so this position may not have existed as early as the fourth century. Nonetheless, Ratié’s investigation into the Bhartṛhari’s discussion of the six elements acts as a salutary reminder that modern philologists should not be too quick to discount the perspectives of medieval Indian exegetes in favour of a reductionist scepticism.