Kumārila on deities

Did Kumārila believe in the language-independent existence of deities? In their efficacy within sacrifices? I believe he did not. Sacrifices work independently of deities who at most might be Epicurean-like entities, with no function in human lives. For this purpose, I am going to examine a passage in Kumārila’s Ṭupṭīkā ad 9.1, adhikaraṇa 4, p. 1652ff (Śubbāśāstrī 1929 edition).

The whole passage starts because Śabara is discussing the role of the deity in sacrifice. He explains that the deity does not promote the sacrifice (na devatā prayojikā and na devatāprayuktāḥ pravartiṣyāmahe). An opponent mentions the fact that the deity is mentioned as the target (sampradāna) of the action (given that it is in the dative case). Śabara quickly dismisses the point as related to the fact that the sacrifice is the real instrument to realise the result and moves on. Kumārila elaborates on it and comes to the conclusion that the deity has a primary role because of its grammatical function, which makes it needed for the performance of the sacrifice.

“[Obj:] But the deity is mentioned as the target (sampradāna) [and hence needs to be the one that prompts one to sacrifice, like the teacher to whom one gives a gift prompts one to give it].

[R:] The sacrifice, which is enjoined with regard to a result. requires a performance. And this performance is not possible without a deity and a sacrificial substance (to be offered). And the substance and the deity which are required (by the performance) are required only on the side of the complement, not as the thing to be realised [by the sacrifice] (which is the real motivator). Among the two, the substance becomes a complement through the third-case ending. The deity through a suffix or through the fourth-case ending.”

Here Pārthasārathi’s Tantraratna adds an intermediate objection explaining that this makes the deity seemingly into the principal element. The response is that this is not a problem.

“And if this deity-complement did not reach the condition of being primary with regard to the sacrifice (as required by the dative ending expressing the target), the sacrifice would not be performed at all. And without the performance of the sacrifice, there would not be the complement either. And if the sacrifice did not reach the condition of being its (complement’s) secondary element (guṇa), it would not come into being.

Therefore the sacrifice needs to reach the condition of being secondary as something/through an activity which is unavoidably concomitant (to the grammatical form used). And also the deity-complement [needs to reach] the condition of being primary as something unavoidably concomitant (to the grammatical form used).”

The addition of “unavoidably concomitant” (nāntarīiyaka) may seem puzzling, since neither Kumārila nor Pārthasārathi explain it. Clooney 1997 makes the bold move to interpret it as na antarīyaka (not… intermediary). I can see his motivation, but antarīyaka is not a Sanskrit word I am aware of. (NB: The L manuscript and the Tantraratna read tasmād yāgena nāntarīyakeṇa vyāpāreṇa guṇabhāvaḥ pratipattavyaḥ (instead of tasmād yāgena nāntarīyako guṇabhāvaḥ pratipattavyaḥ)).

At this point, Pārthasārathi’s Tantraratna adds that this does not mean that the sacrifice has become because of that for the sake of the deity. The following explains why.

“Nor is it the case that the unavoidable functioning is the cause for being primary or secondary, since it is not what is enjoined (and only the injunction determines real primary status). For, [one thinks:] “Since I have been enjoined towards the result, I realise the result through the sacrifice, not otherwise” (and this shows how the injunction puts the result as the primary thing and the sacrifice as its instrument).”

The Tantraratna has an interesting variant here, namely yāgo hi phale coditaḥ. This fits the beginning of the passage, which also read yāgaḥ phale codito…. The translation, in this case, would be as follows: “The sacrifice, being enjoined with regard to the result, does not realise the result otherwise [but through the deity as target]”.

“In this way, the unavoidable functioning needs to be the secondary element with respect to the deity.”

PSM adds: “in order to realise the sacrifice’s being an instrument towards the result”. It also specifies: “the secondary status is not enjoined”.

“Hence, in all cases there are two parts (a principal one, and a secondary one). Among them, we need to understand which one is what is wished to be expressed and which one is not. Among these two possibilities, in worldly experience what is wished to be expressed is determined by the force of things. In the Veda, by contrast, by language. And through language the sacrifice is the primary element, because it realises the result, given that it is proximate [to the result], The deity, by contrast, is understood to be the secondary element. Nor is it the case that the subordination (śeṣatva) is characterised as being an auxiliary (upakāra). Rather, it is established that it is characterised by the Vedic injunction.”

PSM explains that the “Nor…” sentence is the cause of the previous one.

“[UP:] What is the purpose of this investigation?

[R:] If the sacrifice had the purpose of gladdening the deity, then the deity were the one to be worshipped and the sacrifice would be a worship. And the worship is a thing known in worldly experience. Within a worship, what would be the confidence [one could have] in the claim that Sūrya is worshipped in the same way as Agni? The very opposite might be the case (namely that Sūrya dislikes what Agni likes). (Hence, the ectype for Sūrya should not be performed as the archetype sacrifice for Agni!)”

Thus, Kumārila concludes, the fact that the material trumps the deity when it comes to determining the procedure to be followed shows that the sacrifice is not a worship aimed at the deity and that the deity’s seeming predominant role is due to grammar only. I don’t see any important difference between Kumārila’s and Śabara’s conclusions here.

On LLMs, publishing houses and our volunteer work for them

I will not be able to take part in any new project hosted by publishing houses that are ready to send my work to LLMs (I have a few ongoing and will conclude them). Allow me to explain why.

I am deeply concerned by the LLMs being a big risk for the environment, our students’ mental health and deskilling as well as their being based on intellectual theft. Thus, I will not volunteer my time and energy to help publishers that will then give my work to feed LLMs.

I asked various publishing houses about their politics with regard to LLMs and received (disappointing) answers on how “LLMs are the future”, “LLMs are inevitable” etc., all leading to the same conclusion, namely that I cannot opt out from my work being used to feed them. Such being the case, I am sorry to say that I prefer to pass.

I apologise for not being able to help the various editors who asked me to contribute or peer review for their volumes, but time is limited and I prefer to volunteer my time to help publishers who have higher standards. If a publisher wants to just focus only on profit, they should start paying their contributors, editors, peer reviewers… I know that my decision will not change anything (alternative peer-reviewers or contributors will be found etc.), but perhaps if enough people were to refuse working for free for publishers that comply to LLMs’ demands, then some change could be achieved.

UPDATE:
—Journals and publishing houses that have answered that LLMs are unavoidable etc.: CUP (author can opt out, but not in the case of open-access publications), OUP (basically, LLMs are the future, like google search is the present), Springer Nature (“peer review reports and unpublished manuscripts are not used for training LLMs, while accepted articles are”), Taylor and Francis (“In terms of licensing we do permit some trusted partners to use specific content for the purposes of training AI. We feel this is important as Publishers need to engage with these companies and control the use of content – for instance making sure it is used appropriately, within licence terms, with authors being fully attributed for their work, and to be paid for this use where contracts specify royalties. If we do not do this there is a real risk that these firms will simply access our content without permission and try to establish this as a form of fair use. In fact this has already happened with one major firm doing precisely this.”).
—Journals and publishing houses that have answered that they offer authors the option to opt out of LLMs, but cannot guarantee that it will be respected: Brill
—Journal and publishing houses that have answered that they don’t feed our work to LLMs: University of Hawai’i press (Philosophy East and West)

Cognition of the self

How does one know about the self, according to the three main schools discussed in my last post?

Buddhist Epistemological School (Dharmakīrti): the self does not exist. The only thing that exists is a stream (santāna) of causally linked momentary cognitions. Cognitions are self-aware of themselves qua cognitions (svasaṃvedana). This is not contradictory, because each cognition has a perceiver and a perceived aspect (grāhaka and grāhya-ākāra respectively).
Nyāya: the self is known only through inference (Vātsyāyana, Jayanta); it is known also through perception (Uddyotakara, Udayana)*
Bhāṭṭa Mīmāṃsā (Kumārila): we have direct access to our self through ahampratyaya `cognition of the I’. No need to infer it, since perception trumps inference and Mīmāṃsā authors require novelty as a criterion for knowledge, so that repeating what is already known through ahampratyaya would not count as knowledge.

The first Nyāya position might lead to problems if connected with the acceptance of yogic perception. Yogins can indeed perceive the self, according to all Naiyāyikas. Why not all other beings, given that perception requires a conjunction of self+manas+sense faculties, that the self is pervasive (vibhu) and that spatial limits are not needed for perception, as shown by the case of absence? Jayanta explains that the self is partless and that a partless thing cannot simultaneously be perceiver and perceived (cf. Kumārila’s argument against the Buddhist idea of cognitions’ having a perceiver and a perceived aspect and Kumārila’s claiming that this does not apply to the self, which is complex and not partless).

The Mīmāṃsā position requires the joint work of intrinsic validity and falsification: some I-cognitions are not about the ātman, since they are indeed falsified (e.g., “I am thin”, which only refers to the body).
Other I-cognitions are not, e.g., cognising ourselves qua knowers and recognising ourselves as the same knower who knew something in the past.

*I am grateful to Alex Watson for discussing the topic with me per email, on top of his decades of work on the topic!

Intro to Sanskrit philosophy

Background: This year I taught again a class on Sanskrit philosophy (for the first time since 2021). I only had 12 meetings, of three hours each, hence I had do made drastic choices. The following is the result of these choices (alternative choices could have been possible, e.g., focusing on the Upaniṣads and their commentaries). Comments, as usual welcome!

There is a time within Sanskrit philosophy, approximately around 500 to 1000 CE, without which all later discussions do not make sense (whereas one can understand later discussions without referring to, e.g., the Brāhmaṇas, the Pāli canon etc.).
I am thinking of this core of Sanskrit philosophy as the period of time in which philosophers interacted with each other in a dialectical way, learning from each other and being compelled by each other’s points. In other words, as the time in which philosophy was constrained by the need to give reasons for each claim. In this sense, I am not focusing on the Pāli Canon or on the Upaniṣads.

At the core of this period lies the interaction between three schools, namely Mīmāṃsā, Nyāya and Buddhist epistemological school. No matter the topic, the interaction among these three is always at the center and always needs to be taken into account. According to the various topics, further schools might need to be taken into account. For instance, discussions about atomism will need to take into account the Vaiśeṣika school, discussions about language need to take into account the Vyākaraṇa school.

At the center of this core moment are discussions about epistemology and philosophy of language. It is interesting to note that ontology does not necessarily logically precede epistemology and that the opposite can be the case, especially in the case of Mīmāṃsā. This is particularly evident in the case of discussions about prāmāṇya `validity’.

Sanskrit philosophy developed through debates among thinkers commenting and responding to each other. In this way, they showed that ‘novelty’ is overestimated as a criterion to assess philosophical value and its consistent presence among the criteria reviewers of grants and projects are asked to assess is more the result of a fashion than of inner-philosophical reasons.

This does not mean that individual authors did not deliver substantial contribution to philosophy. Philosophy develops through its history and its history is made by individual thinkers. Nonetheless, these individual thinkers contribute under the garb of a school, downplaying their disagreements with their predecessors and often enveloping them within a commentary on a predecessor’s text, which is meant not just to explain it, but also to enfold all its potential meaning. Some scholars did move from one school to the other (e.g., possibly Vasubandhu or Maṇḍana), others just introduced in one school the elements of the other school they more strongly agreed with (e.g., Jayanta).

Key authors to be kept in mind:
• Dignāga (Buddhist epistemological school), introduced the threefold check, later accepted by all thinkers
• Kumārila (Mīmāṃsā), introduced the concept of intrinsic validity, explained that cognitions are not self-aware, challenged the Dignāga framework, systematised the discussions about absence and the other sources of knowledge (found already in his predecessor, Śabara).
• Dharmakīrti (Buddhist epistemological school), younger contemporary of Kumārila, adjusted the apoha theory and several other epistemological points in the light of Kumārila’s cricitism.
• Jayanta (Nyāya), modified the Nyāya epistemology in the light of Kumārila’s criticism, explained that cognitions are intrinsically doubtful, unless proven right, but that this does not lead to a paralysis, because one can act based on doubt.

dharma and adharma are not “virtue” and “vice” (in Medhātithi)

G. Jhā was an amazing scholar and translator, but he produced so much that he could not revise in detail his translation choices and some infelicities are kept in the published versions of his translations.

One such cases is “sinful” or “vicious” for adharma and “moral” or “virtuous” for dharma in the translation of Medhātithi’s commentary. I am sure that there are cases in which such a translation could make sense, but not in Medhātithi.

Medhātithi follows the Mīmāṃsā approach and defines dharma as what is prescribed by the Veda or is in line with the duties prescribed by the Veda and adharma as its opposite. Translations such as “sinful” or “virtuous” suggest that actions have an intrinsic moral value logically prior to the commands applied to them. By contrast, this is not the case. Violence is not adharma because it is intrinsically “sinful” or “vicious” and in fact Medhātithi explains that the Jyotiṣṭoma violence is not adharma at all.

Punishing? Yes, but not through violent means (acc. to Medhātithi)

The reasoning by Medhātithi on 8.316 (on punishment and why it does not violate the prohibition to perform any violence) is quite complex and multiple opponents discuss. I will list them below attributing a number to each one of them.

1. A first opinion being discussed is that, since punishment is enjoined (presumably: as part of the duty to protect people) and prohibited (by the prohibition to inflict any violence), it is optional. “Option” (or “free picking”, vikalpa) is in general to be avoided and Mīmāṃsā authors would accept it only for details bearing no great significance. Hence, the opinion is dismissed.

2. Next, a second speaker observes that no vikalpa applies, because violent punishment is clearly forbidden by the prohibition to perform any violence.

3. Rebuttal: It is not forbidden because it is performed as part of fulfilling one’s duty.

2a. Second speaker again: in order for the prohibition to perform violence not to apply, the act of (violent) punishment should be enjoined, as it happens in the case of the Agnīṣomīya (a sacrifice which involves violence, and which is not blocked by the prohibition to inflict violence).

3a. Violent punishment is enjoined through worldly experience, hence no prohibition affects it.

2b.The second speaker then expands on the above point and observes that there is no specific prescription about violent punishment, since punishment is only derived from the duty to protect people, but this duty could be fulfilled also through other means (e.g., reprimanding people). Hence, inflicting violent punishment is not enjoined as a duty and it is just a worldly act (and, as such, liable to be blocked by the prohibition). Similarly, if one were to perform a sacrifice leading to a violent result because one desires it, as it happens in the case of the Śyena, that sacrifice would be blocked by the prohibition to perform violence.

2c. The second speaker also belabours on the difference between instrumental and resultative violence, saying that in the case of the śyena violence is prompted by desire, whereas in the case of the Agnīṣomīya, it is merely subsidiary.

2d. If violent punishment were specifically prescribed, then we would have to resort to vikalpa (as discussed above).

Thus, the conclusion appears to be that violent punishment should not be resorted to as a consequence of the duty to protect, although (as discussed in my article on corporal punishment) it can be resorted to in order to purify the culprit.

Thinking about Johannes Bronkhorst (UPDATED)

On May 15, Harry Falk announced on the Indology mailing list that Johannes Bronkhorst had “left this world”. In the following weeks the mailing list (and, I am sure, other online forums) has been virtually monopolised by people remembering the man and his endless contributions to Sanskrit studies and connected fields. In fact, Johannes has been extremely prolific (Greater Magadha was written in just one semester!) and his contributions have been impactful with almost no comparison.

He had studied first mathematics and physics and then moved to studying Sanskrit in India, Pune. In a recent interview with Vincent Eltschinger (on April 21 2025) he commented the choice to travel to India as due to his desire not to serve as a soldier —a choice which was deeply important to him. But, whatever the initial motivation, his years-long stay in India was meaningful and influential for his life and he never grew out of his fascination for Indian thought.

The fact that he started studying Sanskrit while in India is key to understand the role of Vyākaraṇa in his first many decades of work, given that Vyākaraṇa (or Sanskrit linguistics) is still studied and lively engaged with in contemporary India in general and in Pune in particular. Vyākaraṇa demands deep and almost complete dedication because of its technical character. One needs to know by heart or at least to be able to navigate all the 4000 aphorisms of Pāṇini’s seminal work for the school, together with their punctual glosses by Kātyāyana and the commentary by Patañjali, and this before even being able to open one’s mouth in a symposium of Vaiyākaraṇas. Bronkhorst has been able to contribute to this very technical field, especially to its perhaps most original thinker, Bhartṛhari, but without being swallowed up by the labyrinth of Vyākaraṇa. In contrast, he learnt from its method and contents, but retained his untameable intellectual curiosity.

For scholars of Bhartṛhari, Bronkhorst’s articles are indispensable. But even the ones among of us who never specialised on Bhartṛhari have probably been influenced by Bronkhorst and by his unique blend of thought-provoking ideas and thorough knowledge of the sources. In fact, Bronkhorst was an avid and fast reader, who read hundreds of pages of both Sanskrit scholarship and contemporary, mainly scientific, papers. His ideas looked at first sight almost too thought-provoking, almost like balons d’essay (trial balloons). However, when one tried to refute them, one was forced to see that Bronkhorst knew the Sanskrit sources of the relevant period thoroughly and that his bold ideas were in fact also well-grounded. (Apologies for not discussing here whether they were also ultimately right and completely so. I want to focus more on what we can learn from him than on correcting the occasional typos or on disagreeing with specific points.)

For instance, in May 2021 Dominik Wujastyk organised a (virtual) conference on the topic of Johannes Bronkhorst’s Greater Magadha (2007), which possibly remains his most influential book. Bronkhorst himself had been invited as a respondent for talks which all engaged with his hypothesis. I was only in the audience, but was astonished to see how, almost twenty years after the book’s composition, Bronkhorst was still able to discuss each of its aspects and to respond (again, I will let to others to assess whether successfully) to each criticism raised by the speakers, through precise references to the epics and/or to Vedic texts.

Let me know enter into some details about a few of Johannes Bronkhorst’s contributions. Again, let me emphasise that there are too many to discuss even a significant percentage of them and that therefore the choice will be partly whimsical. I will focus on

  • a) The sceptical Johannes Bronkhorst looking at the development of Sanskrit philosophy: The Greater Magadha hypothesis, the “discovery of dialogue” and its significance for the history of Sanskrit philosophy
  • b) The sceptical Johannes Bronkhorst looking at the role of authors in Sanskrit philosophy: his hypothesis about a unitary Yogaśāstra and dis-unitary Mīmāṃsāsūtra and its importance for how we assess Sanskrit aphoristic texts
  • c) His hypothesis about a radical difference between Sanskrit thought and European thought
  • d) His general sceptical-scientific methodology
  • a) Greater Magadha is one of those books about which we remember a moment before and a moment after. Before the book, scholars and lay people alike took it for granted that there was a single line of development within Indian though and that since the Buddha and his thought postdated early Vedic texts by centuries, these needed to contain the seeds which would have later led to the development of Buddhist thought. The texts which were conceptually closer to ancient Buddhism, namely the Upaniṣads were therefore dated to before the Buddha.

    The Greater Magadha takes the opposite point of view and looks at the evidence available with fresh eyes and notices that they are less uniform than we might think. They thus point to a different line of development, one in which there were different roots for Indian culture, which developed in parallel and not just a single line. On the West, the brāhmaṇic culture produced the Vedic texts. On the East of the Indian subcontinent, around Magadha, the culture he provisionally called “śramaṇic” produced Jainism and Buddhism, as well as key ideas that were later absorbed in the Brahmanic fold, such as karman and rebirth. By the way, the presence of an Eastern border for the Brahmanical culture is also attested by Patañjali’s definition of Āryavarta, which has an Eastern boundary (unlike Manu’s description of the same, only a few centuries later).

    The Greater Magadha can explain why karman and rebirth make a sudden entry in the Upaniṣads although they are virtually absent from the preceding Vedic texts. They enter the Brahmanical culture so well-developed and all at once because they had been elaborated for centuries outside of the Brahmanical culture. If Bronkhorst is right, one can stop looking for faint traces of possible forerunners of karman and rebirth in the Vedic Saṃhitās and start focusing on how the theory was already developed in Buddhist texts and then imported into the Upaniṣads. One can also invert the chronology of the Upaniṣads, which post-date the encounter with śramaṇic culture (this does not mean that they need to postdate the life of Siddhartha Gautama, since he was only one exponent of that culture, as is clear through the parallel of Jainism). The same applies to the claim that “Yoga” was practiced by the Buddha. In contrast, the similarities between the PYŚ and the Buddha’s teachings should be. according to Bronkhorst, interpreted as an influence of Buddhism into Yoga.

    Although I am here mainly focusing on philosophical issues, let me emphasise again that Bronkhorst’s reconstruction is extremely detailed and covers also aspects like the different funerary practices (round stūpas in the East vs. quadrilateral moulds in the West), the approach to medicine and the conception of a cyclical time, as well as the opposition between a urban (Magadha) and rural (brahmanical) culture. Last, it has the advantage of providing a methodology to identify what is original in the teaching of the Buddha and to explain why asceticism is both endorsed in the Pāli canon and criticised by the Buddha (it was part of his cultural milieu).

  • a2) Distinguishing communities and not looking for historical links when they are virtually absent was at the basis of another of Bronkhorst’s contributions, namely the idea that the roots of Indian dialectics should be placed in the Buddhist communities in the northwest of the Indian subcontinent (which might have been influenced by the Greek tradition of public debate in the Indo-Bactrian kingdoms) and that it was useless to consider Upaniṣadic dialogues as the forerunners of the dialectical engagements which became standard in Sanskrit philosophy. Upaniṣadic dialogues are just something different (closer to the instruction by a wise person).
  • b) Bronkhorst was (to my knowledge, as always) the first one to propose the idea of a unitary composition for what is known as the Yogasūtra and the Yogabhāṣya He spoke accordingly of a unitary Yogaśāstra. Like in the previous case, the idea is mind-blowing. Up to that point, many scholars had tried to reconstruct the worldview of the Yogasūtra as divided from the Yogabhāṣya and the Sāṅkhya intervention of the latter. If Bronkhorst’s hypothesis is correct, by contrast, the division into sūtra ‘aphorism’ and bhāṣya ‘commentary’ is only a polarity within a single text. This explains what could have otherwise been considered an anomaly, like the complete absence of an autonomous transmission of the Yogasūtra. Like in the Greater Magadha case, one could find alternative explanations, but Bronkhorst’s hypothesis has the advantage of showing a possibility for streamlining explanations and avoiding unnecessary additional steps (in Sanskrit, one would call that kalpanāgaurava). I should add in this connection that Bronkhorst’s hypothesis was presented in just an article (1985), but has thereafter been embraced by Philipp Maas (see especially Maas 2006 and Maas 2013) who found many evidences corroborating it, from manuscripts to the syntax of the sūtra-bhāṣya connecting links.
  • b2) A similar case is that of the relation between the so-called Pūrva and Uttara Mīmāṃsā Sūtra, also known as Mīmāṃsā Sūtra and Brahma Sūtra. Authors before Bronkhorst had discussed their relation and chronology, Bronkhorst (2007) suggested that the latter imitates the style of the former, though not emerging from the same exegetical milieu.
  • c) In the occasion of Ernst Steinkellner’s retirement, a symposium on the topic “Denkt Asien anders?” (Does Asia think differently?) was organised. Bronkhorst’s intervention led to a later book chapter and finally a book on the topic of what is different in Sanskrit thought. Bronkhorst proposed, as usual, a thought-provoking thesis, namely that there is indeed a radical difference, namely the reliance on language by Sanskrit philosophers.
    He explained how the various causation theories within Sanskrit philosophy (from Vaiśeṣika to Vedānta etc.) and the puzzled they involved (such as how could it be possible to bring into existence something that previously did not exist) are all due to thinking about the problem in linguistic terms. Their answers, in other words, were oriented by the Sanskrit form of basic sentences such as “the potter makes a pot”. In fact, how can the pot figure as the object of a sentence, given that it does not exist yet? Bronkhorst thought that this was a linguistic problem, namely one occasioned by the structure of language and not an ontological one. Westerners, according to Bronkhorst, would have immediately labeled the pot as non-existing until it is realised by the potter and would not have paused on its ontological status, whereas Indians never distinguished between linguistic and external reality.

    This is an interesting insight, and in fact there are several elements suggesting (as Karl Potter maintained) that the “linguistic turn” occurred in India much earlier than in Europe (note that I am saying the same thing Bronkhorst said, but looking at it from a more favourable perspective), such as the insistence on the analysis of linguistic data in order to solve epistemological or ontological issues (cf. the insistence on the linguistic use śabdaṃ kṛ- within the debate about the ontological status of śabda).
  • d) Bronkhorst was a convinced asserter of the scientific approach. This does not mean that he was an a-priori believer in natural sciences. Rather, he thought that the scientific method is based on a healthy form of scepticism and thus can never lead to fanatical beliefs nor to any form of “scientific traditionalism” (if correctly applied). For this very reason, he also thought that the scientific method was not “Western”, it had proven to work because of its ability to ask questions and thus to be universal. He took seriously Yoga and meditation techniques and thought that they could be analysed with the scientific method and possibly lead to new discoveries.
  • d2) Similarly, Bronkhorst clearly looked down on blind believers and thus praised Sanskrit philosophers for their ability to distinguish myths from arguments. In “What did Indian philosophers believe?” (2010) he noted that Sanskrit philosophers did not attack each other based on myths (although, one may add, some Buddhist philosophers did have fun at criticising some passages of the Veda and Kumārila made fun of the walls-speaking argument), but rather their arguments (“These philosophers, while criticising each others’ views, never attacked each others’ myths. Yet these myths would have been easy targets, if they had been seriously believed in”). In short, the reliance on the scientific method meant a radical openness to defeasibility of one’s beliefs and to a data-based approach.

Let me add a few words about Johannes Bronkhorst as a human being. The Indology list was full of “Bronkhorst stories” and therefore I will not need to take too much of your time with them (you can read them on the Indology archives). Let me just point out how Bronkhorst was generous and supportive with younger scholars and even students, but in a very unique way. I still remember our first meeting. I was an undergraduate student and he immediately asked me which were my key interests (I was unable to give a specific answer, at that point I was just busy learning Sanskrit and reading as much as possible of any text my professors read). I read or hear similar stories from others, all pointing to how Bronkhorst took people seriously, even young people. He was supportive, but not patronising. He was interested in one’s opinion, but would not refrain from saying that it was wrong if he thought so, according to the scientific method discussed above. He would not mince words to attack a view, but not so when coming to the person holding it, and I have seen him greeting warmly people with whom he had had violent disagreements on specific issues.

On sacrificial violence

One cannot solve the Agnīṣomīya problem (the clash between the prohibition to perform any violence and the prescription to slaughter an animal as an offer to Agni and Soma) via an appeal to suspension (bādha) of the prohibition to perform violence.

Using suspension would be based on the fact that the prescription to perform the Agnīṣomīya is more specific than the prohibition. However, if this were a viable move, then it would apply also to the case of the Śyena (a sacrifice to be performed in order to achieve the death of one’s enemy). But all Mīmāṃsā authors agree that the śyena should not be performed.

Thus the Agnīṣomīya case cannot be solved through suspension, as this would have been applicable also to the Śyena scenario and this would be an unwanted output.

How else can the Agnīṣomīya riddle be solved? By explaining that the prohibition against violence was only about violence-as-part-of-the-result and not about violence-as-part-of-the-instrument. Thus, sacrificial violence, which is only part of the sacrifice qua instrument, was never forbidden, whereas violence as part of a sacrifice’s result is forbidden.

Now, you may suggest that this reasoning leads to the unwanted consequence that violence which is instrumental to a different result would also not be prohibited. Does this mean that beating someone in order to take their wallet would not be prohibited, because it is instrumental? No, because here violence would be part of the result (you want the person to be made harmless/unable to react).

This is, by the way, consistent with my previous studies on deontic conflicts, according to which suspension can only be applied to prescriptions, and not to prohibitions.

In memoria, Raniero Gnoli (1930-2025)

I met Raniero Gnoli for the first time at the beginning of my university career, when I started attending his class on Abhinavagupta’s commentary on the Paratriṃśikā. I had only had one year of Sanskrit and I spent each class taking as many notes as possible, which I then re-read in the following days, trying to reconstruct the translation of a text that was way beyond my competence.

Raniero Gnoli had published his translation of Abhinavagupta’s commentary in 1985, but had decided to get back to it and while reading it with us (he did most of the reading) he partly revised his previous understanding. I was impressed by how learned he was, by how he seemed to understand Sanskrit without any difficulty (for instance, if he ever hesitated about a word, he would mentally check its usages in other Sanskrit texts he seemed to know almost by heart and thus understand it better), but also by his willingness to revise his previous understanding and to let us look at how he engaged with his past self and with Abhinavagupta himself.

The following year he read with us a Sanskrit text together with its Tibetan translation (he knew Tibetan, but refused to use the contemporary Tibetan pronunciation and would just pronounce “rgyu” and the like as if they were Sanskrit words, namely as र्ग्यु). During my first 1–2 years he was also the director of the Department of Oriental Studies (the name might have been different at that point, I am no longer sure) and I asked him his advice about what to specialise on. I told him that I was also thinking about Indian art and archaeology. His answer was clear-cut, namely that Chineses excelled in art, Indians not so much. There was thus no point in specialising in it. (I don’t know whether he said it as a test or a form of upāyakauśalya, but it did the trick.)

For many years to follow, Raniero Gnoli had stopped teaching at the University, but would still welcome us in his castle to read Sanskrit together. “We” means the students of Sanskrit at that point, a group of undergraduates who were still struggling with even simple Sanskrit. It took us years to get better, but he welcomed us at all stages. He read with us whatever we asked him to read, from the Nyāyabhāṣya (which was understandable for us, hence it made us overconfident) to Prajñākaragupta’s commentary on Dharmakīrti (too difficult for us!).

Although he was among the most productive scholars imaginable, always studying new things (often unheard for, such as how to execute the “punto Milano” embroidery), he appeared to always have time for reading Sanskrit with us. For the first months, he just read himself, then at a certain point he must have understood (how? Out of our pupils’ movements?) that we had become good enough and just asked us to translate (I was shocked, but it worked) and then corrected us. We could come for weeks in a row with the same text or just show up with a new text, he was open to everything.

He had strong opinions about Sanskrit philosophy and seemed to be guided by a crystal clear intellect. This must have been the reason he could recognise the philosophical genius of Abhinavagupta at a time virtually no one had even noticed him. For instance, he is the origin of the translation “aesthetic experience” for rasa, thus understanding what was at stake with Abhinavagupta’s insistence on it, at a time when other translators still spoke of “juice” or the like. Similarly, Gnoli would have definite opinions about how to put together a critical edition (which needs to be “critical” and not a mere collation) and about priorities for the discipline. Once, for instance, we asked him what prompted him to prepare the critical edition of the Pramāṇavārttika+svavṛtti. He said that it was needed.

After reading, he would chat with us for a while. We got to know about his journeys through India, often looking for stones (like a certain lumachella) or manuscripts, or about his Greek literature exams during his university studies, or about Tucci’s eating restrictions. He was the kind of superior being who does not need to show their superiority, since it is apparent, and he was therefore humble, often wearing old and worn trousers and even making fun of himself. For instance, when he turned 70 Raffaele Torella put together two volumes in his honour (“Le parole e i marmi”) and organised a small ceremony for him once the volumes were published. Gnoli started his speech saying that it was nice to feel younger (since the volumes ended up being published two years after his 70th birthday) and later commented with us that the whole ceremony had been a “preview” of his funeral —little did he know that he would have lived and worked for further 23 years.

Personally, my debt towards him is inextinguishable. I learnt most of what I know about Sanskrit philosophy from illustrious teachers, and Raffaele Torella and Raniero Gnoli were not only the first ones, but two absolute models of scholarship. I can’t fathom how much good karman I must have accumulated in the past to have been worthy of reading with them, learning from their pāṇḍitya, which was so deep it looked spontaneous, and even listening to their fearless way of expressing ideas —very different from what many describe as “typical Academia”.

Unrequested advice for students translating (Sanskrit) texts (will be updated)

This list will be updated and I will add here links to other posts where I discuss parentheses, brackets, footnotes and so on.

1) On parentheses and brackets: https://elisafreschi.com/announcements/how-to-use-parentheses-and-brackets-in-your-translation-my-tips/#more-3372

2) On transcribing Sanskrit passages: https://elisafreschi.com/2018/05/28/dividing-words-in-transcription/

3) I am very much in favour of being more explicit in highlighting objections and replies, e.g.:

[Objection:]…

PARAGRAPH BREAK

[Reply:]…

PARAGRAPH BREAK

Just saying “you may object” is not enough, because it does not identify the end of the objection and often leaves the beginning of the reply unmarked. Moreover, adding specific markers forces us to be clear in identifying objections and replies, whereas just translating Sanskrit markers like “nanu” might leave ourselves and the readers in the dark about who is talking (“nanu” usually introduces an objector, but not necessarily so!).

4) How to cite Sanskrit words: https://elisafreschi.com/2020/01/20/how-to-deal-with-sanskrit-words-in-an-english-article/

5) Refer to specific sūtras, not just to page numbers (e.g. “TV ad 1.1.2”, not “TV 1929, p. 101”). In case of longer passages, adding the page number might be necessary, in which case using the most authoritative edition is the thing to do. For Kumārila’s TV and ṬṬ it has become customary to refer to Subbāśāstrī’s 1929 edition. For NM, Mysore edition (whenever there is not a critical edition available, such as Harikai or Graheli respectively).

6) Scholastic Sanskrit is primarily about nominal sentences, not so English. Hence, break up nominal sentences into verbs+nouns ones. Don’t say “because of the absence of the possibility of X” but “because X is impossible” (for X-upapatti-abhāvāt). Repeating the nominal structure of the Sanskrit in an English sentence is not “being faithful to the Sanskrit”, since the Sanskrit sentence did not sound awkward to its readers/listeners, whereas the allegedly faithful English translation that preserved its nominal style is unbearably cumbersome.

7) I disagree with experts completely prohibiting the use of passive sentences in English. Nonetheless, the use of the passive is not marked in Sanskrit (meaning: it is considered as absolutely normal by readers/listeners), whereas it is often marked in English. These cases of undesirably marked passive need to be rendered into active in English. For instance: “Who is reading the book?” rather than “By whom is the book being read?”. This also has the advantage of forcing us to reflect on the agent (is it the siddhāntin saying it? The pūrvapakṣin?). By contrast, unmarked passive can be rendered as passive.

8) As a rule of thumb, Sanskrit philosophers were smart guys. If your translation attributes them an inconsistent thought, you must rethink it.

Btw, unrequested advice on bibliography: https://elisafreschi.com/2025/04/15/unrequested-advice-on-bibliography/