Naive referentialism and Indian philosophy. A Guest post by Johannes Bronkhorst

In a number of publications, I have had the audacity to propose an explanation for certain developments in the history Indian philosophy.1 A simple assumption made clear how and why Indian thinkers had adopted a number of at first sight counterintuitive views, including the satkāryavāda, ajātivāda, śūnyavāda, anekāntavāda, etc. Very simply put, the thinkers concerned found it problematic that in situations described by statements such as “the potter makes a jar” there is no jar. The above-mentioned views were believed to solve this problem.
It goes without saying that these solutions will convince few modern readers, because most of them do not find it problematic that there is no jar in the situation described by “the potter makes a jar”. And without problem there is no need for a solution. Most modern readers will therefore conclude that the problem of the ancient thinkers was ill conceived and perhaps based on a somewhat naive understanding of the relationship between language and reality.
Given all this, it is not a little surprising that some scholars attribute naiveté, not to the Indian thinkers referred to above, but to the researcher who drew attention to their way of thinking, i.e., to me. Consider the following:

… we wish to note that naive referentialism, which is implicitly adopted by several Sanskritists who attempt to analyse the ancient Indian philosophy of language (see for instance the way the problem is framed in Bronkhorst 1999), does not, or at least not always, represent a good explanatory framework. This is so not only because it is not tenable in general, but also because, in particular, it does not represent a fitting parallel to the ancient Indian thinking on this topic (Freschi & Keidan 2017: 260).

I have some difficulty with his passage. It appears to accuse me, rather than certain early Indian thinkers, of implicitly adopting naive referentialism. This, of course, would be a case of killing the messenger, and therefore bizarre.
Let us therefore assume that what the authors of this article really wanted to say is that I attribute naive referentialism to certain early Indian thinkers. This does not in their opinion “represent a good explanatory framework … because it is not tenable in general”. This observation seems to presuppose that all the views and opinions of Indian thinkers were “tenable”. It excludes the possibility that they could ever be wrong or mistaken. I fail to believe that the authors of the article seriously accept this position.
One point remains: One should presumably not attribute naive referentialism to early Indian philosophers “because … it does not represent a fitting parallel to the ancient Indian thinking on this topic”. This point might be taken seriously, were it not for the fact that it contains a generalization that is hard to sustain. What is this “ancient Indian thinking” on cognition, language and reality that was accepted by all during all the phases of the development of Indian philosophy? What we do come across in the literature is a variety of opinions that are not always in agreement with each other (or indeed with modern thought).
My observations so far might be considered irrelevant if the attribution of more sophisticated ideas to the early Indian thinkers could provide a better explanation of the appearance of satkāryavāda and related doctrines. I cannot a priori exclude that this may be possible, but have not seen evidence that anyone has made the effort. Until and unless someone will successfully make that effort, I find it hard to see what is wrong with my explanation.
To illustrate my position, let us look at one passage that “proves” satkāryavāda. Śaṅkara on Brahmasūtra 2.1.18 states:

prāg utpatteś ca kāryasyāsattve utpattir akartṛkā nirātmikā ca syāt /
If the effect does not exist before it arises, the arising would lack an agent and would be void (taken from Bronkhorst 2011: 66).

The context of this line shows that it concerns the statement “the potter makes a jar”. Śaṅkara clearly thinks that the jar is there before it arises. He links this conviction to the linguistic expression, for he then states that, if there were no jar in the situation, one would say “the potter, etc., which are causes, arise” (kulālādīni kāraṇāni utpadyante) instead of “the jar arises” (ghaṭa utpadyate).
I conclude from this example that Śaṅkara somehow believed that the presence of the word “jar” guarantees that there is a jar in the situation described by “the potter makes a jar”. This conclusion is not waterproof, of course, but I do not succeed in coming up with a more plausible and more appropriate interpretation of this passage. If this means that I attribute to Śaṅkara “naive referentialism”, so be it. I am more than willing to change my mind if a better interpretation of Śaṅkara’s remarks is proposed that attributes to him more sophisticated notions regarding the relationship between cognitions, words and things. So far I have not seen any.
The critical remarks discussed above occur in an article that makes the point that a specialist acquaintance with modern ideas can help the interpretation of ancient texts. Indeed, “[a]ncient scientific treatises can be very sophisticated and anticipate modern insights; they should therefore not be considered ‘primitive’ or ‘naive’ by default” (p. 251). This is, in my opinion, a valuable observation, and the article gives some examples of misunderstandings in modern scholarship that are due to the neglect or ignorance of this principle.
However, there is no guarantee that all that is said in ancient Indian scientific treatises anticipates modern insights. An overzealous application of the principle may induce scholars to impose interpretations on ancient texts and overlook direct interpretations that do not anticipate any modern insights. The study of Indian philosophy should not become a game in which overly clever scholars “find” one modern insight after another in texts whose authors, to the best of our knowledge, were unaware of those insights. The presence of modern ideas in ancient texts remains something that is in need of demonstration, not something that can be taken for granted.


Bronkhorst, Johannes (1999): Langage et réalité: sur un épisode de la pensée indienne. Turnhout: Brepols. (Bibliothèque de l’École des Hautes Études, Sciences Religieuses, 105.)
Bronkhorst, Johannes (2011): Language and Reality: On an episode in Indian thought. Translated from the French by Michael S. Allen and Rajam Raghunathan. Revised and with a new appendix. Leiden – Boston: Brill. (Brill’s Indological Library, 36.)
Bronkhorst, Johannes (2013): “The correspondence principle and its critics.” Journal of Indian Philosophy 41(5), 491-499.
Freschi, Elisa & Keidan, Artemij (2017): “Understanding a philosophical text: The problem of ‘meaning’ in Jayanta’s Nyāyamñjarī, book 5.” Reading Bhaṭṭa Jayanta on Buddhist Nominalism. Ed. Patrick McAllister. Wien: Verlag der Österreichischen Akademie der Wissenschaften. Pp. 251-290.


  1. Primarily Langage et réalité: sur un épisode de la pensée indienne (1999) and its English translation Language and Reality: On an episode in Indian thought (2011); further “The correspondence principle and its critics” (2013).

Author: Johannes Bronkhorst

About elisa freschi

My long-term program is to make "Indian Philosophy" part of "Philosophy". You can follow me also on my personal blog:, on Academia, on Amazon, etc.

5 Replies to “Naive referentialism and Indian philosophy. A Guest post by Johannes Bronkhorst”

  1. Hi Johannes!

    I don’t find your translation of Śaṅkara matching the two instances of the particle ‘ca’ in the sanskrit, only the second one. Śaṅkara seems to me to say something like, “if the effect does not exist and / with that which (as-ya!) arises. . .”

    This problem was certainly familiar in ancient philosophy: Aristotle introduced potential existence to deal with it, placing him closer to Śaṅkara than your Whiggish, modernizing reading, which is just naive Hume, in the fallacious form “post hoc ergo propter hoc”: after the fact, thus because of it.

    The so-called Indian syllogism has been relentlessly misread in this way, to no good effect. And Śaṅkara had the whole real debate behind him, while you have him faltering on entry to the arena of debate, which is gross anachronism.

    I am not here taking modern science as the criterion of truth, merely observing that sanskrit compounds allow expressions of process which are had to translate, but echoed in Whitehead’s process philosophy, as in William James on the stream of consciousness.

  2. I will briefly mention something which has, probably, been taken into account. At some point, in some fashion anyway. Eastern thought, of any sort, is old. The ancient origins do not, of themselves, ensure clarity, or necessarily, wisdom. This feature is likewise evident in expressions of other belief systems, eastern; western or whatother. Things lost in translation(s) may never have been there to lose. No one (or, at least, few) who support(s) time-honored tradition(s) wants to concede that such thinking leaves something out, or unexplained.
    Seems to me, anyway.

  3. The new (Greek neotas) thought widely stereotyped as modern and Western is now dated. Breaking news in current science, has a whole series of modernist conjectures refuted. Latest (recent at Quanta Magazine) is the grand dogma of reductive computationalism (in symbols, P=NP): no, computations do not all beak up into simple independent components (P) there exists a whole class of genuinely hard interactive problems (NP).

    And with that finding goes also the ancient dogma of compositionality. So it’s no longer scientific to assume with Wikipedia & vulgar or Carvaka materialism that reality is a simple thermal ensamble of discrete particles. That doctrine gave us the dark Satanic mills of Victorian industry, now at last headed for obsolescence, along with their high carbon and methane output.

    In this ironic context, I find reason to appreciate the ascetics who long ago dissented from the whole development, and sought a more wholesome way of being. Their vegan and vegetarian foods are now entering the mainstream, welcomed by the health sciences. And that already was the force behind the critical debates in Classical India: it was all set against the monumental enterprise of editing the medical classic of Charaka:.

    Causality in medical problems is notoriously complex, and that was equally clear to the West. “It is absurd to propose that purpose is not present because we do not observe the agent deliberating. Art does not deliberate. If the ship-building art were in the wood, it would produce the same results by nature.” (from the Physics attributed to Aristotle.)

  4. Not being acquainted with present-day discussions on naive referentialism, I will not comment on it. But in regard to the problem that Johannes identifies, I am attracted to the suggestion that “the thinkers concerned found it problematic that in situations described by statements such as ‘the potter makes a jar’ there is no jar.”

    So why did they find it problematic, whereas we do not? My immediate thought is that it is because they accepted the kāraka theory. According to that theory, the jar designated as the object of action should be a contributing causal factor (kāraka or sādhana) in the action of jar-making, and therefore should exist prior to the action in some form or other. To quote from the Sādhanasamuddeśa:

    निर्वर्त्यो वा विकार्यो वा प्राप्यो वा साधनाश्रयः।
    क्रियाणांएव साध्यत्वात् सिद्धरूपोऽभिधीयते॥79॥

    Iyer’s translation: “Whether the object be something to be made or something to be modified or something to be reached, it is the substratum of power and is presented as an accomplished thing, because it is only actions which are to be accomplished.”

    This raises the further problem of the precise sense in which a jar, which is not yet made, can be a contributing factor in the accomplishment of the action of making. Two solutions are presented in the Vākyapadīya. If the meaning of a word like “ghaṭa” is the jāti, then the jar as a causal factor contributing to the jar-making exists in the form of ghaṭatva-jāti, and what is made is a particular (vyakti) jar (Jātisamuddeśa 27). However, if the meaning of a word is the vyakti, then the particular jar as a causal factor contributing to the jar-making exists in a mental form (buddhirūpa), which is then produced through the jar-making (Sādhanasamuddeśa 7).

    As for Śaṅkara’s statement, if that was intended to be an argument for satkāryavāda, I think it shows that he found kāraka theory to be a supporting ground for satkāryavāda.

  5. Thank you, Paul, Johannes, Boram and all!

    The quote above is taken from an article where Artemij Keidan and myself were trying to reconstruct what Jayanta meant when he spoke of artha, viṣaya, abhidheya, vācya… within NM 5 (part of the discussion of śabda). The entire article is available here:āyamañjarī_Book_5

    A short methodological comment: In general, I think that it is far to assume that authors using different terms within the domain which is the domain they are focusing on (say, different terms about śabda while discussing language) mean something (perhaps slightly) different with each one of them. It seems to me that assuming that they just mean the same is not a good methodological starting point. One might end up deciding that a certain author is speaking of, e.g., varṇa, śabda, pada and vākya synonymously, but I would not start with this hypothesis. My starting hypothesis is, rather, that each word is used with regard to something else and that *I* am the problem if I am not seeing the difference. The advantage of this approach is that even if I have to end up saying that there is no real difference (say, between the usage of śabda and pada), I have certainly improved my understanding of the passage in the process. Whereas starting with the default assumption that the authors at stake just used words imprecisely or whimsically will not improve one’s understanding and entails the risk of one’s overlooking slight differences in use. Once again, this methodology only makes sense if it is applied to a philosophical text and it is about the topic the author is focusing on. A kāvya text will use many synonyms, e.g., for metrical reasons. A philosophical text dealing with ontology might use different synonyms of pada just for the sake of variation. But I would not expect a linguist to use different terms for the same reason.

    Thus, this is what Keidan and myself were doing while looking at Jayanta and, imho, we reconstructed Jayanta’s understanding of each term in a way which accounts for each use. If this reconstruction is correct, then Jayanta was not just a naïve referentialist and was able to distinguish between external objects and meanings.
    Why would he have achieved such level? Perhaps because of his exposure to Mīmāṃsā, which forced him to consider arthas such as kārya and apūrva, for which there is no external world referent matching them at the moment in which they are said. The liṅ in yajeta means something that ought to be, but that does not exist yet. Moreover, it is not even something future (at least if you follow Prabhākara, whose views Jayanta knew and discussed).

Leave a Reply

Your email address will not be published. Required fields are marked *