Could Upaniṣadic sentences be interpreted as prescriptions? A debate within Maṇḍana’s Vidhiviveka

Within the Vidhiviveka, a Prābhākara-inclined Mīmāṃsaka debates with a Vedāntin about the meaning of Upaniṣadic sentences on the self.

The Prābhākara insists that all sentences should be injunctive in character, and that Upaniṣadic sentences should also be interpreted in this way. But what exactly could they prescribe? They could enjoin one to know their content, they suggest. But this option is not viable, since there is no difference between “s is p” and “you ought to know that s is p” insofar as one knows that s is p through the first sentence, too.

A further option would be that the Upaniṣadic sentences enjoin one to know their meaning in a definite way (niścaya), but this option is also ruled out. For, if the niścaya were obtained through language alone, then it would occur automatically, without the need to enjoin it. If it were not obtained through language alone, then it would occur because of something other than language, but then the Upaniṣads would no longer be the instrument for knowing about the ātman, which runs against other fundaments of the school.

At this point, the Prābhākara-like objector suggests a further possibility, namely that Upaniṣadic sentences prescribe arthaparatā, i.e., the ‘intentness on the meaning’. This also does not go, because prescriptions need to have preferably a visible purpose, like ritual prescriptions. And understanding sentences as aiming primarily at their meaning would put at a disadvantage exactly prescriptive sentences, which aim not only at conveying their meaning, but also at urging someone to perform a given activity. Moreover, a prescription should aim at a certain goal and knowledge cannot be a goal separated from the content to be known (as explained above with regard to the case of whether the understanding itself could be enjoined): ज्ञानस्य ज्ञेयाभिव्याप्तिफलत्वात् फलान्तरानभ्युपगमात्।

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.

8 Replies to “Could Upaniṣadic sentences be interpreted as prescriptions? A debate within Maṇḍana’s Vidhiviveka”

  1. I am enjoying these weekly puzzles on Mimamsa positions, Elisa. These positions seem to defy common sense, but the survival of the Mimamsa tradition through debates with various schools and across centuries suggests that it has developed time-tested defenses for them that are not easy to defeat. So, in this case, too, I am interested in knowing whether (and if so how) an actual Prabhakara Mimamsa proponent wrote a later work responding to the Vidhiviveka.

    In my state of ignorance, I can only speak to how I think a modern-day Mimamsaka might respond, apprised of developments in contemporary Western philosophy.

    (1) An imperative statement has world-to-word direction of fit: if the world fails to fit the statement, it ought to be changed to fit the statement. More precisely, stating the propositional content ‘p’ in the imperative mood has the effect of giving ‘p’ a world-to-word direction of fit, enjoining that ~p ought to be changed to p if there is a mismatch such that ‘p’ and ~p.

    (2) An indicative statement has word-to-world direction of fit: if the statement does not fit the world, it ought to be changed to fit the world. More precisely, stating the propositional content ‘p’ in the indicative mood has the effect of giving ‘p’ a word-to-world direction of fit, enjoining that ‘p’ ought to be changed to ‘~p’ if there is a mismatch such that ‘p’ and ~p.

    On this basis, a Mimamsaka can propose two further ad hominem arguments based on specific premises that Mandana himself advances, or that Advaitins generally ought to find acceptable.

    First, one step in Mandana’s second argument is interesting: “if the niscaya were obtained through language alone, then it would occur automatically, without the need to enjoin it.” Now, if something occurs automatically, that precludes the failure of its occurrence. So this suggests to me the following maxim: “Ought implies the possibility of failure”. This maxim, implicitly accepted by Mandana, can show how an indicative statement implies an ought. An imperative statement implies an ought, and that ought is implied in part because the *world* can fail to fit the statement. However, an indicative statement also implies an ought, and that ought is implied in part because the *statement* can fail to fit the world. This is true even of Upanishadic statements, even on the Advaita view, as I shall propose below.

    Second, according to Advaitins, no word nor statement is adequate to describe ultimate reality. Nirguna Brahman being ineffable, any descriptive statement about it is bound to fail, introducing duality where there is none. Ultimately, then, all statements in the Upanishads are false, and the best way to describe Brahman is “neti, neti”, or it is better yet to be silent. The above direction-of-fit explanation can explain why this should be so of indicative statements in the Upanishads: if there is any mismatch between indicative statements and reality, the statements must be revised to fit the reality. But no positive statement can ever get it right. Hence the negations, hence the silence.

    (p.s. Later I will write up a response to your suggestion in the previous post.)

    • Hi Boram and thanks for your comments. I think you are generally right, I would just add the role of the adhikāra. Basically, the adhikāra makes sure that the imperative is not addressed indefinitely, but to a precise person (the adhikārin), who therefore *might* fail but is not a condition in which they *must* fail (i.e., ought implies can). If they just cannot realise the command, they are not the adhikārin of it.

      • Elisa, thanks, that is a very important point! It might be best for me to answer in three stages, gradually building up to a more adequate response.

        (1) Imperative sentences state that such-and-such ought to be done, and as you say, these statements are made to eligible agents who are able to perform the stated actions, and hence are responsible for performing them. By analogy, then, indicative sentences state that such-and-such is the case, and these statements are made or accepted by eligible cognizers who are able to verify the stated scenarios, and hence are responsible for making or accepting true statements. (And by “ability to verify” indicative statements, I mean to refer to the means of knowledge or the pramana-s.)

        (2) To make sense of the Prābhākara claim that indicative statements are also injunctive, I construed “injunction” in a broad sense to include directions of fit. A word-to-world direction of fit enjoins that the propositional content of indicative statements conform to the world. A world-to-word direction of fit enjoins that the world conform to the propositional content of imperative statements. Although these two directions of fit are different, they are both normative rules, and prescribe goals that are not only regulative but also constitutive of indicative and imperative statements. Because directions of fit are constitutive rules for indicative and imperative statements, they are ontologically and conceptually prior to indicative and imperative statements.

        By the same token, the adhikaarinah for these constitutive rules are prior to the speakers and hearers of indicative and imperative sentences. They are subpersonal or subagential mechanisms that help constitute a cognizer or agent. The constitutive rules for indicative statements and cognitive mental episodes apply to subpersonal belief-forming mechanisms, the pramana-s. These mechanisms can fail to produce true cognitions due to various defects, but they are also capable forming true cognitions. As for constitutive rules governing imperative sentences, they apply to the desire-forming and action-guiding mechanisms of agents.

        To account for the normativity of these rules, I would appeal to Ruth Millikan’s teleosemantics and naturalistic, evolutionary account of proper functions for biological and cognitive mechanisms. I doubt that Mimamsakas would be attracted to Millikan’s account, because they believe in the beginningless nature of Vedic sentences, and that would seem to rule out an evolutionary, historical account.

        (3) Aside from the indicative and imperative mood, there are also benedictive, optative and subjunctive moods that can be used to express a blessing or a hope or a wish. Perhaps, for benedictive sentences, an adhikārin is required, namely gods or goddesses that have the power to give the blessings. But when optative and subjunctive moods are used to express a hope or a wish, such as “Would that the world were free of natural disasters!”, that expresses a desirable state of affairs that cannot obtain, then no adhikārin could be addressed by it.

        So I wonder if Mimamsakas also assimilate benedictive, optative and subjunctive statements under injunctions. In my view these statements all have world-to-word direction of fit, and although they express unrealizable hopes and wishes that could not have directly contributed to the survival of organisms that had them, they are byproducts or spandrels of desire-forming and action-guiding mechanisms that do have a survival value.

        • Thanks again for these thoughts, Boram.
          As for Mīmāṃsā in general, content prevails over form, in the sense that vidhis can be expressed with any mode (indicative included, in cases such as “The ladle is made of parṇa wood”, meaning “Make a ladle of parṇa wood”). For the Prābhākaras, vidhis are just the main form of language and indicative sentences have only value as auxiliaries of them. Your idea of “directions of fit” seems appropriate to depict Maṇḍana’s theory (in the project, we used stit (‘see-to-it-that’) logic so far).
          Thanks for pointing out Ruth Millikan’s work. Would you start with her Language, Thought, and Other Biological Categories?

          • I think Millikan would refer to cases like “The ladle is made of parna wood” as pushmi-pullyu representations. (One of her own examples is, “This meeting is adjourned” said at the end of a meeting.”) These representations have both descriptive and directive content. In her view, such representations predate the bifurcation of content into descriptive and directive, or indicative and injunctive. For example, the honeybee’s figure-8 dance is descriptive and directive at the same time, indicating the location of the nectar and directing other bees to go there.

            For me personally, LTOBC is up there with the Critique of Pure Reason as the most difficult books I’ve read in grad school. You can read the first two chapters on proper functions, and then decide whether to read on. It would be easier to come back to the book after reading her later ones like the Varieties of Meaning or Language: A Biological Model (this last book is a collection of previously published articles, and so is White Queen Psychology and Other Essays, so you can just look at the chapters and pick the chapters/articles you want to read, and with Varieties of Meaning too you can dip in wherever you want. It will be more difficult to do that with LTOBC).

  2. So there is limited use in importing the syntax / semantics distinction: this Mimamsa is relational, and C.S. Peirce’s way with that (he was of Brahmin extraction) was already misunderstood. Or more specifically, connectionist, working in the relationship of speaker and auditor: and something of that *does carry to Aristotle’s Posterior Analytics, also much misunderstood. I’m here investigating those slender links.
    It’s remarkable how Samkhya lore tracks the Mimamsa development; and there, as it influenced Ayurveda, you find some sixty Yuktis, points of logical clarification, prefacing the Charaka-samhita. And reflecting the kind of logico-linguistic wrangling still ongoing here! That also tracks West, with the art of disputation, to Eristic in the school of Megara, where Plato passed through after the death of Socrates. We may owe the method of Analysis to that pass, as a formal topic.

  3. Thanks for the reference to Millikan’s work and to the “meeting is ajourned” example. A Prābhākara would say (you would guess it) that it is purely prescriptive. A speech-act solution sounds also interesting.

Leave a Reply

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