7. Mīmāṃsā Classification of Vedic Texts and the Upaniṣads

We saw, thus, how the Mīmāṃsakas understood ritual as a means of human felicity. By and large, they considered the Vedas to be only about this: the purpose of the Veda was to enjoin ritual and communicate in some way that its performance is a means of human happiness. For that purpose, a text had to be worked out which would cover all the details of the performance as well as the knowledge and know-how necessary for it. This text would need to contain knowledge about the agent, details about offering preparation, the use of implements and the role of the recipients, a script for the performance, an incentive for the sacrificer, all teleologically driven by the injunction which introduces the ritual. The text should be constituted from the Veda, and Mīmāṃsakas classified the entire Veda by types of passages that it contains in terms of their ritual applicability. In this section, we will outline this classification, and then we will try to pinpoint the role of the Upaniṣads in it.

Mīmāṃsakas classified the Veda in four kinds of texts: (1) vidhi or injunctions; (2) mantra or sacrificial chants; (3) arthavāda or descriptive passages; and (4) nāmadheya or names. It is no surprise that the injunctions were considered the central texts: only they were directly related to dharma. The injunction group referred to the portions of the Brāhmaṇas which enjoin the performance of sacrifices and included all kinds of injunctive texts, which Mīmāṃsakas meticulously classified. Vidhi covered not only the principal injunctions, but everything that is enjoined in the sacrifice, for instance the sannipātyopakārakas, the ārād-upakārakas and all the individual actions they involved and auxiliaries they required. Crucially, they involve everything that is to be understood “just as it is heard.”

Injunctions were classified primarily based on two criteria: (1) what they enjoin; and, (2) how they enjoin it. Under the first rubric, one common division is into utpatti-vidhi, an originative injunction in which the predicate is the principal element of the ritual performance; and viniyoga-vidhi, an applicative injunction which affirms that some auxiliary is related to the principal. The two are commonly referred to as a pair by Kumārila in the Kumārila’s Tantra-Vārttika on Śabara’s Mīmāṃsā-Sūtra-Bhāṣya, and we can think of them as puruṣārtha-vidhi and kratvartha-vidhi, enjoining that which is for the good of man or that which is for the good of the ritual, respectively. Good instances would be agnihotraṃ juhoti, “He performs the daily fire ritual,” and dadhnā juhoti, “He uses yoghurt as the oblation.” To this pair, a third is commonly added, adhikāra-vidhi, a statement that introduces the ritual agent who is entitled to reap the fruits of the sacrifice, for instance the famous svarga-kāmo yajeta, “He who wants heaven should sacrifice.”

It is not necessary that these be stated in separate sentences. However, an injunction can affirm only one thing, and in complex injunctions the other elements are considered qualifiers of that one thing, producing thus a viśiṣṭa-vidhi, a qualified injunction. When these three injunctions in a ritual are put together through the three features of syntactic expectancy, ākāṅkṣā, in a hierarchy ascertained through the Mīmāṃsā principles of interpretation, they produce a prayoga-vidhi, a whole manual that delineates the integral organization of the ritual and the manner of its performance, as well as all the other details required. This is the text that I referred to above as the final product that covers the whole ritual procedure.^[1] This is commonly listed as the fourth type of injunction in this classification, and we should take a good note of it because it will play an important role in the mahā-vākya idea. This fourfold classification is given in later manuals, but the individual types of injunctions are common currencies in Śabara’s and Kumārila’s works.^[2]

The second classification of injunction asks the question, how is something enjoined. By this criterion, injunctions are commonly classified in three types: (1) apūrva-vidhi; (2) niyama-vidhi; and (3) parisaṅkhyā-vidhi. The locus classicus on these is Kumārila’s Tantra-Vārttika on Śabara’s Mīmāṃsā-Sūtra-Bhāṣya 1.2.34, although Kumārila there does not mention apūrva-vidhi, but rather talks about vidhi, niyama and parisaṅkhyā.^[3]

An apūrva-vidhi is a statement that enjoins by disclosing an otherwise unknown causal relationship. Take, for instance, the statement vrīhīn prokṣati, “he besprinkles the rice,” that is an action of the type of saṃskṛti or consecration, and adds an excellence of some kind to the substance over which it operates. That there is some causal relationship between the action of besprinkling and the excellence that obtains subsequently in the rice is not empirically knowable, and is solely due to the injunction. We can look at this from the point of view of the desired result. We need an element of excellence in the rice so that it can be used in preparing the sacrificial cake. Because this excellence is invisible, no action is empirically related to it. In Kumārila’s words, such relationship does not obtain “absolutely.” For all we know, it may be the action of arranging the rice into the image of LeBron James that will furnish the required excellence. Nothing of the kind is known to us “before the sentence,” and for this reason this type of injunction is “pure,” fully in the domain of the Veda.^[4] It is exclusively related to unseen results. It is eminently clear that this injunction is related to the original meaning of the notion of apūrva, unprecedented as knowable only from the Veda.

There may be cases, however, when two ritual elements are commonly related, in multiple possible ways. Keeping with the rice example, once the rice has been consecrated, its husk needs to be removed so that it can be used in making the cake. We know how to do that, and we could imagine more than one appropriate ways—this is not empirically unavailable. An injunction in relation to this reads, vrīhīn avahanti, “he threshes the rice.” The predicate of this injunction is not the action of rice preparation, but its specific mode of threshing qualified by the natural consequence of excluding all other possibilities, never mind if they are all accounted for or not. The important thing is that optionality obtains in general. This type of injunction is called niyama, restriction.

When, however, there is a similar situation but one in which the whole scope of what can be affirmed is known, and the point is not to affirm the stated element as intended but to exclude whatever is not stated, this injunction is called parisaṅkhyā, exclusion. Take, for instance, the statement aśvābhidhānīm ādatte, “he takes the horse’s bridle.” While we need not go into involved details, the statement as it stands is problematic because the action of holding is supposed to be performed alongside the recital of a certain mantra, but the statement seems to reiterate something already affirmed in a related text, which puts its purpose in jeopardy: a pramāṇa cannot repeat something known. The solution is to take the statement as intending not to enjoin the holding of the horse’s bridle while reciting the mantra, but as intending to exclude the holding of the donkey’s bridle which presents itself as an assumed alternative. The restriction still denotes the action of holding: holding as qualified not by what is said, but by what is not said when it could have been said.

This, obviously, leaves a lot of leeway for dissent, and it is often a matter of disagreement whether a specific statement is a restriction or an exclusion. We need not worry about this, but we should note well how the three are defined, because they will play a formative role in Śaṅkara’s making sense of meditation vs. reflection in Vedānta. To summarize, the first classification provides for the structure of the ritual: it is set in motion by a sentence that presents the central ritual element; ritual details are related to the central element; the agent entitled to the results is pointed out; the structure itself is given. The second classification, on the other hand, is concerned with knowing causal relationships that govern the elements of the ritual.

The second group of Vedic texts, mantras, referred to versified composition, generally in the Vedic Saṃhitās, which are recited in a sacrifice accompanying parts of the ritual and are intended as markers of these parts, the respective deities, the offertories, etc. “[M]antras allude to what is going on in the sacrifice as the priest executes it. Thus, recited in the proper sequence, they help the priest see what he is doing and remind him of what has yet to be done. They provide a running narrative of the rite.”^[5] Taber points to a crucial feature of theirs: “[T]heir meaning is usually evident as soon as they are pronounced. They are grammatical; they make sense of themselves. But, still, when a mantra is presented in the Veda as a formula to be uttered in the context of the ritual, one may take it to express what it means, or one may not.”^[6] Unlike the injunctions that must be taken as heard, mantras are more like a soundtrack. They do not say anything about Indra, Agni, the sacrificial fire, etc., but indicate what is happening at a given moment in the ritual and help the priest recall that detail, much as Wagner’s “Wedding March” played at a wedding to accompany the entrance of the bride is not about Elsa, but marks the entrance event. The reference of the waltz is the event, not Elsa.

As for the arthavāda or stories and descriptions found alongside the ritualistic sections of the Brāhmaṇas, they were a problem for Mīmāṃsā because they evidently do not enjoin an action—all that the Vedas are valid for—but are part of the Vedas and cannot be discarded without compromising the validity of the corpus. An additional problem with these passages was that many of them were just contrary to sensory evidence. Think, for instance, of the many bandhus or correlations in the Brāhmaṇas and the Upaniṣads, such as the famous identification of the sacrificial horse with the universe at the opening of the Bṛhad-āraṇyaka Upaniṣad. Given the premise that the Vedas are valid only regarding (ritual) action and do not teach about things that we can see, what to do with such descriptions many of which are plain false and all of which are not about action?

Śabara’s solution was to treat such passages as not being truth claims at all. Consider any story. A story can do two things: (1) it can give an account of past events; or (2) it can cause attraction or repulsion to something else, like an advertisement that makes you want the product no matter how accurate it is. This is what arthavādas do: they advertise the ritual action. Their validity does not consist in whether what they say is true or not—though one can always interpret them to avoid contradiction with the evident—but in aiding the performance of the sacrifice by making it look good.^[7] To take the standard arthavāda example: “One who wants prosperity should immolate a white animal to Vāyu. Vāyu is the swiftest deity. Vāyu comes with his own property and leads him [the sacrificer] to prosperity.” In Kumārila’s words, knowing that the cause and the effects are alike, one is made to believe that the sacrifice to the swiftest deity will make the result arrive without much delay.^[8] This makes the arthavādas purposeful and, therefore, valid, not in terms of truth, but action.

Kumārila brought his hallmark sophistication to the issue of arthavādas.^[9] I mentioned that he distinguished two kinds of bhāvanā or verbal productivity, arthī or actual and śabdī or verbal bhāvanā. In the paradigmatic injunction svarga-kāmo yajeta, this bhāvanā had the form of “bring about” for the arthī bhāvanā and “should” for the śabdī bhāvanā. The arthī bhāvanā further had three points which it required for its completion: an object (heaven), an instrument (the sacrifice) and a procedure of sacrificial performance. Now, the śabdī bhāvanā similarly needs to become complete in the same three points in order to accomplish its objective, which is to get the man perform the sacrifice. The śabdī bhāvanā is all about the taking up of the sacrifice, not its accomplishment, so naturally its object, answerable to the kim-feature of the ākāṅkṣā or the verbal need, is the taking up of the action, for which reason the sacrificer must be induced.

Further, the śabdī bhāvanā requires an instrument for effectuating this, corresponding to the kena-feature of the ākāṅkṣā, and for this it must give rise to an understanding of the injunction on the part of the sacrificer, one that is contingent on experiencing that there is a causal relation between the action to which he is prompted and the result that he expects. Kumārila calls these two—the kim and kena features—puruṣa-pravṛtti and vidhi-jñāna respectively, engaging man in the sacrifice through understanding the injunction.

Finally, the śabdī bhāvanā needs to find a way to do that, corresponding to the katham-feature of the ākāṅkṣā. The question, then, is how the optative suffix which expresses bhāvanā—and Kumārila is quick to point that the insentient suffix operates through the sacrificer’s awareness—can convince the sacrificer that the ritual action can furnish the result. “For, a man acts led by reason, and as long as he does not understand something as good, he will not act upon it.”^[10] We shift perspective now, from the optative suffix to the sacrificer, because the sacrificer must see the desirability of the sacrifice. This, Kumārila claims, can happen in two ways. The sacrificer can see, first, the excellence of the ritual action; it is an action laid down by a Vedic injunction, and the Veda is faultless. In this case, the śabdī bhāvanā operates solely through the optative suffix. Or, he can realize how the sacrifice is good because some of the deities or substances that are its part are excellent in some way. This is accomplished by the arthavāda sections: hearing, as I already said, how Vāyu is the swiftest deity, one’s understanding that the cause and the effects are alike is activated and a conviction that the sacrifice to the swiftest deity will make the result arrive quickly is born. The fact that an arthavāda is juxtaposed to an injunction makes these two seek each other for completion, and while the suffix could perform the same function alone, the presence of the arthavāda suspends that. Kumārila calls this feature of the śabdī bhāvanā corresponding to the itikartavyatā, the katham feature, the knowledge of excellence or prāśastya-jñāna.^[11]

The question now presents itself: where is the place of the Upaniṣads and the knowledge of the Self as its domain in this classification? The seemingly easy answer is: the Upaniṣads are part of the Brāhmaṇas and they do not enjoin action (or so it seems); ergo, they must be arthavādas.^[12] This is how scholars tend to present the Mīmāṃsā understanding of the Upaniṣads. For instance, Halbfass writes: “Kumārila mentions the Upaniṣads side by side with arthavādas, and he tends to see the Upaniṣadic teaching about the Self as being auxiliary to dharma, that is, to the performance of ritual actions, insofar as the notion of a noncorporeal permanent self is a condition and an incentive for performing such acts which are supposed to bear fruit in another life or birth.”^[13] Hirst explicitly identifies the Upaniṣadic statements about the as self as arthavāda in Mīmāṃsā: “These [non-injunctive statements] last were classified as arthavāda, secondary statements whose real function was to encourage a person to undertake ritual action. So, for example, all statements about the self were seen, not primarily as descriptions of the self, but as motivators to action, the self being the one who would accrue the result of the sacrifice performed.”^[14] “The application of the category of arthavāda (secondary passages) (v) is slightly more complicated. The Ritualists developed this notion to account for apparently descriptive passages, particularly those found in the Upaniṣads.”^[15] Rambachan gives a similar explanation:

The Pūrva Mīmāṃsā contention that the Upaniṣads have no independent purpose but are merely an appendage to the main body of injunctive text was a formidable challenge to Śaṅkara. … Many Vedic texts, for example, including the sentences of the Upaniṣads (vedānta-vākyas) are seen as having their purpose only in praising what has been enjoined in the injunctions (PSM 1.2.7). … According to Pūrva Mīmāṃsā, the Upaniṣads are merely an appendage to the main body of injunctive statements. The utility of the Upaniṣads lies only in praising the prescribed action or in providing some useful information, such as knowledge of the deity or agent for performance of a particular rite.^[16]

The fact of the matter is more complex than this simple identification of “being subordinate to dharma” with “being arthavāda,” and its corollary “the Upaniṣads as a unit are subordinate, ergo they are arthavāda.” Let us examine carefully what Kumārila says about the Upaniṣads. To begin with, Kumārila’s understanding of the complex ritual causality found a place for the Upaniṣads as providing knowledge about the agent in the sacrifice, ultimately serving the purpose of action but having truth value. We should recall here the puruṣārtha/kratvartha and principal/auxiliary organization of the sacrifice. The ritual agent in the sacrifice was an auxiliary factor, a kratvartha, and the Upaniṣadic texts which present knowledge of this agent were ultimately absorbed in the principal ritual action through the agent. Unlike the arthavādas, which for Kumārila were strictly in the realm of śabdī bhāvanā where truth values do not matter, the Upaniṣadic passages about the Self were absorbed in the arthī bhāvanā where accurate knowledge was important, although ultimately made use of in action. Ascertaining the details of procedure that involved the ritual agent was not related to like or dislike, and the success of the sacrifice was predicated on knowing such details. It is significant that Kumārila placed the Upaniṣads right there. “The Upaniṣads discharge their need (ākāṅkṣā) through presenting the agent that is subordinate to the ritual action.”^[17]

Thus, Kumārila included the Upaniṣadic description of the Self in the arthī bhāvanā, before he had the occasion to introduce the arthavādas and the śabdī bhāvanā as their domain. The issue of arthavāda appears with passages which are fanciful or do not contribute anything obvious to the action. They are not required as part of the sacrificial procedure, but are present in the text and must be accommodated because of that. Vāyu may be the fastest deity for all we know, but the point is that this does not matter in the sacrificial procedure. The situation with the Self is different, and in one sense can be compared to the sacrificial cake: the cake is subordinate to the action of offering, yet the passages which enjoin how to prepare it are not arthavāda, because they are predicated on having truth value. They enter the itikartavyatā. With the Self, of course, the issue was somewhat more crucial, since without a permanent Self that enjoys the results of the sacrifice the authority of the Veda would be compromised. Kumārila thought that such knowledge of the Self as an eternal agent and enjoyer of ritual action follows even from the bare injunctive statements through scriptural postulation—there must be an eternal Self that will enjoy the results, or otherwise what the injunction says would be false—but such knowledge becomes firm through the study of the Upaniṣads.^[18]

However, in other places Kumārila does treat Upaniṣadic texts as arthavāda. For instance, in his account of the origin of smṛti, he attributed the various theories of creation and dissolution common among Vedic folks to ideas that originate in the mantras and arthavādas.^[19] This certainly includes sections of the Upaniṣads. He traced even the origin of some Buddhist ideas—idealism, momentariness, the doctrine of no-Self—to the Upaniṣads and arthavāda, explicitly paired, and meaning to prevent excessive attachment to sensual matters.^[20] Reasoning also had origin in the Upaniṣads and arthavāda in pair.^[21] These ideas of his were not particularly revolutionary either: for most of them he had a precedent in Bhartṛhari.^[22] We may, further, venture to guess that he would have classified the Self-Brahman identification in the Upaniṣads as arthavāda, since such a doctrine, as noted by Nakamura, makes the Mīmāṃsā doctrine, predicated on a plurality of Selves, fundamentally impossible. “The eternal existence of the individual ātman, from the standpoint of the highest truth, is absolutely necessary and indispensable as the presupposition on which the Mīmāṃsā philosophy can establish their rites.”^[23] Finally, Kumārila refuted Vedāntic theories of the origin of the world and rejected the very possibility of a creator, which is so prominent in Vedānta.^[24]

Therefore, it seems to me that it is a notional mistake to talk about a general Mīmāṃsā attitude to the Upaniṣads as a single corpus, as it is commonly done. Kumārila clearly had an idea that they are distinct, focused on knowledge of the Self, but they contained the same types of sentences as the Brāhmaṇas: injunctions and their auxiliaries, and arthavādas. The fact that both are treated as auxiliary to dharma is not the characteristic that calls for putting an equation sign; there is, rather, a crucial difference, insofar as one give information that must be taken literally and the other is for the purpose of inspiration.

The key distinction in attitude between Mīmāṃsā and Advaita Vedānta was not about the Upaniṣads being arthavāda, but about what kind of Self they presented: A Self that is essentially a ritual agent and an enjoyer, or a single aloof Self, one for all. A corollary to this concerned the status of passages that talk about liberation from saṃsāra: are they true statements of result, or just statements of praise? We will see this conflict already in the next chapter.

Footnotes and references:

[back to top]

[1]:

See Tachikawa, Bahulkar and Kolhatkar 2001 for a ritual modeled on the Darśa-pūrṇamāsa and a corresponding prayoga-vākya serving as a manual of its performance, the Pavitreṣṭi-prayoga.

[2]:

A good overview is available in Pandurangi 2006:177-8.

[3]:

Kumārila’s definition, which he then expands on, says: “A sentence is an injunction when [some causal relationship] absolutely does not obtain [by other means of knowing]. A restriction happens when there is [general] optionality, whereas an exclusion when such optionality obtains in regard both to one and another.”—vidhir atyantam aprāpte niyamaḥ pākṣike sati | tatra cānyatra ca prāpte parisaṅkhyeti kīrtyate. Kumārila’s Tantra-Vārttika on Śabara’s Mīmāṃsā-Sūtra-Bhāṣya 1.2.34, I.152.

[4]:

tatra yo ‘tyantam aprāpto na ca prāpsyati prāg vacanād ity avagamyate tatra niyogo śuddha eva vidhir. Kumārila’s Tantra-Vārttika on Śabara’s Mīmāṃsā-Sūtra-Bhāṣya 1.2.42.

[5]:

Taber 1989:149.

[6]:

Ibid, 145.

[7]:

See Śabara ad Mīmāṃsā-Sūtra attributed to Jaimini 1.2.10.

[8]:

sādhanānurūpa-sādhyotpatti-dvāreṇa kṣipra-devatā-sādhyaṃ karma kṣipram eva phalaṃ dāsyati. Kumārila’s Tantra-Vārttika on Śabara’s Mīmāṃsā-Sūtra-Bhāṣya 1.2.7, I.115.

[9]:

This section is based on Kumārila’s Tantra-Vārttika on Śabara’s Mīmāṃsā-Sūtra-Bhāṣya on Mīmāṃsā-Sūtra attributed to Jaimini 1.2.7

[10]:

buddhi-pūrva-kāriṇo hi puruṣā yāvat praśasto ‘yam iti nābudhyante, tāvan na pravartante. Kumārila’s Tantra-Vārttika on Śabara’s Mīmāṃsā-Sūtra-Bhāṣya 1.2.7, I.114.

[11]:

The fourth part of the Veda, Nāmadheya or names, refers to texts which give references to particular sacrifices through their names, and seems to have been posited as a category just to avoid double injunctions. See Jha 1964:182-6.

[12]:

For a reliable study of arthavāda in Mīmāṃsā, see Harikai 1994. For a short overview, see Jha 1907:xxxv; for a longer overview, Jha 1964:177-182.

[13]:

1991:150.

[14]:

2005:38.

[15]:

Ibid., p.63.

[16]:

1992:34.

[17]:

etena kratvartha-kartṛ-pratipādana-dvāreṇopaniṣadāṃ nairākāṅkṣyaṃ vyākhyātam. Kumārila’s Tantra-Vārttika on Śabara’s Mīmāṃsā-Sūtra-Bhāṣya 1.2.7, I.114.

[18]:

Kumārila’s Śloka-Vārttika Ātmavāda 141, 148.

[19]:

yāś caitāḥ pradhāna-puruṣeśvara-paramāṇu-kāraṇādi-prakriyāḥ sṛṣṭi-pralayādi-rūpeṇa pratītās tāḥ sarvā mantrārthavāda-jñānād eva. Kumārila’s Tantra-Vārttika on Śabara’s Mīmāṃsā-Sūtra-Bhāṣya 1.3.2, I.168.

[20]:

sarvatra hi tad-balena pravartate, tad-uparame coparamatīti vijñāna-mātra-kṣaṇa-bhaṅga-nairātmyādi-vādānām apy upaniṣad-arthavāda-prabhavatvaṃ viṣayeṣv ātyantikaṃ rāgaṃ nivartayitum ity upapannaṃ sarveṣāṃ prāmāṇyam. Ibid.

[21]:

tatra lokārthavādopaniṣat-prasūtais tarka-śāstraiḥ sarva-vipratipatti-mukha-pradarśanam. Ibid.

[22]:

Aklujkar 1991.

[23]:

1983:363.

[24]:

Kumārila’s Śloka-Vārttika Sambandhākṣepa-Parihāra 83-4. The whole chapter is one massive statement against a first principle of the world.