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.
One of the things I’ve always found most intriguing about Mīmāṃsā is the way they are nearly “fundamentalist” in their commitment to scripture… yet without a deity involved! It’s a combination that would be unimaginable in the West. And it is intriguing to see how they deny even the deities mentioned in the scripture they revere.
I completely agree, it’s one of the things I find most fascinating too.
And so, from a distance, it was observed: “In the beginning was the Word, and the Word was with God, and the Word was God, and the Word became flesh and dwelt amongst us,” in the form of the sacrificial animals. Thus John the Baptist introducing his Gospel.
Now in a philosophical vein I’m not convinced that questions of motivation or effect are at all relevant: rather the process was always about *living the meaning of what was remembered as well said, for in *lived meaning you have something intrinsically of the nature of value.
Hi Elisa, thanks for posting this. As I’m sure you already know, Kumārila’s discussion of “Mahendra” is relevant, too, in which he argues, pace Śabara, that the compound does not refer to a different deity from Indra. Because of vyapadeśa, “Indra” and “Mahendra” can be used interchangeably, although, I think, not without some difference in their semantic contributions, since one is just “Indra” and one is “Indra” qualified by greatness. (So not strict synonymy.)
At one point in his discussion, Kumārila says
ekas tāvad ātikleśair indro ’stīty avagamyate |
mahendras tv aparaḥ kalpyaḥ punaḥ kleśāntarair bhavet ||
(See Subbhaśāstri’s MD, vol 2, page 421)
Jha takes this as “we find a great difficulty in believing in the existence of Indra; and it would entail a much greater difficulty to assume the existence of another and analtogether distinct deity in the shape of Mahendra” (page 549 in his Tantravārttika, https://archive.org/details/in.ernet.dli.2015.282920/page/n719/mode/1up)
In what follows, though, Kumārila adds, after explaining the principle that we shouldn’t multiply meanings for a single word even when we can see the object (and even more so when we have to posit its existence, like apūrva, devatā, etc):
tad ihendraśabdaprayogānyathānupapattyendras tāvad adṛṣṭo ‘pi kalpyate, tatra yadi punar mahendraśabdaprayogānyathānupapattir bhavet tato ‘rthāntaraṃ mahendro nāma kalpyeta.
He seems a bit unwilling to be pinned down: some artha needs to be posited to make sense out of the Vedas, but it is difficult to understand the existence of these deities (he doesn’t say it is impossible, though). And in another relevant passage, at MS 1.4.24, he explains “A brahmin is indeed Agni” by saying that, even if the event of a birth of both brahmins and Agni from Prajāpati never happened (using the word “vastu,” something like state of affairs?), since it’s described in another arthavāda, that can be the basis for the similarity between brahmins and Agni. But it’s couched in a conditional that suggests he might be able to wiggle out of denying the birth never happened. It’s just, well, even *if* it didn’t, we have no problem.
Kiyotaka Yoshimizu has a nice, short paper on mahendra (https://www.jstage.jst.go.jp/article/ibk/62/3/62_KJ00009323471/_pdf/-char/ja).
Anyway, lots of interesting questions for Kumārila’s compositional semantics here, if these terms don’t refer, in addition to the question of deities.
Many thanks, Malcolm. I am in the end-of-the-academic-year mode already (because of the various workshops planned for December), but I will definitely go back to your suggestions. Thank you again!
One more note, as I’ve been looking at Kumārila’s Ṭupṭīkā ad MS 10.4.23, p. 1925ff (Śubbāśāstrī 1929 edition) for another project. This is a section where Śabara says explicitly that he’s fine if the devatā is “śabda eva,” as this doesn’t impact the Mīmāṃsā’s view negatively. Interestingly, Kumārila doesn’t comment explicitly on this section of text, and I think he implicitly rejects it, maybe in light of a worry from an opponent.
I believe there are three positions here:
Opponent 1 says synonyms are acceptable in mantras (so “Indra” and “Mighty” are equally okay, since they both refer to Indra) but notes that that the word “Indra” has a dual purpose, bringing to mind the word itself and the object.
Opponent 2 says the word “Indra” refers to the object only, as the object is connected to the kārya. And so synymonyms are fine. But they add something that targes Śabara’s view (that “Indra” plays only a linguistic role, identifying which deity an action is done for, without that entailing there is some object that receives the action):
yadi śabda eva devatā bhavet tathā sati devatāpadam arthaśūnyaṃ syāt | arthaśūnye ca tasmin, asyaśabdena yo’rtha ucyate, sa eva kevalaḥ syāt | tatra sākāṅkṣatvād anirākāṅkṣīkṛtatvāc ca vākyam eva na syāt.
If the deity were the word only, then the deity-word would have an empty referent. And were the referent empty in this case, then what is expressed by “asya” (of it) would be independent. And because since it has expectancy and that expectancy hasn’t been removed, there would not be a sentence.
Kumārila doesn’t respond at this point, interestingly (and I don’t think what I’ve quoted is in his voice; I think it is an opponent). But he says later, siddhāntas tu agniśabda evāṣṭākapālena saṃbadhyamāno’rthāt tasya devatātvaṃ pratipādayati, “But the settled view is just the word ‘Agni’ itself is connected with the eight-cake-pan, of which its having a deity (as its indirect object) is communicated by implication.” What he seems to say (though I’d be interested to know if people disagree) in this short section is that the mantra’s specific language is what is commanded, but it is commanded because there is first a deity, although what is primary in the intention and command is the language. And so, the meaning of “Agni” in the context of a mantra is not just the word but also its object. This is in contrast to accepted “śabda eva” as Śabara does, and it responds to the worry that we have some semantic failure:
tatra prathamam arthasya devatātvam | paścāc ca śabdoccāraṇaṃ vidhīyata iti śrutyarthābhyām arthadvayaṃ bodhyate |
“In this case, first, there is a deity for the object. And afterwords, the utterance of the word is commanded. Thus, one understands two meanings: the scripture and the referent.”
Of course it could be that “artha” here is not a mind-independent divine being, but he does seem to push back against the śabda-only view.
I’m writing in a small comment box, so I’ll leave it here. Curious what others think if they check out the passage.