What is “perception”? For Buddhist epistemologists, it includes:
- sense-perception (indriyapratyakṣa)
- yogic perception
- self-awareness of cognitions (svasaṃvedana)
- mānasapratyakṣa (used to access immediately preceding cognitive events)
For Nyāya epistemologists, it includes:
- sense-perception (indriyapratyakṣa)
- yogic perception
- mānasapratyakṣa (used to access immediately preceding cognitive events) (including anuvyavasāya)
svasaṃvedana is refused, because cognitions are not transparent. Instead, they are perceptible through anuvyavasāya, which is a form of mānasapratyakṣa.
For Mīmāṃsā epistemologists, it includes:
- sense-perception (indriyapratyakṣa)
Please notice that sense-perception includes, for all Sanskrit philosophers I am aware of, six senses, one of which is manas, which is meant to grasp internal events, typically sukha or duḥkha.
Now, where does Kumārila’s ahampratyaya ‘cognition of the I’ fall into? It cannot be a case of mānasapratyakṣa, because this one is not accepted as a separate way to accept cognitive events (for thorough refutations of it from a Mīmāṃsā point of view, one can check Śālikanātha’s Pramāṇapārāyaṇa, pratyakṣapariccheda or Vācaspati’s Nyāyakaṇikā, on VV chapter 8). It could be a case of indriyapratyakṣa, with the manas working as the sense faculty, but this is odd, given that manas as a sense faculty should grasp a sensory object, like sukha or duḥkha and it is unclear how the aham could qualify as one.
Okay that’s helpful, thanks Elisa. I think I’m seeing better where you’re coming from now. I think we’re inching forward and some mutual understanding is emerging.
Let’s focus on the distinction you are making between:
– The Nyāya view, in which perception includes indriyapratyakṣa and mānasapratyakṣa
– The Mīmāṃsā view, in which perception includes only indriyapratyakṣa, not mānasapratyakṣa
I think this makes it easy to understand why I was attributing to you (https://indianphilosophyblog.org/2026/04/17/is-there-a-seeming-contradiction-in-kumarilas-account-of-i-cognition/) the view that Mīmāṃsā / Kumārila denies that mānasapratyakṣa is perception. You hadn’t put it as baldly as this before, but I had identified it as an underlying presupposition necessary to justify your claim that there is a contradiction in Kumārila if he accepts I-cognition as perception. (If he accepts mānasapratyakṣa, then I-cognition can be mānasapratyakṣa so there is no contradiction.) I followed this up by saying that I didn’t think you could “really think that Kumārila denies that mānasa-pratyakṣa is pratyakṣa”: https://indianphilosophyblog.org/2026/04/17/is-there-a-seeming-contradiction-in-kumarilas-account-of-i-cognition/.
Now this latest response, where you clarify your position by making this contrast
– Nyāya: perception includes indriyapratyakṣa and mānasapratyakṣa
– Kumārila: perception includes only indriyapratyakṣa, not mānasapratyakṣa
certainly makes it look like you’re denying that Kumārila includes mānasapratyakṣa in perception. But it’s clear that in one sense you’re denying it and in another sense you’re not. You’re distinguishing “manas-pratyakṣa” from “mānasapratyakṣa”, and saying that Kumārila accepts the former but denies the latter.
Here is my more detailed understanding of what you’re doing in this latest response.
You are
– making/seeing a firm distinction between indriyapratyakṣa and mānasapratyakṣa
– placing cognition of pleasure and pain within the category of indriyapratyakṣa, not mānasapratyakṣa
– reserving the term mānasapratyakṣa for the kind of cognition that accesses the immediately preceding cognition
– using the term manas-pratyakṣa for perception of pleasure and pain
– making a firm distinction between manas-pratyakṣa, which is a kind of indriyapratyakṣa, and mānasapratyakṣa, which is not a kind of indriyapratyakṣa
That raises the following questions for me.
1. Does this reflect Sanskrit usage?
2. If it doesn’t, does it matter? Does this new terminology have some advantages?
3. If we put this together with your earlier claim (that the I cannot be perceived, because indriyapratyakṣa is the only kind of perception for Kumārila), where are we now?
First question:
Wouldn’t you say that this goes against Sanskrit usage? Not that that is a problem if there is some benefit to seeing things in this way.
I’m guessing you would agree that this does not reflect Sanskrit usage, in that:
– we don’t find in Sanskrit usage a terminological distinction between manas-pratyakṣa and mānasapratyakṣa
– we do not find perception of pleasure etc. exclusively referred to as indriyapratyakṣa
– we commonly find perception of pleasure etc. named as mānasapratyakṣa
– we commonly find perception of the self referred to as mānasapratyakṣa, even though it does not access the immediately preceding cognition
– the things you are separating out usually get lumped into mānasapratyakṣa, this commonly being the term used to describe (1) perception of pleasure, etc., (2) perception of I, and (3) perception of the immediately preceding cognition.
– the things you are lumping together as indriyapratyakṣa – perception of an external object and perception of pleasure etc. – are frequently distinguished from each other through the former being referred to as indriyapratyakṣa and the latter as mānasapratyakṣa.
When indriyapratyakṣa and mānasapratyakṣa are juxtaposed in Sanskrit sources, wouldn’t you agree that A is much more common than B? –
A. indriyapratyakṣa means perception by means of the 5 external faculties and mānasapratyakṣa means perception by means of the manas
B. indriyapratyakṣa means perception by means of the 6 faculties and mānasapratyakṣa means preception of a preceding cognition.
I’ve never actually seen B. Here are a few examples of A:
1. Nyāyamañjarī I 484,10–13:
tathā hīśvarasadbhāvo na pratyakṣapramāṇakaḥ /
na hy asāv akṣavijñāne rūpādir iva bhāsate //
na ca mānasavijñānasaṃvedyo ‘yaṃ sukhādivat /
yoginām aprasiddhatvāt na tatpratyakṣagocaraḥ //
“To explain: The existence of God is not supported by perception, for He does not appear in sense-perception in the way that something like colour does. And He is not known through mental (mānasa-) perception in the way that pleasure etc. are. [And] He is not accessed by yogic perception, because Yogins [with the power of extra-sensory perception] are not commonly accepted.”
So we here we have a triple division of perception into sense-perception (akṣavijñāna, indriyapratyakṣa), mental (mānasa-) perception and yogic perception – the three types that you list in your post as accepted by Naiyāyikas. Jayanta understands them in the way that strikes me as typical:
– sense-perception means perception by means of the 5 external senses, the standard example being perception of colour
– mental perception means perception by means of the manas, the standard example being perception of pleasure
But your understanding of this triple division between sense-, mental and yogic is different in that you see sense-perception as including perception of pleasure and pain, and mental perception as only including awareness of a previous cognition.
2. You might say that Jayanta’s is a Naiyāyika perspective, and that Kumārila, unlike him, would see sense-perception as including perception of pleasure and pain, and mental perception as only including awareness of a previous cognition? But it is actually a Bhāṭṭa Mīmāṃsaka speaking there (as can be seen from the rejection of yogic perception); Jayanta is representing Kumārila’s view as faithfully as he can. And see Śālikanātha’s characterization of the Bhāṭṭa view at Prakaraṇapañcikā, p. 333, 1–3:
tatra ke cid āhuḥ—mānasaṃ pratyakṣaṃ sukhādiṣv ivātmani pramāṇam iti
“Mental perception is the means of knowing the self, as it is the means of knowing pleasure and the like.”
Here too we find that
– Kumārila is depicted as accepting mānasaṃ pratyakṣaṃ
– the standard example of this mānasaṃ pratyakṣaṃ is perception of pleasure and the like.
For you mānasa-cognition is something altogether different from the kind of cognition through which we know pleasure and the like. But here it is exactly that. It is not a name only for the kind of cognition that accesses the previous cognition.
3. Throughout his discussion of recognition (Nyāyamañjarī II 307–336), Jayanta distinguishes two views – that recognition (pratyabhijñā) is sense-perception (indriyajaṃ jñānam) or that it is mental (mānasa): see pp. 1394,129; 1396,144; 1396,147; 1396,148 in the edition in Watson (2022). See Nilanjan Das (2018) and Watson (forthcoming) for a discussion of these two views.
That indriya is used throughout this passage to refer exclusively to the 5 external sense-faculties – to contrast them with the manas – is beyond doubt. See for example §2.3.4 and §2.4 (in the edition in Watson 2022). If indriya were being used to refer to the 6 sense-faculties, i.e. to include the manas, the intended contrast between the indriyaja view and the mānasa view would not come across.
4. When Kamalaśīla is defending, against non-Buddhists, his view that pleasure etc. are brought to awareness by svasaṃvedana, he names the means by which pleasure is perceived for the non-Buddhists as mānasa-cognition, not indriya-cognition; see TSP ad TS 1332–1339:
sukhādīnāṃ mānasenaiva cetasā vedyatveneṣṭatvāt …
sukhādīnāṃ mānasenaiva cetasā vedyatvenābhyupagamāt.
If indriya-cognition meant perception by means of the 6 faculties and mānasa-cognition meant perception of a preceding cognition, then we would expect perception of pleasure to be termed indriya-cognition, but it is not: it is termed mānasa-cognition.
5. If your definitions of indriyapratyakṣa and mānasapratyakṣa were the way those terms were used, then we would expect I-cognition to be classified by its advocates as indriyapratyakṣa. But it is not. It is classified as mānasapratyakṣa. That should mean, sticking to your usage, that I-cognition perceives an immediately preceding cognitive event. But none of the upholders of I-cognition thought that it was perceiving an immediately preceding cognitive event.
6. Concerning the use of the term indriya in Nyāya sources:
The manas is often treated as something over and above the ‘indriyas’, something not included among the referents of the word indriya. For examples of this use of indriya, where it excludes the manas, see:
– Nyāyasūtra 1.1.12: it lists the indriyas and leaves out the manas.
– Vātsyāyana ad 1.1.4 defends the fact that the Naiyāyika definition of perception includes the qualifier “is definitive” (vyavasāyātmakam), on the grounds that otherwise doubtful cognitions would count as perception. For the other qualifiers in the definition, such as “arises from contact between indriya and object”, apply to doubtful cognitions. To avoid the definition undesirably including doubtful cognitions, we need it to include the stipulation that perception must be “definitive”, i.e. non-doubtful. Vātsyāyana sees this as inviting the following objection, which he articulates and then answers: But doubtful cognitions do not arise from contact with an indriya, they arise from contact with the manas. (So they would not be included by the definition even if it lacked the qualifer ‘is definitive’.) Clearly this would not be a cogent objection if the word indriya was understood to include the manas among its referents.
– Further on in the same sūtra, Vātsyāyana articulates the following objection. The stipulation in the definition that perception must ‘arise from contact between indriya and object’ makes the definition too narrow, because it excludes perception of the self and perception of pleasure, given that neither of them are produced by contact with an indriya (ātmādiṣu sukhādiṣu ca pratyakṣalakṣaṇaṃ vaktavyam, anindriyārthasannikarṣajaṃ hi tat). A good definition of perception must apply to perception of pleasure and perception of the self, but the requirement of indriya-object contact excludes these two valid forms of perception. Vātsyāyana’s response is that the faculty that brings about perception of the self and of pleasure is the manas – implying that the manas is a kind of indriya; but again the existence of the objection assumes an understanding of the meaning of indriya as excluding the manas.
– Vātsyāyana ad 1.1.9 defines the manas in opposition to the indriyas, on the grounds that indriyas are not capable of apprehending all objects, but the manas is: sarvārthopalabdhau nendriyāṇi prabhavantīti sarvaviṣayam antaḥkaraṇaṃ manaḥ.
These passages do not negate the existence of other passages in which the manas is said to be a kind of indriya, but they do support the idea that when Naiyāyikas claim to accept indriyapratyakṣa and mānasapratyakṣa, they do not – as stated in your post – include perception of pleasure and pain within indriyapratyakṣa; they include it within mānasapratyakṣa.
***
So what we find in these 6 kinds of passages are:
1. mānasapratyakṣa rather than indriyapratyakṣa being used for perception of pleasure and pain.
2. mānasapratyakṣa rather than indriyapratyakṣa being used for perception of the self
3. indriyapratyakṣa being used to refer exclusively to perception by means of the 5 external sense-faculties, in order to contrast with mānasapratyakṣa
Your restriction of mānasapratyakṣa to only cognitions that “access immediately preceding cognitive events” strikes me as counter to Sanskrit usage. Have you seen it defined in such a narrow way?
The way I see it, which I think reflects Sanskrit usage more:
– both Nyāya and Mīmāṃsā accept indriyapratyakṣa (in the sense it is used in the passages above: perception by means of the 5 external senses) and mānasapratyakṣa (in the sense it is used in the passages above: perception by means of the manas)
– but there is one thing that Nyāya includes in mānasapratyakṣa that Mīmāṃsā rejects: anuvyavasāya.
In other words, Kumārila and his followers accept mānasapratyakṣa, but assign it a more limited role. If you could provide me with an example of Kumārila saying that mānasapratyakṣa is not valid, then I’ll reconsider.
***
Perhaps you will say (?) that although
– mānasapratyakṣa is not usually defined as perception of a previous cognition
– you have not seen passages where Kumārila and his followers deny the validity of mānasapratyakṣa in all its forms
there are advantages to stating things in this way.
If so, what do you think the advantages are in comparison to characterizinig things in this way:
– Kumārila accepts mānasapratyakṣa
– he denies svasaṃvedana
– he rejects the Buddhist understanding of mānasapratyakṣa
– he denies one subtype of mānasapratyakṣa that the Naiyāyikas accepted, namely anuvyavasāya
– he accepts the other subtypes of mānasapratyakṣa: perception of pleasure and pain, etc.*
*For another kind of mānasa-cognition that he takes to be valid – simultaneous apprehension of all of the phonemes making up a word after cognizing the last phoneme of the word – see Ślokavārttika sphoṭavāda 113abc: sarveṣu caivam artheṣu mānasaṃ sarvavādinām / iṣṭaṃ samuccayajñānaṃ. And how about conceptual perceptions / perceptual judgements (such as “this is a pot”); aren’t they a kind of mānasapratyakṣa accepted by Kumārila (or does he think their primary instrument is the external sense-faculty)?
***
Das, Nilanjan (2018) “Object reidentification and the epistemic role of attention.” Ratio 31: 402–414. https://doi.org/10.1111/rati.12214
Watson, Alex (2022) ‘Pratyabhijñā: Recognition’s Nature, Cause and Object. Critical Edition and Annotated Translation of a Portion of the Nyāyamañjarī’ in Francesco Sferra and Vincenzo Vergiani (eds.) Verità e bellezza. Essays in honour of Raffaele Torella. Series Minor XCVII.1–2, Università di Napoli “L’Orientale”, Dipartimento Asia, Africa e Mediterraneo, Volume II, pp. 1325–1398. Naples: Unior Press. ISBN 978-88-6719-209-0.
Watson, Alex (Forthcoming) ‘Jayanta on whether recognition can refute momentariness’ in P. Bilimoria, A. Rostalska, D. Singh & R. K. Sharma (eds.) Conceptualizing Categories: Texts and Context in Indian Philosophy. Festschrift in Honor of Professor ShashiPrabha Kumar. Sophia Studies in Cross-cultural Philosophy of Traditions and Cultures. Cham, Switzerland: Springer Nature.
I said in my previous response: “If you could provide me with an example of Kumārila saying that mānasapratyakṣa is not valid, then I’ll reconsider.” But your latest post does actually give two sources for the idea that Kumārila and his followers reject mānasapratyakṣa: Śālikanātha’s Pramāṇapārāyaṇa, pratyakṣapariccheda, and Vācaspati’s Nyāyakaṇikā on Vidhiviveka 8. You know both of those better than I do, but I had a quick look to see if I could indeed find evidence there of a Bhāṭṭa rejection of mānasapratyakṣa.
Śālikanātha’s Pramāṇapārāyaṇa
– Surely this is not a source for the Bhāṭṭa view of mānasapratyakṣa, but for the Prābhākara view?
– Śālikanātha himself does not accept mānasapratyakṣa as a way of perceiving either cognition or self, because for him cognition is self-luminous, and it can perceive both cognition and self, so there is no need for mānasapratyakṣa – for a second mental perception after self and cognition have already anyway been perceived in the first moment. But for Kumārila neither cognition nor self are self-luminous, leaving room for other pramāṇas to bring about awareness of these two: arthāpatti for cognition and mānasapratyakṣa for the self. There are two further, resultant differences: (1) For Śālikanātha, the self is only ever perceived as subject (grāhakatvenaiva), never as object (grāhyatayā). I-cognition, being a kind of mānasapratyakṣa, perceives the self as an object. (2) For Śālikanātha the self is perceived non-conceptually; for Kumārila and his followers it is perceived conceptually.
– Apart from that section, the other sections where Śālikanātha discusses mānasapratyakṣa seem to be targetting the Buddhist view of it.
– I’m therefore yet to find evidence in this text that the Bhāṭṭas rejected mānasapratyakṣa.
Vācaspati’s Nyāyakaṇikā on Vidhiviveka 8
Here I find evidence that Vācaspati does accept mānasapratyakṣa. See for example:
– āntare ca sukhādau tat svatantram iti sukhādiviṣayaṃ tat (where tat is picking up mānasam pratyakṣaṃ)
– tasmāt sukhādiviṣayaṃ mānasaṃ
both of which also affirm that, for Vācaspati, cognition of pleasure, etc., takes place by mānasapratyakṣa, not by indriyatyakṣa.
The context is a critique of omniscience, and what I find here is a refutation of mānasapratyakṣa’s ability to know everything. It is capable of knowing sukhādi, but not everything. Surely there is no critique here of the validity of mānasapratyakṣa per se?
In both of these texts it looks like the bulk of the treatment of mānasapratyakṣa is taken up with long critiques of the Dharmakīrtian doctrine that mānasapratyakṣa arises immediately after a moment of indriya-cognition, has as its samanantarapratyaya that indriya-cognition, and has as its sahakāri the next moment of the object.
So is it possible that one or some of the following have led you to your view that Bhāṭṭas reject mānasapratyakṣa? –
1. you find Bhāṭṭa arguments against the validity of Buddhist mānasapratyakṣa, and take them to mean that Bhāṭṭa Mīmāṃsā is against mānasapratyakṣa more widely.
2. you find Bhāṭṭa arguments against the validity of Nyāya anuvyavasāya, and take them to mean that Bhāṭṭa Mīmāṃsā is against mānasapratyakṣa more widely.
3. you find a tendency in Prābhākara authors to reject mānasapratyakṣa as a separate subtype of pratyakṣa, and to dissolve perception by means of the manas into indriyapratyakṣa, and you extrapolate this tendency to the Bhāṭṭas too.
But (1) the Buddhist view of mānasapratyakṣa is eccentric. Rejecting it does not entail rejecting all forms of mānasapratyakṣa. (2) Kumārila and Vācaspati reject Nyāya’s anuvyavasāya not by denying the validity of mānasapratyakṣa, but by asserting the imperceptibility of cognition (the object of anuvyavasāya). As we saw in the passage above, Vācaspati explicitly accepts mānasapratyakṣa. (3) Surely what Prābhākara authors accepted is not a pramāṇa for what Bhāṭṭa authors accepted.
Typo in my previous response:
“… both of which also affirm that, for Vācaspati, cognition of pleasure, etc., takes place by mānasapratyakṣa, not by indriyatyakṣa”
Replace the last word with:
indriyapratyakṣa
My first response above said that your post raised 3 questions for me. I now come to the 3rd question: If we put this latest post of yours together with your earlier claim (that the I cannot be perceived, because indriyapratyakṣa is the only kind of perception for Kumārila), where are we now?
This latest post of yours, if I understand you correctly, is intended to be a defence of your original claim (https://indianphilosophyblog.org/2026/02/16/does-kumarila-accept-i-cognition-as-a-kind-of-perception/). Let’s call that claim “Elisa 1”: There is a seeming contradiction in Kumārila. The contradiction is between:
– the claim that I-cognition is perception
– the claim that only indriyapratyakṣa is perception
“Alex 1” (https://indianphilosophyblog.org/2026/04/17/is-there-a-seeming-contradiction-in-kumarilas-account-of-i-cognition/)
involved laying out a dilemma for you: (A) If by indriyapratyakṣa you mean perception by means of one of the 6 faculties, there is no contradiction; I-cognition can be perception by means of the manas. (B) Since you think there is a contradiction, you must mean, by indriyapratyakṣa, perception by means of one 5 external sense-faculties. But then your claim that Kumārila regards only indriyapratyakṣa as perception means that you think he denies that cognition of pleasure and pain by means of the manas is perception. Yet we know that he does not deny this.
In “Elisa 2” (17 April 2026 response here https://indianphilosophyblog.org/2026/04/17/is-there-a-seeming-contradiction-in-kumarilas-account-of-i-cognition/)
you say that you take indriyapratyakṣa to include perception by means of the manas.
“Alex 2” (22 April 2026 response to “Elisa 2”)
reasserted the dilemma. The fact that you take indriyapratyakṣa to include perception by means of the manas means you are taking horn A of the dilemma. So you can avoid the unwanted consequence associated with horn B: You are not subject to the view that Kumārila denies that cognition of pleasure and pain by means of the manas is perception. But the fact that you take horn A means you are still subject to the unwanted consequence of that horn: There is no seeming contradiction in Kumārila. By opting for horn A, you seem to be contradicting the main point of your post, that there is a seeming contradiction in Kumārila holding the self to be perceptible. If indriyapratyakṣa includes perception by means of the manas, what contradicts perception of the self?
In order for there to be a contradiction in Kumārila, you would have to take indriyapratyakṣa to exclude perception by means of the manas. But the fact that you take indriyapratyakṣa to include perception by means of the manas contradicts the claim that there’s a contradiction.
Your last post, as I understand it, is your explanation of why your position is not subject to this conundrum. It is your explanation of how you can maintain that there is a seeming contradiction in Kumārila. So let’s consider how / whether it does indeed enable you to continue to maintain that these two are contradictory:
(1) For Kumārila perception is only indriyapratyakṣa, which means perception by means of one of the 6 sense-faculties
(2) The self is perceived
Using the terminology you introduce in your latest post, (1) comes out as:
(1*) For Kumārila perception is only manas-pratyakṣa plus perception by means of one of the 5 external sense-faculties
But how does this new terminology help to justify the contradiction or to move things forward? Even using the terminology in (1*), we are still in the position that you are maintaining a contradiction between two things that are surely not contradictory: the self being perceived and perception being only that which comes about either by the 5 external senses or the manas.
It doesn’t seem to me that the new terminology helps. But there is one other claim in (the last sentence of) your latest post that does help. Let me build up to that:
Even though there is no contradiction between either (1) or (1*) and the claim that the self is perceived, perhaps you will want to say that perception of the self is contradicted by (1) / (1*) in combination with a further claim. Is it contradicted by the combination of these two?
(1) only indriyapratyakṣa is perception
(2) indriyapratyakṣa (although including perception by means of the manas, e.g. of pleasure and pain) excludes the kind of mānasapratyakṣa that involves perceiving the previous cognition.
It would only be contradicted by these two if perceiving the self meant perceiving the previous cognition. But no one equated perceiving the self with perceiving the previous cognition. I have got the impression from some of your posts that you think perception of the self, if it happens, will have to resemble perception of the previous cognition more than it will resemble perception of pleasure. If that were the case, then I see there would be a difficulty (I would use a word like this rather than “contradiction”) in accepting perception of the self. But I don’t think Kumārila, other Mīmāṃsakas, or any upholders of perception of the self, thought that perception of the self resembles perception of the previous cognition more than it resembles perception of pleasure.
What I’m guessing from your latest response is that you would explain the contradiction in the following way. Perception of the self is contradicted by a combination of these claims:
(1) only indriyapratyakṣa is perception
(2) indriyapratyakṣa is only capable of perceiving “a sensory object” (taken from the last sentence of your post)
(3) pleasure, pain and external objects are sensory objects; the self is not a sensory object
Have I reconstructed your thought correctly?
If so, I think that what is carrying the weight here is just the claim that the self is not a sensory object – not something that is suitable to be brought to awareness by either the external faculties or the manas. If we want to hang on to the word “contradiction”, we could say there is a contradiction between the perceptibility of the self and the fact that the self is not a sensory object. But I don’t see that Kumārila’s rejection of mānasapratyakṣa, which for you means perception of the previous cognition, is relevant at all: perception of the self is not perception of a previous cognition.
So if you were willing to drop your original claim
– There is a contradiction between perception of the self and the exclusion of mānasapratyakṣa
and replace it with something like
– There is a contradiction between perception of the self and the fact that the self is not a sensory object
then we would have reached a point of agreement.
Now the issue becomes whether the self is “a sensory object”. It’s of course question begging to assume it’s not against someone who thinks the self is perceptible. Is there a non-question-begging way of defending that it is not? Potentially, yes. One could argue, for example, that the self has more in common with the previous moment of cognition than it has with pleasure or pain. But I’m not sure what reasons one could give in support of that.
My first response above said that your post raised 3 questions for me. I now come to the 3rd question: If we put this latest post of yours together with your earlier claim (that the I cannot be perceived, because indriyapratyakṣa is the only kind of perception for Kumārila), where are we now?
This latest post of yours, if I understand you correctly, is intended to be a defence of your original claim (https://indianphilosophyblog.org/2026/02/16/does-kumarila-accept-i-cognition-as-a-kind-of-perception/). Let’s call that claim “Elisa 1”: There is a seeming contradiction in Kumārila. The contradiction is between:
– the claim that I-cognition is perception
– the claim that only indriyapratyakṣa is perception
“Alex 1” (https://indianphilosophyblog.org/2026/04/17/is-there-a-seeming-contradiction-in-kumarilas-account-of-i-cognition/)
involved laying out a dilemma for you: (A) If by indriyapratyakṣa you mean perception by means of one of the 6 faculties, there is no contradiction; I-cognition can be perception by means of the manas. (B) Since you think there is a contradiction, you must mean, by indriyapratyakṣa, perception by means of one 5 external sense-faculties. But then your claim that Kumārila regards only indriyapratyakṣa as perception means that you think he denies that cognition of pleasure and pain by means of the manas is perception. Yet we know that he does not deny this.
In “Elisa 2” (17 April 2026 response here https://indianphilosophyblog.org/2026/04/17/is-there-a-seeming-contradiction-in-kumarilas-account-of-i-cognition/)
you say that you take indriyapratyakṣa to include perception by means of the manas.
“Alex 2” (22 April 2026 response to “Elisa 2”)
reasserted the dilemma. The fact that you take indriyapratyakṣa to include perception by means of the manas means you are taking horn A of the dilemma. So you can avoid the unwanted consequence associated with horn B: You are not subject to the view that Kumārila denies that cognition of pleasure and pain by means of the manas is perception. But the fact that you take horn A means you are still subject to the unwanted consequence of that horn: There is no seeming contradiction in Kumārila. By opting for horn A, you seem to be contradicting the main point of your post, that there is a seeming contradiction in Kumārila holding the self to be perceptible. If indriyapratyakṣa includes perception by means of the manas, what contradicts perception of the self?
In order for there to be a contradiction in Kumārila, you would have to take indriyapratyakṣa to exclude perception by means of the manas. But the fact that you take indriyapratyakṣa to include perception by means of the manas contradicts the claim that there’s a contradiction.
Your last post, as I understand it, is your explanation of why your position is not subject to this conundrum. It is your explanation of how you can maintain that there is a seeming contradiction in Kumārila. So let’s consider how / whether it does indeed enable you to continue to maintain that these two are contradictory:
(1) For Kumārila perception is only indriyapratyakṣa, which means perception by means of one of the 6 sense-faculties
(2) The self is perceived
Using the terminology you introduce in your latest post, (1) comes out as:
(1*) For Kumārila perception is only manas-pratyakṣa plus perception by means of one of the 5 external sense-faculties
But how does this new terminology help to justify the contradiction or to move things forward? Even using the terminology in (1*), we are still in the position that you are maintaining a contradiction between two things that are surely not contradictory: the self being perceived and perception being only that which comes about either by the 5 external senses or the manas.
It doesn’t seem to me that the new terminology helps. But there is one other claim in (the last sentence of) your latest post that does help. Let me build up to that:
Even though there is no contradiction between either (1) or (1*) and the claim that the self is perceived, perhaps you will want to say that perception of the self is contradicted by (1) / (1*) in combination with a further claim. Is it contradicted by the combination of these two?
(1) only indriyapratyakṣa is perception
(2) indriyapratyakṣa (although including perception by means of the manas, e.g. of pleasure and pain) excludes the kind of mānasapratyakṣa that involves perceiving the previous cognition.
It would only be contradicted by these two if perceiving the self meant perceiving the previous cognition. But no one equated perceiving the self with perceiving the previous cognition. I have got the impression from some of your posts that you think perception of the self, if it happens, will have to resemble perception of the previous cognition more than it will resemble perception of pleasure. If that were the case, then I see there would be a difficulty (I would use a word like this rather than “contradiction”) in accepting perception of the self. But I don’t think Kumārila, other Mīmāṃsakas, or any upholders of perception of the self, thought that perception of the self resembles perception of the previous cognition more than it resembles perception of pleasure.
What I’m guessing from your latest response is that you would explain the contradiction in the following way. Perception of the self is contradicted by a combination of these claims:
(1) only indriyapratyakṣa is perception
(2) indriyapratyakṣa is only capable of perceiving “a sensory object” (taken from the last sentence of your post)
(3) pleasure, pain and external objects are sensory objects; the self is not a sensory object
Have I reconstructed your thought correctly?
If so, I think that what is carrying the weight here is just the claim that the self is not a sensory object – not something that is suitable to be brought to awareness by either the external faculties or the manas. If we want to hang on to the word “contradiction”, we could say there is a contradiction between the perceptibility of the self and the fact that the self is not a sensory object. But I don’t see that Kumārila’s rejection of mānasapratyakṣa, which for you means perception of the previous cognition, is relevant at all: perception of the self is not perception of a previous cognition.
So if you were willing to drop your original claim
– There is a contradiction between perception of the self and the exclusion of mānasapratyakṣa
and replace it with something like
– There is a contradiction between perception of the self and the fact that the self is not a sensory object
then we would have reached a point of agreement.
Now the issue becomes whether the self is “a sensory object”. It’s of course question begging to assume it’s not against someone who thinks the self is perceptible. Is there a non-question-begging way of defending that it is not? Potentially, yes. One could argue, for example, that the self has more in common with the previous moment of cognition than it has with pleasure or pain. But I’m not sure what reasons one could give in support of that.
One more point about the argument in the last paragraph of your post. Would this be a fair summary of that paragraph:
– perceiving the self either has to be (1) like perceiving pleasure, or (2) like perceiving a previous cognition.
It can’t be either.
– It can’t be (1), because the self is not a ‘sensory object’ in the way that pleasure is
– It can’t be (2), because Kumārila does not accept perception of a previous cognition
Certainly Kumārila would agree that it can’t be (2). The reason it can’t be (2) is more than just that Kumārila does not accept perception of a previous cognition. That is somewhat irrelevant, because even if he did accept perception of a previous cognition, perception of the self would not be like that: The advocates of I-cognition, whether Mīmāṃsakas or non-Mīmāṃsakas, do not claim that it is like perceiving a previous cognition. So Kumārila would certainly agree that it can’t be (2).
I see no obstacle to him saying that it is like (1): the self is a sensory object like pleasure.
But even if we were to grant that you are right – grant that the self is not the same kind of object as pleasure, so it cannot be perceived in the way that pleasure is perceived –, then we can just defend Kumārila’s position by saying that for him the self is neither like pleasure nor like a previous cognition. It is not that for him perceiving the self either has to be (1) like perceiving pleasure, or (2) like perceiving a previous cognition. Those two possibilities are not collectively exhaustive: I-cognition can fall into a third category – a kind of manas-pratyakṣa that differs slightly from perception of pleasure. Differs from it in that
– whereas in perception of pleasure the most significant contact is that between manas and pleasure
– in I-cognition the most significant contact is that between manas and self.
***
I’ve just considered the two possibilities that, for Kumārila and his commentators, perception of the self (1) was like perceiving pleasure or (2) fell into a category distinct from either perceiving pleasure or perceiving the previous cognition. But on both of these views it is ‘a sensory object’, insofar as it is perceived by indriyapratyakṣa (as you define that term, i.e. as including manas-pratyakṣa). So we come back to the issue of whether or not it is a sensory object, whether it is a suitable object of indriyapratyakṣa. I’ll say something about that in my next response below.
In my responses above, I have argued against the existence of the seeming contradiction that you see in Kumārila – a contradiction between
– asserting that the self is perceptible
– holding that perception excludes mānasapratyakṣa (in the way you define it: perception of the previous moment of cognition)
I said that your latest post perhaps suggests that you’re not really relying on that contradiction, but rather:
– a contradiction between perception of the self and the fact that the self is not a sensory object.
So what to make of that?
The first thing to say is that we’re no longer dealing with a contradiction between two things that Kumārila says. It would only be such a contradiction if Kumārila somewhere says that the self is not a sensory object.
While I feel strongly that you are being unfair to Kumārila in accusing him of a contradiction, I do not feel so strongly opposed to your claim that the self is not a sensory object. But I do think it’s hard to argue for it in a non-question-begging way. And I do think there are some arguments one could give for the opposite conclusion: that there is nothing to disqualify the self from being a sensory object.
I was also curious to ask how far you want to push this. Do you think
1. Kumārila thought the self was not a sensory object?
or
2. Kumārila must have accepted it as a sensory object, but that is puzzling, because (a) he ignored important considerations against it being one or (b) other commitments of his go against it being one?
In the intellectual climate surrounding Kumārila, it was not uncommon for it to be accepted as a sensory object – accepted as a natural object of pratyakṣa, something suited to be perceived by the manas, in the same way that pleasure etc. are so suited.
1. There is Vaiśeṣikasūtra 9.13: ātmany ātmamanasoḥ saṃyogaviśeṣād ātmapratyakṣam (there are variant readings that do not significantly affect the meaning).
“There is direct perception of the self as a result of a particular conjunction in the self of manas and self’.”
The earliest earliest commentators on, and quoters of, the sūtra take it as referring to yogic perception. But yoga or yogic perception is not mentioned in this or the surrounding Vaiṣeśikasūtras, so it is not impossible that originally it was intended as a description of the perception of ordinary people. See Isaacson (1993) for detailed discussion. Even if it was intended to refer to yogic perception, it is clearly not the kind of yogic perception that infringes the requirement of normal perception that there be contact between a sense-faculty and the object of perception (here, manas and self).
2. When Vātsyāyana is discussing ad 1.1.3 whether one and the same object can be known through more than one pramāṇa, he gives the example of the self as something known by speech, inference and perception. The kind of perception of the self he has in mind here is yogic perception, but again there is explicit mention of faculty-object contact.
3. When Vātsyāyana is discussing the definition of perception in Nyāyasūtra 1.1.4, he comments that if it’s a good definition it must allow for perception of the self and of pleasure, etc. Therefore the definition’s requirement that there be contact between indriya and object of perception must be fulfilled by manas-object contact. That way cognition of self and of pleasure can count as perception as per the definition. So it’s clear here that the self is being deemed a perceptual object. And the grouping together of perception of the self and perception of pleasure suggests that perception of the self was envisaged as taking place in the same kind of way as perception of pleasure. That, I think, was Kumārila’s view too.
4. Apart from Jayanta, all Naiyāyikas from Uddyotakara onwards whose works survive thought that this self-perception should not be restricted to yogins but is something that all of us are undergoing every time we have an I-cognition. I-cognition, Uddyotakara explains (ad 3.1.1), is just as perceptual as perception of a colour, for both contain the two crucial hallmarks of perception: they do not rely on memory and they detect and conform to the different natures of their objects (viṣayasvabhāvabhedānuvidhāyin). The same for Śaivas – for them too I-cognition was a kind of perception involving sense-object contact (manas-ātman contact). So if you think the self is not a perceptual object, there are a lot of people who disagree with you.
Perhaps you will respond by saying something like:
– “Whether I think the self is a perceptual object is not the issue. The issue is whether Kumārila thought it. Naiyāyikas and Śaivas may have thought it, but Kumārila had certain specific commitments that should have disinclined him from going down the path of the Naiyāyikas.” What commitments? “His rejection of mānasapratyakṣa.”
Have I reconstructed your stance correctly? In case I have, here is my response:
I could understand this stance if the Naiyāyika acceptance of the self as a sensory object presupposed the validity of anuvyavasāya. In that case Kumārila’s rejection of anuvyavasāya might have, or should have, dissuaded him from accepting the self as a sensory object.
But I feel that your view (if I’m correct in attributing it to you) rests on an equivocation on the meaning of mānasapratyakṣa. You think that Kumārila’s rejection of mānasapratyakṣa should have prevented him from viewing the self as a sensory object. That is plausible if mānasapratyakṣa means perception by means of the manas. For it is indeed the manas that is the appropriate faculty for perceiving the self. But we know that Kumārila did accept perception by means of the manas, so you explain your claim that Kumārila rejected mānasapratyakṣa by defining mānasapratyakṣa as perception of cognition. Now if that is all that Kumārila’s rejection of mānasapratyakṣa means – that he denies perception of cognition – then it is no bar to him holding the self to be a sensory object.
Or to put it again in terms of a dilemma that your position is subject to:
– Either you define mānasapratyakṣa as perception by means of the manas, in which case your claim that Kumārila rejects mānasapratyakṣa is not true
– Or you define mānasapratyakṣa as perception of cognition, in which case Kumārila’s rejection of mānasapratyakṣa is no bar to him agreeing with the Naiyāyikas (and later the Śaivas) that the self is a sensory object.
I can see two ways in which rejection of anuvyavasāya would be a bar to regarding the self as a perceptual object. (1) When we perceive the self we are perceiving a previous cognition. But that is not what advocates of I-cognition thought: they thought we are perceiving something currently existing, with which the manas is in contact, analogous to the way the manas is in contact with pleasure (sukha), or the visual faculty is in contact with a colour.
(2) Even if perceiving the self did not involve perceiving a previous cognition, it might still be the case that regarding the self as a perceptual object is bound up with, or presupposes, the validity of anuvyavasāya. But it isn’t the case. When Uddyotakara and other Naiyāyikas defend the perceptibility of the self, they do not in any way appeal to the validity of anuvyavasāya.
If rejection of Nyāya’s anuvyavasāya contradicts, in some way, acceptance of perception of the self / acceptance of the self as a sensory object, how do you explain that Kumārila’s Bhāṭṭa followers accept perception of the self / accept the self as a sensory object?
One final point here. Someone might argue that since Kumārila
– rejects the perceptibility of cognition
– thinks there is a porous boundary between self and cognition (the two being related through a relation of bhedābheda)
he should reject the perceptibility of the self.
Addressing that particular argument will involve points that I have not yet made in any of these posts, but this is a different argument from ones that you have put.
Isaacson, H. (1993). “Yogic Perception (Yogipratyakṣa) in Early Vaiśeṣika.” Studien zur Indologie und Iranistik, 18, 139–160.
Since in actual Sanskrit usage indriyapratyakṣa is often used of perception by means of the 5 external senses, and mānasapratyakṣa is often used of things like perception of pleasure, I will use *indriyapratyakṣa, and *mānasapratyakṣa to denote those words as you are using them.
What I’m wondering is how far apart our positions are after this discussion; how strong a thesis you are still wanting to put forward. You say in the last paragraph of your last post that I-cognition cannot be *mānasapratyakṣa because Mīmāṃsakas did not accept that, and it cannot be *indriyapratyakṣa because it is not a sensory object. Is your current position that:
1. Neither Kumārila nor his followers accepted I-cognition to be perception
2. Kumārila did not accept I-cognition to be perception, but his followers did
3. Kumārila did not accept I-cognition to be perception while he was writing the perception chapter of Ślokavārttika, but he did when he subsequently came to write the ātmavāda chapter
4. He did accept I-cognition to be perception, but it is contradicted by other of his commitments
5. There are no strong reasons left for thinking that he denied I-cognition to be perception, and none of his other commitments contradict it
1. Your second post (https://elisafreschi.com/2026/02/26/again-on-ahampratyaya-in-kumarila-using-watson-2010-and-2020/) did concede that for Kumārila’s commentators I-cognition is mānasapratyakṣa. So I’m guessing that you reject 1.
2. If 2, what do you make of the evidence here (https://indianphilosophyblog.org/2026/04/22/is-i-cognition-for-kumarila-a-kind-of-perception/) that Kumārila accepted the self to be perceptible? (There is also Ślokavārttika śabdanityatādhikaraṇa 337–339, which I think Devansh mentioned in his talk at the Kumārila workshop in May.)
3. Re. 3, what did you make of my other explanations here (https://indianphilosophyblog.org/2026/04/22/is-i-cognition-for-kumarila-a-kind-of-perception/) for the absence of mention of I-cognition in the pratyakṣa chapter? If they are plausible, isn’t it uncharitable to interpret Kumārila as thinking the self to be imperceptible while writing the pratyakṣa chapter, and then realizing he was wrong, and changing his mind in the ātmavāda chapter? Surely he would already have read the ātmavāda section in his source text, the Śābarabhāṣya, where the self is argued to be perceptible, before writing his commentary on the pratyakṣa chapter? So if you take this option 3, I guess you think that while writing the pratyakṣa chapter he was thinking that when he got to comment on that ātmavāda section of the Śābarabhāṣya he would just ignore its claims of the perceptibility of the self? Or argue against that one part of the Śābarabhāṣya?
4. Re. 4, I argued here (https://indianphilosophyblog.org/2026/04/17/is-there-a-seeming-contradiction-in-kumarilas-account-of-i-cognition/) that there is no such contradiction, and your response is to distinguish manas-pratyakṣa and *mānasapratyakṣa, to say the self could be perceived by the latter but not the former, and to say that Kumārila denies the latter. Now since Kumārila does accept manas-pratyakṣa, and it is manas-pratyakṣa that is held to be the kind of perception that perceives the self (by all advocates of ahampratyaya), there is no contradiction. We would only be dealing with a contradiction if Bhāṭṭas
– denied *mānasapratyakṣa
– said I-cognition were a kind of *mānasapratyakṣa
But they don’t do that.
5. Would you be okay with 5, as long as it is qualified by something like: Although none of his other commitments contradict perception of the self, it is surprizing that he accepted the manas to be capable of perceiving it because it is not a sensory object?