Tuesday, February 15, 2011

Can we connect meant entities acquired through different sources in a single notion?

We can doubtlessly get a composite notion of something through sensible data acquired at different times and places. For instance, my notion of the Colosseum consists of data acquired through many cases of linguistic communication and of sense perception, through inference and perhaps also through analogy.
But the Bhāṭṭas also claim that this is simultaneously possible. I would be able to get a composite notion at once, if I were to touch and see the same thing I am hearing you talking about, for instance. This view has a linguistic consequence, too, since it makes it possible for the entities meant by words to connect and convey a sentence-meaning. On the contrary, Prābhākaras claim that only connected words can yield a unified sentence-meaning. For them, there is no unitary notion out of data acquired through different means of knowledge. Once something has been known, it can connect to something else only through a second-order epistemological operation, such as a further inference.

I happened to read an odd passage on this theme in Rāmānujācārya's Tantrarahasya. The passage runs as follows:

«On the contrary, it is commonly seen that also [meant entities] conveyed (pratipad-) by different instruments of knowledge can cause to know a composite (saṃsarga) [notion]. For instance, the composite (saṃsarga) notion “a white horse is running” is commonly seen in the case of one who perceives (pratipad-) through direct perception (pratyakṣa) a whiteness, whose substratum (āśraya) is unknown; infers through the noise (rava) of neighing (heṣā) a horse, whose visible form is [for him] unknown; and infers through the sound (śabda) of hooves’ (khura, puṭa) clashing a going, whose author is unknown. It has been said [by Kumārila]:

For one who sees a white form (ārūpa) and hears the sound of neighing |
and the sound of hooves clashing, the notion (dhī) “a white horse is running” ||
is commonly seen, [even] without a sentence |
(ŚV vākya 358-359a)
Hence, Kumārila claims that by seeing an indistinct white form, hearing a sound of neighing and hearing the sound of clasping hooves one gets at once the composite notion "a white horse is running". On the other hand, the Prābhākaras maintain that such notion is the end-result of a sum of inferences.
The Prābhākara opponent in the Tantrarahasya maintains that one

  1. 1. sees a white shape moving
  2. 2. hears hooves clasping
  3. 3. infers a substrate which is both running and a horse out of the sound of hooves
  4. 4. connects the inferred white-thing with the inferred horse and ascertains through cogent evidence (anyathānupapatti) ``A white horse".
  5. 5. connects the white horse with the running, which must have a substrate, and ascertains through anyathānupapatti ``A white horse is running".

The substrates of the sensible data, namely, of the colour white and of the sound of clasping are inferred because there is no colour without a substance, nor sound without a substance. The second substance is more precisely inferred due to one's experience of hooves' clasping.

The gist of the argument is clear: the notion that a white horse is running is the result of a second-order epistemological activity. However, the form of the argument seems unsound. At stage 3, one was already able to connect horse and running. Hence, why should one go back to the horse alone at stage 4, thus needing also stage 5? On the contrary, Kumārila's version had one infer the horse out of neighing and the going out of the sound of clasping. Rāmānujācārya has quoted Kumārila some lines above, hence why should he misrepresent the theory, in a way which, moreover, does not ultimately favour the Prābhākara view?

Can readers see better than I do?


Anonymous said...

I know nothing of the topic beyond this post, but this is how I would understand the described 5-step process:

1. [From senses] White shape

2. [From senses] Clashing hooves

3. [From 2] (a) Horse (b) is running

4. [From 1 and 3a] White horse

5. [From 4 and 3b] White horse is running.

Of course, step 3 is not the end, we need at least a step 4... the reason for having two steps after 3 is probably because there are two parts, 3a and 3b.
(Or it may be some distinction between nouns/adjectives/verbs — it becomes a sentence only when the verb "is running" is also incorporated?)

elisa freschi said...

Thank you anonymous reader,
I agree with you, but I cannot see why should one separate 3a from 3b.
The notion "aśvo dhavati" (a horse is running) is surely made of two parts, but once we have seized them as belonging to the same substrate (and this surely is the case at point 3, since we infer them out of a single piece of evidence, the sound of clasping hooves), why would we have to connect just the horse-substrate with whiteness? Should not we already know that the substrate of "horse" is the same as that of "running"?

Anonymous said...

Well, again, I've strayed here by mistake and should leave it for others to comment, but anyway, since it was interesting…

Firstly, is the question of whether the number of steps is 4 or 5 crucial to Rāmānujācārya's argument? If it isn't crucial, then it's perhaps best skipped. :-)

(Also, the step marked as 3 really seems to involve two separate inferences to me out of the same piece of evidence (that's why I wrote 3a and 3b).)

Now assuming it is... one guess is that, once you imagines that you are thinking in language, then there's the fact that adjectives describe nouns and not sentences. (I say this naively, without knowing the prior philosophers' views on the topic.) Thus, in

3. "A horse is running" (aśvo dhavati)

the "whiteness" can be attached only to the horse, not to the sentence — it is the horse that is white, not the running (or the fact of the horse's running) that is white. (In "śvetaḥ aśvo dhavati", it's the "aśvaḥ" that is "śvetaḥ".) Yes it's true that the same thing that is running is also white, but to get to this thought you have to first attach whiteness to the thing.

[This is perhaps a linguistic matter only: if the sentence instead was

3'. "A running horse [exists]" (dhāvat-aśvaḥ asti?)

then we can directly attach whiteness to the (compound) noun: step 4' would be "A white running horse [exists]." But then I don't know if they would insist that a step 5' is necessary to convert 4' to the form(?) "A white horse is running"!]

Of course, mentally we have no problem visualising a running horse and then adding the white colour to our image (though again it's debatable whether we're adding colour to the image or to the horse in it!). But perhaps relying on mental images instead of sentences is closer to the Bhāṭṭa view and not what a Prābhākara would say (or think like); I don't know. :p

Anyway, it's poor form for me to jump in like this without background; apologies for any silliness! I'm uncomfortable with this sort of thing so I'll stop now.

elisa freschi said...

Dear Shreevatsa,
why should not you comment? Personally, I am always glad to read other people's thoughts and a blog is not a professional lecture, where only well-considered arguments can be suggested (a kind way of saying that conferences are usually useless). In short, you are very much welcome to jump in other topics, too.

As for your point, here is R's passage about the Prābhākara rejoinder in my (horrible, I am afraid;-) translation:

[Siddhāntin:] «It is not so. In the case of this person, a composite (saṃsarga) apprehension (pratīti) does not arise. Instead, a white thing is seen. It can be originated by a horse [or] by some other moving [thing], so far alone [is known]. Then (atha), [the person] understands (avagam-) because of [his] expertise (pāṭava) [gained] through repeated exercise (abhyāsa) that the sound of clashing hooves [which he hears] is connected (sambadh-) with a horse . Then, by virtue of an inference only, the notion “a horse is running” occurs. Then (atha), [the person] ascertains that whiteness must have as substratum the inferred horse, and then also because of the conclusive evidence (pariśeṣa) “it cannot be anywhere else but in a horse” there is the notion “a white horse”. In the same way, when one infers that the inferred substrate of whiteness has a common substrate with the action of running, the notion “a white horse is running” arises. Else, it does not. That's all.

And you are right, the concept at point 3 is "aśvo dhavati". It might be that a "dhavadaśvo 'sti" would not have lead to the split of point 3. Yet, I can't understand why the Prābhākara does not precisely follow Kumārila example (which has been quoted by R. in the immediately preceding lines), where "horse" is conveyed by the sound of neighing and "running" by that of hooves. Would not you agree that it would have made better sense? What was the advantage of this misrepresentation of the Bhāṭṭa view?

Licenza Creative Commons
Quest' opera è distribuita con licenza Creative Commons Attribuzione - Non commerciale - Non opere derivate 2.5 Italia.