Tuesday, March 1, 2011

Is the existence of God inferable? Epistemic Liberalism

Consider the following argument:

the earth and the other things [of the world] is a product, hence it has a doer, like a pot.
Can we establish the existence of a doer out of the fact that the mountains, etc. are a product? Of course, a first problem could be how could we be sure that those things are a product. But a second one regards the legitimacy of an inference based on a concomitance we could never grasp (that between an embodied doer and its products).

Over the week-end I read an interesting article by Matthew Dasti, Indian Rational Theology: Proof, Justification and Epistemic Liberality in Nyāya's Argument for God. The article focuses on the proof for the existence of God I mentioned above, which is the standard one in Nyāya texts.
Dasti considers the argument as dealt with by various Naiyāyikas, by its Buddhist opposers (mainly Ratnakīrti), its Western parallels and its significance within the Naiyāyika epistemology. As already pointed out in Parimal Patil's study of Ratnakīrti's argument against God's existence, in fact, the inference about the existence of God is deeply linked with the general problem of the reliability of inference as instrument of knowledge and its ampliative character.
Because of the problems hinted at above, the product-producer inference can hardly apply to God, unless one stresses the ampliative character of inference and admits an inference regarding a producer one has never experienced to be actually concomitant with products whose arousal one has never seen. Buddhists are quite hard in denying the validity of an inference which is not safely based on a sure invariable concomitance. Naiyāyikas, on the contrary, are –in Dasti's opinion– more "liberal".


Jayarava said...

The idea that every process has an agent 'a doer' (kartṛ?) is a logical fallacy. I think it emerges out of a linguistic rather than an ontological observation - and a confusion about which category is which. A verb always has an agent in Indo-European languages. Hence we say 'it is raining' but what is 'it'? It's only a place holder. The same happens in Sanskrit (tad varṣati). Benjamin Lee Whorf pointed out that some Amerindian languages don't do language this way. They don't imply an agent for every process - they might just say "there is rain". This aspect of Indo-European gives us propensity to see the impersonal as personal (or I suppose one might argue that it works the other way around; that because we personalise everything we expect every process to have a doer).

Perhaps the Vedics would argue that Indra makes it rain, or - as in Bṛhadāranyaka Upaniṣad 1.1: yad mehati tad varṣati - that it is the cosmic sacrificial horse urinating that makes rain. The argument is circular: we believe that Indra causes the rain; it is raining; therefore we believe in Indra. On what fact or observation does the belief in Indra rest? To some extent it all rests on this idea that for a verb to be used in language, there must be an agent doing the verb. Whereas this need not be the case for nature. I note the propensity of Vedic grammarians to see everything in terms of verbs, and this can't have helped.

An inference is not proof of anything unless one can demonstrate that no other inference is a possible answer. However there is always the unanticipated possibility - so no inference is ever proof of anything. But Karl Popper was a long time after the period under discussion.

The early Buddhist critique of the belief in the existence of Brahmā was interesting. Since the goal of the brāhmana was a permanent, post-mortem, merger with brahmā (brahmasahavyatā), there was no way any brahmin could honestly claim first person knowledge of brahmā (brahmā sakkhidiṭṭho) since they were still alive.

elisa freschi said...

Dear Jayarava,
you raise many interesting points. The first one (the linguistic and not ontological nature of some seemingly philosophical problems) has been dealt with by Johannes Bronkhorst. Bronkhorst maintains that the problem of causality is, in Indian thought, mainly a linguistic problem. Indian thinkers wondered about the coming into being of items due to sentences like "The potter makes the pot", where the pot is in the accusative, although, in fact, it still does not exist.
However, I do not think that this is the case with the product-producer inference. I'll dedicate a separate post to a detailed answer.

Matthew Dasti said...

Greetings. Elisa sent me a link to her posts on my paper. A quick note on this, if I may. I will spend a little more time responding to Elisa's comments in the second post on my paper. I think that both of you have important points to make, but Jayarava, if you don't mind, I'd take issue with a couple of things you asserted. First, the fact that ontology may follow ordinary language is not enough to refute said ontology. I need not remind you of the dangers of the genetic fallacy in this application. Also, I don't think that the principle 'An inference is not proof of anything unless one can demonstrate that no other inference is a possible answer.' is sound. Rather, an inference is not proof if there is another inference put forth (or at least in, which one should be aware of) with roughly equal or greater epistemic support (or other countervailing considerations). The principle of inference to the best explanation is not inference to the only explanation, after all. Moreover, it is unlikely that any proof would stand given your principle, since it requires that we have the ability to discover and refute all other potential inferences before our original inference is allowed to stand. My feeling is that a better way to attack Nyaya's causal principle is to illustrate how much it depends on the details of Nyaya's atomism. Given that dependence, it is unlikely that others, who don't hold such an atomism, would be persuaded by it. Anyway, just a few thoughts. Thanks for considering the paper.

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