Given Anselm's Insight, How is Empirical Evidence Relevant to the Existence/Nonexistence of God?

ANSELM'S INSIGHT

I take what I call Anselm's Insight to be non-negotiable.  St. Anselm appreciated, presumably for the first time in the history of thought, that a divine being, one worthy of worship, must be non-contingent.  If your god is contingent, then your god is not God. For if your god is contingent, and he exists, then his nonexistence is possible, and nothing like that could count as God. God, by definition, is that than which no greater can be conceived, to use use Anselm's signature phrase, and necessity is a greater or higher modal status than contingency. Contingency in an existing thing is an ontological defect. God is an absolute, and no absolute worth its salt is contingent.  If, on the other hand, your god is contingent and he doesn't exist, then it is worse still: your god is impossible. 

Anselm's point, expressed with the help of Leibniz, is that the concept of God is the concept of a being that either exists in every metaphysically possible world, or in no metaphysically possible world.  This formulation has the advantage of not presupposing the existence of God by the use of 'God.' For it may be that the concept is not instantiated, as it would not be if God is impossible.

Anselm's Insight rests on the assumption that the necessarily existent is 'better' than the contingently existent, which in turn rests on the Platonic sense that the immutable is higher in value than the mutable. This assumption can be reasonably questioned but also reasonably defended. But I won't go into that now. For present purposes I endorse both the Insight and the Assumption upon which it rests.

ANSELM'S ARGUMENT

But it does not follow from Anselm's Insight that Anselm's Argument is probative. I am referring to the modal ontological argument found in Proslogion III. One cannot validly infer from God's non-contingent status alone that he exists.  For it is epistemically possible -- possible for all we know -- that God is modally impossible. 'Non-contingent' does not mean 'necessary.' It means 'either necessary or impossible.'  To arrive at God's existence, one needs a possibility premise to the effect that God  is possible.  But how would you know that the premise is true? Conceivability is no sure guide to real possibility. 

For more on the Argument see here. Today's topic is different.

IS EMPIRICAL EVIDENCE RELEVANT TO THE EXISTENCE/NON-EXISTENCE OF A NON-CONTINGENT GOD?

A question posed to me by a reader requires only Anselm's Insight although he thinks it requires the probativity of Anselm's Argument. I take him to be asking how empirical evidence could be relevant to the existence of a non-contingent God.  

Suppose God does not exist.  Then, by the Insight, he is impossible.  But no amount of empirical evidence could show that what is impossible is actual.  So no miraculous event could show that God exists. If God is impossible, any putative empirical evidence to the contrary could be legitimately dismissed a priori.  Compare round squares and colored sounds. They are impossible. The claims of anyone who said he had empirical evidence to the contrary would be legitimately dismissed a priori.  We know that there  cannot be any round squares or colored sounds. We do not know that God is impossible. But we do know (given Anselm's Insight) that either God necessarily exists or God is impossible.  

If, on the other hand, God does exist, then he cannot fail to exist and no empirical evidence is needed or could be relevant.  Suppose that someone demanded empirical evidence of the truth of 7 is prime or the existence of 7 or the existence of the proposition, 7 is prime.   You would show that person the door.

Now either God exists or he does not. Either way, empirical evidence is irrelevant to the existence of God.

What then of natural and moral evil? Are they evidence of the nonexistence of Anselm's God?  No, and to think otherwise is to assume that God is a contingent being.

The above is my interpretation of my reader's reasoning.  I am inclined to accept it as I interpret it.

Comments

[Bradley Robert Schneider]

Thanks for this post, Bill, responding to my inquiry. As I've thought about this more, I wonder if a qualification is in order: assuming Anselm's Insight is correct (and I think we agree it is), empirical evidence might be relevant insofar as it has a bearing on whether God's existence is possible. For example, if you find the philosophical arguments on the possibility of a necessary being to be inconclusive, it arguably makes sense to consider whether empirical evidence can "tip the scale" in favor of one conclusion or another. But what doesn't make sense is to (1) accept Anselm's Insight and (2) grant the possibility of God's existence while denying that God does in fact existence based on one's weighing of the empirical evidence (e.g., Russell's famous comment that, if questioned by God about his non-belief in the afterlife, he would claim as an excuse "not enough" evidence).

[BV]

>>empirical evidence might be relevant insofar as it has a bearing on whether God's existence is possible. For example, if you find the philosophical arguments on the possibility of a necessary being to be inconclusive, it arguably makes sense to consider whether empirical evidence can "tip the scale" in favor of one conclusion or another.<<

I don't understand the above. If a being is necessary, then of course it is possible. But you can't assume that God is a necessary being. For if you assume that, then you assume that God is actual. What the Insight states is not that God is a necessary being, but that God is a non-contingent being: one that is either necessary or impossible.

Could there be empirical evidence of the possibility of a thing? Yes, if there is empirical evidence of its actuality. For whatever is actual is possible. But what we need in the case of God is evidence of the possibility of a being that we do not know exists. What could be evidence of that?

[Bradley Robert Schneider]

I now see that I was using language too loosely. By "necessary," I meant "non-contingent." With that clarified, my question remains: If I agree that God is either necessary or impossible, how am I to decide which is the case? Suppose I am not convinced either way by non-empirical arguments as to God's possibility. Can't I take into account miracles or religious experiences in assessing whether God is actual?

[BV]

If we know that God is possible, then we can prove that he is actual. What I don't understand is how we can come to know that God is possible by empirical means. For if God does not exist, then he is impossible, and any putative evidence adduced will be unavailing.

[Lukáš Novák]

Bill,

I think you are confusing logical and epistemic modalities. The fact that God is non-contingent is a fact about the logical modality of its existence. Evidence, however, is something that concerns epistemic modality, and the two do not go hand in hand. We can be 100% epistemically certain of a contingent truth, therefore even contingent truths can provide 100% epistemical support to a necessary truth. A very simple example: the contingent, empirical fact that I exist provides 100% epistemical support to the conclusion that a man is a possible being - which is a logically necessary truth. Likewise, the empirical fact that there exists a thing caused can, ultimately, provide 100% epistemical support to the conclusion that God exists. The tempting maxim that a necessary truth cannot be derived from (or receive epistemic support from) contingent premises must be discarded.

Conversely, something can be impossible logically without being impossible epistemically. For example, the negation of Fermat's Last Theorem has always been impossible logically, but it used not to be impossible epistemically. And, so long as something is not epistemically impossible (i.e., of zero/minimal epistemical probability), its epistemic probability can be increased by appropriate evidence. For example, the fact that so many mathematicians tried in vain to find the proof of FLT constituted, back then, certain non-negligible (empirical!) evidence increasing the (epistemic) chances that it might actually be false. Even given that we have always known that logically, it is either necessary or impossible, its epistemic probability was floating somewhere in between and was responsive to evidence, even empirical one.

[BV]

Lukas,

It is always a pleasure to hear from you. I hope you and your family are well.

I grant you that 'I exist' assertively uttered by me expresses a truth that is both logically contingent and epistemically certain to me. But it does not follow that, possibly, a man exists. What follows is that, possibly, an ego (an I, a self, a res cogitans) exists.

More later.

[BV]

Lukas writes, >>Likewise, the empirical fact that there exists a thing caused can, ultimately, provide 100% epistemical support to the conclusion that God exists. The tempting maxim that a necessary truth cannot be derived from (or receive epistemic support from) contingent premises must be discarded.<<

Suppose that the physical universe U exists contingently. Suppose further that there is a probative cosmological argument *a contingentia mundi* that proves that there exists an entity G, distinct from U and all of its parts, such that G causes U to exist. It would then be the case that, necessarily, if U exists, then G exists. But that is distinct from saying that if U exists, then necessarily G exists. The necessity of the CONSEQUENCE is not to be confused with the necessity of the CONSEQUENT. So I ask: how do you know that necessarily G exists?

[Lukáš Novák]

Dear Bill,

thanks, we are fine! To your rejoinders:

(1) I deny that "I exist" epistemically supports "a res cogitans exists". I even deny that "I exist" logically implies "a res cogitans exists" -- for I think that "I exist" is true and "a res cogitans exists" false. I find "men exist" as empirically evident as anything can be; if anything at all ever is empirically evident, then "men exist" is. On the other hand, "res cogitantes exist" is a highly contentional philosophical thesis -- most philosophers of all times would reject it -- and so cannot reasonably be claimed to be evident, either immediately or in virtue of a simple immediate inference like "cogito ergo sum".

(2) But all that is tangential to our present discussion. Perhaps I picked a wrong example; my point is that it is in general possible to derive necessary truths asserting metaphysical possibility from contingent truths asserting existence, which alone suffices to refute the adage that necessary metaphysical truths cannot be derived from contingent premises.

(3) Here comes your distinction. You are right that in case of a necessarily true proposition, one thing is to prove its mere truth and quite another is to prove its necessity. I take it we agree at least so much that you can prove the mere truth of a proposition that happens to be necessary, starting from contingent premises. This much is sufficient to reject the claim that empirical facts are epistemically irelevant with regard to propositions that happen to be necessary. Perhaps you can only infer the mere truth of "God exists" from empirical premises; therefore, irrespective of whether "God exists" happens to be necessary or not, empirical evidence can be relevant to show that.

(4) The Anselmian consideration only shows a fact about the logical or metaphysical modality of God's existence, not about its epistemical modality. As I said, logical non-contingecy does not imply epistemical non-contingency (i.e., either certain truth or certain falsity), and only epistemical non-contingecy implies epistemical non-responsivity to evidence.

(5) However, the Anselmian consideration also shows that there are certain propositions whose truth implies necessity, and that "God exists" is among them. This is the answer to your last question: I know that God necessarily exists because I know that God exists (on empirical grounds) and I also know (on conceptual grounds) that God can only exist necessarily. No necessity of consequence/necessity of consequent confusion here.

(6) Examples of similar arguments: I know, on empirical grounds, that dogs exist, from which I can infer that dogs are S5-possible. Furthermore, I know, on conceptual grounds, that all S5-modal claims are non-contingent; consequently, I can conclude that it is a necessary truth that dogs are possible. Another example: I measure with certain precision that the sum of the angles of this triangle is certainly within the interval [178°; 182°]. This is an empirical observation. I also know that geometrical truths are necessary. Therefore, I can conclude that the sum of the angles of this particular triangle necessarily falls between 178° and 180°.

(7) A historical aside: I doubt that the divine attribute Anselm employs in Proslogion III is God's necessary existence. His words are "quod non possit cogitari non esse", i.e. "that which cannot be thought not to exist". That this "unthinkability of non-existence" is to be taken literally -- that it is impossible for anyone to seriously form the thought that God does not exists -- is evident from the next chapter, where Anselm poses himself the obvious question: how it comes that the fool said in his heart "there is no God". And his reply is that he only proferred the words in his heart, without sufficiently understanding their meaning. (Note that by later scholastics (such as Aquinasú Anselm was regarded as a representative of the thesis that God's existence is per se nota, i.e. impossible to be denied and therefore not in need of a proof; rather than merely as an author of a peculiar proof of God's necessary eixstence.) The "Leibnizian" reinterpretation goes back to Scotus (well, it is more complicated than that, but never mind).

[BV]

Lukas,

I don't agree with your (1), but I won't say any more about it.

As for (2), would this be an example of your reasoning:

a) Tom exists; so
b) Possibly, Tom exists.
c) Possibly p --> Necessarily possibly p. (characteristic S5 axiom) So:
d) Necessarily, possibly Tom exists.

This would seem to show that one can validly infer a necessary truth from premises not all of which are necessary.

But does the possibility of Tom's existence, which exists in the actual world, exist in every possible world? I don't think so. In a world in which Tom does not exist, there is no de re possibility of HIS existence. And at times in the actual world before Tom began to exist, there was no de re possibility of his existence. If this is right, then the above argument doesn't succeed.

[Lukáš Novák]

Dear Bill,

yes,this is an example. Three points:

(1) I disagree, that de re necessities do not "exist" in the worlds where the res does not exist. If you assume trans-world identity of individuals -- which I think is the only option if the "possible-worlds apparatus" is to capture what we mean by "possibility" and "necessity" --, then a possibile in a world A is identical to an actually existing thing in another word B; and its merely-possible-existence located in A is nothing else than its actual possibility-not-to-exist located in B. So non-necessity of an individual existing in one world requires, in my opinion, its individual mere-possibility in some other world: it is THIS individual which can not-exist, and this individual possibility is just its actual-non-existence-but-possible-existence in another world. (And the possible-existence in that world of that individual just is its actual existence in another world). Or very simply: The possibility of x's existence in a world where it does not exist is its existence in a world where it does exist.

It seems to me that your proposal amounts to saying that possibilities are merely generic and not individualized. But since actualities are individualized, and every actuality is an actualization of a possibility (as impossibilities cannot be actualized), there must, trivially, be individualized possibilities, whatever their metaphysical interpretation might be.

(2) Even if it were in some sense correct that the possibility of Tom's existence only "exists" in those worlds where Tom exists, it would have to be read as a metaphysical interpretation of what "Necessarily, possibly Tom exists" means, not as a denial of this proposition. The fact that possibilities are necessary is part of what S5-possibility means. And S5-possibility is the weakest of all possibilities, therefore any other kind of possibility presupposes S5-possibility. So no way out :-).

(3) De re modality is not essential to the example, a de dicto one works as well:

a) Tom exists. [empirical]
b) Tom is F. [empirical]
c) Possibly some F exists. [a,b]
d) Possibly p --> Necessarily possibly p. [characteristic S5 axiom]
e) Necessarily, possibly some F exists. [c,d]

[BV]

Lukas writes, and this seems to be the heart of the matter: >>I know that God necessarily exists because I know that God exists (on empirical grounds) and I also know (on conceptual grounds) that God can only exist necessarily.<<

At best, a cosmological argument takes us from the contingent universe U to a divine creative act that explains the existence of U. Now this creative act is itself contingent: God might not have created anything. If God is simple, then he is identical to the creative act. Since the act is contingent, God is contingent, and therefore not God by the Anselmian criterion. On the other hand, if God is necessary, then the creative act and U are necessary, which is unacceptable. The following cannot all be true:

1. God is simple
2. God is noncontingent
3. God's creative act is contingent

[Lukáš Novák]

Dear Bill,

God's creative act need not be contingent. It only needs to contingently bring about its effect.

God's efficiency is distinct from created efficiency. A created cause is itself changed by causing (by eliciting de novo the productive act as its accidental form), God is not changed by causing (being for eternity identical to any of its timeless creative acts). God would be the same in all respects had He not caused the world into existence. This is the requirement of His perfection.

Lukas