God, Necessity, and Truth

Jacques e-mails:

You think that if God exists, He exists necessarily, and if He does not exist, He does not exist necessarily.  But suppose that God does not exist.  We agree, I think, that we can't rationally rule out the possibility?  For instance, you've often argued that our evidence doesn't settle the question of theism versus atheism.  But then, supposing that God doesn't exist, and supposing that He might not exist in the actual world (for all we know), isn't it evident that regardless there are lots of truths?  For instance, even if God does not exist, it would still be true that He does not exist, or that He does not exist necessarily.  I'm not sure that you'd agree with this, but if you would, shouldn't you also agree that if God does not exist, there are some truths?

That is not quite what I said. I accept what I call Anselm's Insight: if God exists, then he exists necessarily; if he does not exist, then necessarily he does not exist.  What does not exist necessarily might be contingent; what necessarily does not exist is impossible. I know you understand the idea; it is just that your formulation suffers from scope ambiguity. Anselm's Insight, then, is that God is either necessary or impossible. He is necessarily non-contingent. (The non-contingent embraces both the necessary and the impossible.) In the patois of possible worlds, either he exists in every, or in no, world. If you wonder why I don't capitalize 'he,' it is because I hold that while piety belongs in religion, it does not belong in philosophy of religion.

Agreed, we cannot rationally rule out the possibility of God's nonexistence. I would say we cannot rationally rule it out or rule it in. "But then, supposing that God doesn't exist, and supposing that He might not exist in the actual world (for all we know), isn't it evident that regardless there are lots of truths? "

I would rewrite your sentence as follows:

It is epistemically possible that God not exist. Nevertheless, it is evident that there are truths.

I agree with the rewrite.  It is evident that there are truths, but for all we can claim to know, God does not exist. But this leaves open how God and truth are related.  Here are five different views:

1) There is truth, but there is no God.

2) There is truth, and there is God, but God is not the ontological ground of truth.

3) There is truth, there is God, and truth ultimately depends on the existence of God. There is truth because there is God.

4) There is no truth, because there is no God.

5) There is God, but no truth.

Ad (1). This I would guess is the view of  many. There are truths, and among these truths is the truth that God does not exist.  This, I take it, would be the standard atheist view.

Ad (2). This, I take it, would be the standard theist view among analytic philosophers.  Consider a philosopher who holds that God is a necessary being and also holds that it is necessarily the case that there are some truths, but would deny the truth of the subjunctive conditional, If, per impossibile, God were not to exist, then truths would not exist either. 

Ad (3). This is the view that I am inclined to accept.  Thus I would affirm the subjunctive conditional lately mentioned. The difference between (2) and (3) is subtle. On both sides it is held that both God and truths are necessary, but the Augustinian -- to give him a name -- holds that God is the ultimate  'source' of all truth and thus all intelligibility, or, if you prefer, the ultimate 'ground' of all truth and intelligibility.

Ad (4). This is Nietzsche's view.  

Ad (5). I have the impression that certain post-Nietzschean POMO-heads hold this. It is view not worth discussing.

I should think only the first three views have any merit.  

Each of the three has difficulties and none of the three can be proven.

I will mention quickly a problem for the admittedly plausible first view.  

Among the truths there are necessary truths such as the laws of logic. Now a truth is a true truth-bearer, a true proposition, say. Nothing can have a property unless it exists. (Call this principle Anti-Meinong). So no proposition can have the property of being true unless the proposition exists. A necessary truth is true in every metaphysically possible world. It follows that a necessarily true proposition exists in every possible world including worlds in which there are no finite minds.  But a proposition is a thought-accusative that cannot exists except for a mind.  If there is no God, every mind is contingent. A contradiction ensues: there is a world W such that, in W, there exists a thought-accusative that is not the thought-accusative of any mind.

Here are some ways an atheist might 'solve' the problem:

a) Deny that there are necessary truths.

b) Deny that truth is any sense a property of propositions.

c) Deny Anti-Meinong.

d) Deny that propositions are thought-accusatives; accept some sort of Platonism about propositions.

But each of these denials involves problems of its own which I would have no trouble unpacking.

 What say you, Jacques?

Related articles

Maverick Philosopher: God, Truth, Reality Denial: A Response to Some Questions

Maverick Philosopher: Lukáš Novák on Use and Mention

Comments

[Tony Hanson]

Hi Bill,

Hope all is well! Looking forward to Jacques' reply. I don't think you need minds for truths to exist, but I am more interested in another question. Assuming your theory of truth is true, it seems like the most you can prove is that there must be at least one mind in every possible world to apprehend necessary truths. Mind is necessary, but why a perfect mind (God)? Your Anselmian view is, if God exists, he must be a necessary being, but is it also your view that if a necessary being (mind) exists, it must be God? What is your argument for that? How do you derive all the other omni-attributes from merely a necessary being?
Or will you say with the cosmologist and teleologist that God is the most likely candidate?

Cheers!

[BV]

Hi Tony,

It's been a while. Good to hear from you!

Maybe I can address your challenge later when I have more time. But the challenge is a reasonable one: How might I show that it is the same mind in every world?

[Jacques]

Hi Bill,
You're right that my initial comment was a bit unclear. I'm going to think over some of these issues and get back to you! For now let me try to state the argument I had in mind more clearly:

(1) It is not epistemically possible that there are no truths.

(2) It is epistemically possible that God does not exist.

We agree on 1 and 2. But now assume for reductio that

(S) If God does not exist, there are no truths.

Then (3) it is epistemically possible that there are no truths (i.e., 1 is false).

Therefore, (4) it's not the case that if God does not exist, there are no truths (i.e., S is false).

This seems right to me. Do you disagree?

[BV]

Very interesting, Jacques.

You are assuming the following principle:

If it is epistemically possible that p, and p entails q, then it is epistemically possible that q.

I am not sure that this principle is true. Suppose little Johnny knows about addition but hasn't heard about subtraction. It is epistemically possible for Johnny that 2 + 3 = 5. (The proposition is not ruled out by what he knows/believes.) Now 2 + 3 = 5 entails that 5-3 = 2. But it is not epistemically possible for Johnny that that 5 - 3 = 2. This is because Johnny has no grasp of the latter proposition.

On the other hand, I am not sure I have provided a convincing counterexample.

And then there's the question as to the exact definition of 'epistemic possibility.'

[Jacques]

Hi Bill,
The principle is true if 'epistemically possible for S' means 'consistent with what S knows' but I also have doubts about that definition. (And I agree that propositions about subtraction aren't 'epistemically possible' for Johnny in some important sense of the term). I doubt that it defines what we're trying to characterize with phrases like 'For all we know...'. (One big problem is that in real life discussions of _that_ kind of thing we certainly don't have to assume that what is relevantly 'possible' constitutes any logically or metaphysically possible world--or even that consistency with the person's true beliefs is necessary, rather than just consistency or seeming consistency with any beliefs that the person takes seriously...)

On the other hand, as you seem to be hinting, my argument would probably work on some other definitions of 'epistemically possible'.

Or, even without a definition, we might agree on something like this:

- p is epistemically possible for S if S reasonably believes that p is consistent with propositions that S reasonably believes (or reasonably believes himself to reasonably believe).

- p is epistemically impossible for S if S reasonably believes that p is inconsistent with the propositions that S reasonably believes (or reasonably believes himself to reasonably believe).

The same kind of argument works given these assumptions.

Roughly, I could put it this way:

"For all I know, God does not exist.

For all I know, if God does not exist there are no truths.

Therefore, for all I know, there are no truths.

But I _do_ know that there are truths.

Therefore, either it's false that (a) for all I know, God does not exist, or that (b) for all I know, if God does not exist there are no truths.

But (a) isn't false.

Therefore, (b) is false."

We reasonably believe that 'God does not exist' is consistent with what we reasonably believe. We reasonably believe that 'There is no truth' is not. But then, if we grant 'If God does not exist, there are no truths' and 'Possibly, God does not exist' we can reasonably believe that 'There are no truths' is consistent with what we reasonably believe. (It would be reasonable: if I reasonably believe that p is consistent with q, and p obviously implies r, the reasonable conclusion is that r is consistent with q.) But we can't reasonably believe that 'There are no truths' is consistent with what we reasonably believe.

That's how I'd defend the reasoning, I think. But now I'm thinking this is the wrong way to approach the issue. Or the present notion of 'epistemic possibility' makes it wrong, maybe? Thanks for an interesting discussion! Let me know if you have more ideas...

[BV]

I'm stumped for now, Jacques. I'll think about it some more tomorrow.

[Markus]

Dear Bill, (I love your blog and read it regularly)

Call it cosmic synergy, I grappled with this exact same issue yesterday whether truth can exist without a mind.

I was thinking about mind, states of intentionality, semiotics, reason/logic, and truth, which from a naturalist perspective, would only have come to exist in the last 200 000 years or so in human minds. (I was thinking that atheists in particular rely on something, e.g. reason, that would only have arrived on the scene very late, because how can reason (with mind, states of intentionality, semiotics, reason/logic, and truth) exist when only stars were around?)

And if truth is discovered (and not invented), then truth (along with all the rest I mentioned above) must have existed before Homo sapiens were around. So how can truth exist apart from mind(s), except as Platonic forms? I concluded it cannot, only to read here that it can!

I must confess, I don't understand what truth can possibly mean without a mind to create or appropriate it.

[BV]

J. writes, >>The principle is true if 'epistemically possible for S' means 'consistent with what S knows' but I also have doubts about that definition.<<

Here is a counterexample. It is necessarily false that (P) no set has a proper subset of the same cardinality. (The evens are a proper subset of the naturals even though both sets have the same cardinality, aleph-zero.)

Now P is epistemically possible for Tom, whose math education is limited. It is possible for all he knows. But P is not logically consistent with what Tom knows. Propositions q, r are consistent just in case they can (logically) both be true. P, however, being necessarily false, is not consistent with any proposition.

So given the above definition of 'epistemically possible,' P is NOT epistemically possible for Tom, which contradicts the datum that it is.

So what is epistemic possibility?

[BV]

Hi Markus,

Thanks for reading.

Surely there were truths before man evolved. It was true that dinosaurs roamed the earth, that the speed of light was 186, 282 mi/sec, that Planck's constant was such and such, that Earth had exactly one natural satellite, that the Moon had craters and was not a perfect sphere . . . .

And then there are necessary truths such as LNC.

Truth is not a physical property or physical thing. The Moon is not true. A particular crater is not true. The Big Bang is not true. But it is true that the Moon has craters.

Now we have a structure that can't be physical even if it it contains physical things. We have a proposition or state of affairs.

This points us beyond the physical . . . .

[BV]

If one were a strict materialist who held that the material exhausts the real, there would be no room in this scheme for the truth that this is so.

This is why truth is such a magnificent and ennobling object of contemplation. It pulls you right out of Plato's Cave. But whither? To the mind of God or to mere 'abstract' objects?

[Christian]

Dear Bill,

Like Markus I also read your blog regularly and learn a lot from it. Thank you.

This may be a tad simplistic, but I may have an objection to something you said in the comment on Wednesday at 4:52am. Is it really true that a necessarily false proposition is inconsistent with any proposition? Wouldn't any necessarily false proposition be consistent with the proposition that necessarily false propositions contain at lest one error? Wouldn't necessarily false propositions also be consistent with the proposition that any argument that contains a necessarily false proposition is a flawed argument?

Also, how would you evaluate the following definition of epistemic possibility: P is epistemically possible for S if P is not inconsistent with what S knows to be true.

[BV]

Thanks for the kind words, Christian.

>>Wouldn't any necessarily false proposition be consistent with the proposition that necessarily false propositions contain at lest one error?<<

No, because a nec. false proposition cannot be true whereas two propositions are consistent only if both can be true. The same holds for your second example.

>>Also, how would you evaluate the following definition of epistemic possibility: P is epistemically possible for S if P is not inconsistent with what S knows to be true.<<

That is equivalent to the definition Jacques gave that succumbs to the cardinality counterexample.

[Jacques]

Typical discussions of epistemic possibility have always bothered me. The approach of analytic philosophers is to come up with some precise, rigorous definition in terms of concepts they like--consistency, knowledge--which just obviously doesn't fit the ordinary language uses they claim to be 'analyzing'. Or else they offer a mix of that kind of thing plus some hopelessly vague gesturing--e.g., 'consistent with what S knows and S has some relevant practicable method for determining whether p...' (And then they pretend the hand waving is meaningful.)

Maybe these typical ever-more-complicated 'analyses' are definitions of something, some possible concept. But why should we care about it? What can we do with it? Has anyone ever had any need to talk about what is 'possible' in this sense (and, if we wanted to, could we ever know what we were talking about or whether we were saying anything true)? As in so many other areas, it's as if they really just want to have these intellectual toys they can fully control, even if the game has no bearing on reality; what matters is 'clarity', 'rigor'... i.e., power, mastery, pretending to understand everything... Okay that's a bit unfair of course, but that seems to be the psychology. Otherwise it's really hard to understand what they think they're doing, how they can imagine that their toys are real.

When people say 'Trump might win' or 'The keys might be on the desk' they just obviously are not saying something that's true dependent on what they know (so what is actually true) or what is consistent with what they know (what is actually consistent). For reasons we've touched on already.

I think when we say 'Maybe p' or 'p could be true for all I know' we're saying something that's true or appropriate to say if and only if it _seems_ to us that we have no very good reason for denying that p. And that's just really vague, not much of an analysis, but it's all we really mean. Whatever goes on when it seems to me that I have no strong reason for denying that p, and whatever it may be for a reason to be strong (or seemingly strong). Do we need to have a more precise, rigorous (pseudo-scientific) account of epistemic possibility? I guess it might be nice :) But it's just weird to expect or demand that thinking and communicating can always be reduced to a computer program.

I mean, it would be perfectly appropriate or reasonable in the right circumstances to say things like this:

- 'Trump might be a Russian spy' [Trump just isn't a spy, so this can't be consistent with anything the speaker knows... The person is badly misinformed, kind of paranoid, believes crap he sees on TV...]

- 'There might be a set of all non-self-member sets' [The proposition he thinks might be true is necessarily false, inconsistent with anything he might know, etc... The speaker knows a little about sets but hasn't been introduced to Russell's paradox...]

I'd say that in the right circumstances these claims are true: the propositions in question really are 'possible' in the relevant sense, for that person or from his perspective. Is it not enough just to say something vague like 'p is epistemically possible for S if and only if S (reasonably?) takes himself to have no good reasons for denying p'?

Or consider another ordinary use of 'possible'. It's possible for me to tie my shoes, i.e, that's something I can do or could do; I'm able to do it; it's in my power. There's not a lot we can say to 'analyze' this kind of claim. And we shouldn't expect that an analysis, if there could be one, would provide truth conditions sort of like the truth conditions for 'It is logically possible that I tie my shoes' or 'It is legally possible for me to tie my shoes'...

Don't we already know well enough what we mean in these cases? Why should we think there's some other set of more 'basic' concepts (of logic or set theory or whatever) that would make the meaning or truth conditions more clear?

[BV]

Good comments, Jacques.

>>But why should we care about it? What can we do with it? Has anyone ever had any need to talk about what is 'possible' in this sense . . . <<

But wouldn't you agree that to understand metaphysical modality we need to be able to distinguish it from epistemic/doxastic modality? Try this with your wife: Say to her, 'I might not have been born.' You might get the response: So you are not sure you were born? You have some doubt about it?? "No, sweetie, that's not what I'm getting at. You are confusing epistemic with metaphysical possibility . . . ."

In Kripke we find this: if it is possible that I am not identical to my body, then I am not identical to my body. Now it is possible, ergo, etc. Some have tried to undermine this by saying that the possibility is merely epistemic. So it seems we do need to worry about epistemic possibility.

>>When people say 'Trump might win' or 'The keys might be on the desk' they just obviously are not saying something that's true dependent on what they know (so what is actually true) or what is consistent with what they know (what is actually consistent).<<

Not clear. Suppose you can't find your keys. You ask your wife, "Where are my keys?" She says, "They might be on the desk." Suppose unbeknownst to both of you that someone stole your keys and melted them down. It would then be really impossible for the keys to be on your desk. Still, your wife expressed some proposition. Which?

Something like this: we are both ignorant as to the location of the keys, but this ignorance permits the reasonable supposition that there are on your desk perhaps under a pile of student papers.

Your ignorance, not your knowledge, is the truth-maker of the proposition your wife expresses.

Of course, it could be that there is no one analysis that fits all the varied epistemic uses of the modal auxiliaries 'might,' 'could,' 'must.'

'The keys have to be somewhere!' Their being somewhere is entailed by what we normally assume about lost articles, including the assumption that they have not been annihilated.

More later perhaps.