The theological usefulness of arguments for the existence of God

It would be nice if there were any reasonable atheists and agnostics that found them persuasive, no? If you set the bar too low, then this offering is successful in your sense:
(1) Either 7+1 = 834 or God exists.
(2) It's not the case that 7+1 = 834.
(C) God exists.

I'm reasonably certain that as an empirical fact about the meaning of "successful", the assertion "That is a successful argument" is false.

Clayton,

This is a good question. I'm not entirely sure that Alex is setting the bar to low, or low at all. Why think that a successful argument must convince a reasonable atheist? I'm not sure that philosophical arguments succeed or fail only in this way. Couldn't you find an argument illuminating or disabusing even if you are not convinced by it? Don't you find Zombie Arguments illuminating and interesting, though unconvincing? Don't you find most of the arguments in Plato unconvincing, though illuminating and important? When was the last time anyone was convinced by the Cogito? But that's a pretty important argument, wouldn't you say?

Fergus Kerr published a book on Aquinas a few years ago in which he asked the question why Thomas would spend effort on proving the existence of God, when everybody in his world already believed in God. One speculative answer, as I recall (don't have the book available) was that the priests and friars Aquinas hung around were almost _too_ familiar with God, and the cosmological arguments re-introduce a cosmic dimension that goes missing in private devotion. Not being an Aquinas scholar, I can't comment on the merits, but I found it an interesting suggestion.

Clayton,

Does your sample argument "fulfill the theological role of showing the way God is intimately involved with everything"? I didn't read Pruss as saying that truth of premises + validity is sufficient for serving the role for arguments for God's existence that he had in mind, only as necessary.

The previous is a touch too elliptical. More clearly: Pruss is suggesting that these arguments can be explanatory, which requires soundness, though soundness is not sufficient. Your sample argument plainly fails by that criterion.

Alex
You write, "If Christian theism is right, God is everywhere, creation is a self-revelation of his (St Ephraem of Syria talks of the three books of Scripture: the Old Testament, the New Testament and Creation), and he is behind everything that is good. Arguments for the existence of God typically attempt to exhibit a way in which some aspect of reality depends on God, and thus they can increase our theological understanding of God's involvement with creation. This is true of just about every major type of theistic arguments."

Consider this:

When discussing God theists often start by defining God as a perfect being having the characteristics of omniscience, omnipotence, omnipresence, and complete goodness. They also maintain that God is the creator and sustainer of the universe, cares about his creation, and can have a personal relationship with what he created. The problem with this approach is that it is a non-starter as far as us having any knowledge concerning God. Why do we think that such a being exists? What is the evidence?

If it is simply the definition of God that enables us to prove his existence then it begs the question; God exists because he is perfect and he is perfect because he exists. But, as is well known, the fact that we can define something is not a demonstration that what is defined exists. We can define a square as “a parallelogram having four equal sides and four right angles,” (Webster’s Dictionary) but this does not demonstrate that any squares exist. This is, of course, the difference between what Bertrand Russell referred to as ‘knowledge by description’ and ‘knowledge by acquaintance.’ Hemingway’s old fisherman dreaming about lions on the beach does not demonstrate that there are lions on the beach. Being able to describe x does not mean we are acquainted with x.

Consequently arguing from some definition of God will not result is us knowing ‘by acquaintance’ that God actually exists. How about the idea that he created the universe? Let us grant that he did in fact create the universe. What does this prove? Does it prove, or even suggest, that he has all, or any, of the characteristics ascribed to him by theists? No; it simply means that he had the ability to create the universe. What does it mean to say that A can do x? Again, if we rely on the dictionary (Webster’s in this case) we find that ‘A can do x’ means that A is able to do x, has the knowledge to do x, and has the right to do x. Certainly as less then perfect being can satisfy these requirements (after all the universe exists and we granted that he created it) so it is conceivable, based on the evidence, that an imperfect being, a lesser God if you will, created the universe.

If we accept the ‘principle of parsimony’ then we ought to utilize the theory that best squares with the facts but that commits us to the least amount of, or least complicated, ontological entities. In this case, if we grant that the universe was created, this principle seems to warrant us in accepting the notion that a lesser God created the universe. An important consequence of accepting the idea of a lesser God is that many of the traditional problems associated with theism dissolve. No longer do we need to worry about the problem of evil (in its traditional form), or the problem of God creating as less perfect world then he could have created, or if this is the best of all possible worlds, or if there is a best possible world, or if God is free. A lesser God can do what it chooses to do based on his ability, knowledge, and right to do whatever he wants to do. We can conclude then that God created this world because it is the world he wanted to create. We may never know why (the lesser God may no longer exist), but this is not a question that needs to be answered if we focus only on what God can do, which is to create what he has the ability, knowledge, and right to create. He may even care about (if he still exists), and want, a personal relationship with his creation.

So a couple of questions; do we actually give up anything of importance by believing that we are children of a lesser God, and can a person be a Christian without being a theist?

You also write, "What is particularly interesting here is that the arguments do not need to be successful in the full sense of being such that they would convince every reasonable person in order to fulfill the theological role of showing the way God is intimately involved with everything. Of course the arguments need to be sound: the premises must be true and the conclusion must follow from them. But to fulfill this theological role they need not convince the atheists or agnostic. It suffices that the theist see that they are sound and why they are sound."

I think that arguments where the premises are true should convince everyone that the conclusion is also true if it follows from the premises. If not, we simply end up 'preaching to the choir' and while that may be alright for ministers, priests, rabbis, etc., it should not be enough for philosophers.

If it is simply the definition of God that enables us to prove his existence then it begs the question. . .

What is wrong with arguments that include necessary existence as a perfection (hence possessed by a perfect being)? One version of the Ontological Argument does that. But the Argument also includes the premise that God is a possible being (Anselm would say for "it's possible" that "it exists in the understanding"), but Leibniz and the rest up to Godel, Malcolm, Hartshorne and Plantinga make more explicit the premise that God possibly exists. There is nothing remotely question-begging about the argument.

John,

Just because an argument does not prove everything, this does not mean that it proves nothing. Maybe theists have been guilty of extrapolating too much from their argument's conclusion; however, whatever their argument proves it proves. Like you say, if an argument can prove that there is, at the least, a personal being who is powerful enough to create the universe, wise enough to design its parts, sovereign enough to will it into existence, this might not prove God is all-powerful, all-good, all-knowing but surely an argument that could do this would be immensely valuable and make a significant difference in our world.

As far as your desire that sound and valid arguments ought to convince everyone of their conclusion, I agree. However, ought definitely doesn't imply will and that is simply the world we are stuck with. We are probably better off for it too. Our ability to believe otherwise, even when we are wrong, is good. It may even help us understand and discover truths we might not have searched for otherwise, which may allow us to find those arguments that prove the remainder of God's valuable attributes.

"It suffices that the theist sees they are sound..."
Forgive me if this is an ignorant question, but I'm a non-philosopher who enjoys reading this blog. Do you, Dr.Pruss, or other Prosblogions make a distinction yourselves between being philosophers of religion and being a Christian apologist? Is blogging more of an outlet for the latter -- or is this an unhelpful distinction?
Best regards, - Steve Esser

Steve, what is your point?

Hey Mike,

I think zombie arguments are interesting, but I don't find them to be successful for any of the argumentative purposes intended by those who offer them.

Alex was saying that there's a sense of "successful" according to which an argument might be successful even if they didn't convince every agnostic or atheist. That seems right. Surely an argument that convinced 50% of the atheists out there would be pretty darn successful. So that much surely is right. However, the assertion that they might be successful arguments even if no reasonable atheist or agnostic could be moved by them seems too low a bar. However, by suggesting that some sound arguments that only a theist could reasonably regard as sound could be successful, it seems that that's where the bar has been set.

I agree with skeptical that the sample argument I offered didn't capture that idea that these arguments, "fulfill the theological role of showing the way God is intimately involved with everything". That was part of what I was getting at. It showed that there was something odd to the combination of claims that (a) soundness reasonably accepted as such by only those antecedently committed to theism is a sufficient condition for a kind of argumentative success and (b) that the success of such arguments had to do with, "fulfill the theological role of showing the way God is intimately involved with everything".

Anyway, fwiw, it seems to be that running through the argument (such as it is) from the existence of sunny days or the deliciousness of ice cream to God's existence may be useful for all sorts of things without being successful in any way qua argument.

Hello Kevin.
I don’t have a point, I was asking a question about how professionals view the distinction between philosophy and apologetics. Is it a question of higher and lower “bars” for the rigor or persuasiveness of the work, or does it just speak to the audience which is being addressed?

But, looking again at the post, I see Dr. Pruss said he was discussing a “theological and completely non-apologetic” (?) value to the arguments for God. So I would need to try to distinguish theology from both philosophy and apologetics as well.

If this doesn’t make sense or discussing it is too off-topic, please consider the question withdrawn. Best regards, - Steve

Hey Steve,

As I read a lot of Pruss's arguments for the existence of God, found here at Prosblogion and other blogs, I think the difference Pruss is hitting on is that our arguments for God's existence help us understand more fully the particulars of God's character, epecially for those who already accept God's existence. I also do not think these types of "theological arguments" are freed from exemplifying the same logical rigor we would expect for any other philosophical argument. I guess since these arguments are not designed to compel belief in God from those who do not, they could be defined as non-apologetic in value, and more theological in value as they guide those who do believe, hopefully, to a better understanding of the person(s) of God, His attributes, His desires, et cetera.

I see a lot of overlap between philosophy, theology and apologetics. So, I am not sure one branch necessarily excludes the others, but I guess the clearest lines of demarcation are made by their subject matter or goal. If some argument for God's existence compels zero atheists towards belief but compels 50% of theists towards some particular belief about God, this would reveal its theological value as opposed to its apologetic value.

Mike and kevin
Thanks for your comments.

Mike: You wrote. "What is wrong with arguments that include necessary existence as a perfection (hence possessed by a perfect being)?'

Nothing, as long as it is argued for and not simply the result of defining it as such. The point I was trying to make is that defining A as x does not mean that A exists. If we define God as a perfect being with certain characterisitics I have not proven that God exists. That it my only, rather trivial, point. I was trying to get to the evidence that justifies us believing that God as defined by the theists exists and what I am suggesting is that the evidence does support the existence of God, if one grants creation, but does not entail that he has any perfections only the abiliyt to create the universe. This is certainy less then what theists maintain. I take it that theists consider omniscience, omnipotence, omnipresence, and complete goodness to be perfections/ Am I wrong in this?

Kevin, I agree with what you said. I trust no one thinks that I believe theism and theists to be confused and muddleheaded. How can one not appreciate the arguments of Anselm, Aquinas, Plantinga and others even if they are not convincing to some of us. I find the theodicies of Hicks and M. Adams to be thought provoking and inspirational. I am sure many have found guidance and comfort in these works.

One of the reasons I comment on this blog is that I enjoy the exchange of ides. I trust I offend no one.

Steve:

There is at least a stylistic difference between academically working on arguments for the existence of God and doing apologetics, in that most apologetics is not aimed at fellow researchers, while the academic arguments are written for an audience of fellow researchers. For instance, while one must duly acknowledge the contributions of others, it is not necessary to do a full bibliography when writing apologetics. On the other hand, popular writing has much higher stylistic demands.

Clayton:

I think "successful" is relative to a particular agent or set of agents. But the point I am making is that a sound argument need not be successful for it to have value.

Suppose, for instance, that as a sociological fact almost everybody who accepts that marital contraception is always wrong also believes that a deity exists. I am pretty sure this is true, at lest in regard to Americans. However, a sound argument whose only controversial premise is that marital contraception is always wrong (assume for the sake of the example that this premise is true) and whose conclusion is that a deity exists might well, nonetheless, be illuminating, showing a connection between two very different ideas. It would not be successful, but could advance the understand of how sexual ethics ties into theology.

I slightly screwed up my example: the statistical assumption I need is that almost nobody who does not believe in a deity believes in the wrongness of contraception.

However, the assertion that they might be successful arguments even if no reasonable atheist or agnostic could be moved by them seems too low a bar.

Clayton, I don't know, does it ever happen (I can say that I have never once witnessed) that someone S reads an argument A that is inconsistent with their view V on any topic (ethics, philosophy of mind, epistemology, etc.) and S concludes on the basis of A that V is wrong? That just never happens. They might be strongly influenced by A or moved by A to reconsider, and so on. But it seems to me a wildly simplistic criterion that an argument succeeds (as an argument) only if it convinces reasonable people of its conclusion. Given that criterion, virtually no philosophical argument has ever been successful. In that case theistic arguments are no worse off than arguments anywhere else in philosophy.

I appreciate the responses - that's helpful. - Steve

Mike,

I think I'm in agreement with part of what you're saying, but that's why I think the "could" is crucially important:
the assertion that they might be successful arguments even if no reasonable atheist or agnostic _could_ be moved by them seems too low a bar. However, by suggesting that some sound arguments that only a theist could reasonably regard as sound could be successful, it seems that that's where the bar has been set.

Anyway, I'm often convinced by arguments, but that doesn't mean I'm reasonable. It seems that that the bar originally set by Alex was still so low that it didn't matter whether the argument could lead a single, rational atheist or agnostic to reconsider.

Alex, this is a really interesting suggestion.

But...

Let's imagine some argument for God's existence that wouldn't be successful qua argument because one of the premises would be accepted only by those already convinced of God's existence, but which might help illuminate unexpected connections between e.g., contraception and God's existence. OK. In that case, why frame things (misleadingly) in terms of a question-begging, lame-looking argument, rather than straightforwardly making the supposed connection? Making the connection via a question-begging argument seems an oddly indirect and probably less effective way of proceeding.

Mike,

Isn't the premise that God possibly exists the one that atheists will question? Then isn't it question-begging?

Steve,

People mean a lot of different things by apologetics. Some take it to be the desire to get a position to come out true not matter what, even if it means giving bad arguments. Then they contrast that with philosophy, which follows the arguments wherever they lead. While there are certainly apologists who do this, even some who are well-known philosophers, I think it's unfair to define apologetics that way.

My own view is that apologetics is simply defending the faith. You can set out to defend the faith by giving a philosophical argument, all the while being open to discovering that the argument is bad. Thus you are doing both philosophy and apologetics at the same time. You can also set out to explore an argument, not caring whether it has apologetic value, discovering in the end that you think it does. I would say that you weren't originally doing apologetics there, but you might use it apologetically to defend the faith later.

So it's philosophy when it's exploring where an argument leads and whether an argument is good, and it's apologetics if you use it to defend the faith. Neither practice requires the other, but neither practice precludes the other. Both are perfectly legitimate things to do and can involve good motives, and both can be done poorly or with bad motives.

Since theology has come up, I'll say that I think of theology as simply thinking about God. That can be philosophical. It can be apologetical. It can be from the discipline of history of thought. It can be exegetical based on some religious text. It can be based on a systematic construction of a comprehensive theological position. So a lot of things can count as theology as well, and some of them are philosophical, while others are not. I would hold that whether it is philosophical or something else has little bearing on whether it is good thinking. So again we have different approaches that can involve different motivations and different methods, with good and bad uses of those different approaches and methods.

Why not say an argument is successful when it's sound?

The question as to whether it convinces reasonable people strikes me as irrelevant to an argument's success.

Reasonable people with nutty prior probabilities will turn an argument against God's existence into a modus tollens against a premise that other reasonable people will think unlikely given their priors.

At best, we can measure success by the commitments a denyer in the conclusion must revise in order to be consistent. But with different priors it will be question begging as to whether revision is reasonable.

Jeremy, you write,

Isn't the premise that God possibly exists the one that atheists will question? Then isn't it question-begging?

This isn't the premise that John thought was question-begging. I guess a theist might be accused of begging the question on the possibiity of God's existence, but I don't think they simply assert it. Kurt Godel and David Johnson have explicit proofs of the possibility. Others (Plantinga, say)rely on modal intuition. Neither involves just assuming that God exists.

Clayton, you write,

. . . but that's why I think the "could" is crucially important . . .

Right, yes, but the modal claim is bizarre. Isn't it? There is this trivial way of saying it in (1),
1. (Vx)[](if Atheist(x) then x is not convinced by the argument).
But obviously, if you're an atheist, then you're not convinced by the theistic argument. So (1) is trivial. So (2),
2. (Vx)(if Atheist(x) then [](x is not convinced by the argument)).
But that's strange too. It entails that even in worlds where an actual atheist is not an atheist, he's not convinced by the argument. I'm not sure that's possible, unless the argument is so bad that most theists are not convinced by it, either. I might reluctantly concede that any argument such that neither atheists nor theists are convinced by it is an unsuccessful argument. I say reluctantly, because (i) I don't think theists become so because they're convinced by some theistic argument and (ii) I think there could be very good arguments that convince no one.

One of my problems with arguments for/against theism is that they are most often stated in a 'if-then' formulation without ever removing the 'if.'

I am presently reading Rowe's Can God Be Free and am constantly finding statements such as "...I think the following [Principle B] is necessarily true: B. If an omniscent being creates a world when it could have created a better world, then it is possible that there be a being morally better than it."

Grant that this is necessarily true, the question remains does such a being exist? We must remember that from 'it is possible that x exists,' it follows that 'it is possible that x does not exist." It appears to me that many philosophers get wrapped up in analysing the implications and entailments of 'if-then' scenarios and forget that they have proved nothing about whether what they are analysing actually exists.

I would never deny that God's existence is a possiblity; there certainly is no contradiction in saying that God exists, or for that matter, that God does not exist. But that , in itself, is uninteresting. As we all know, one of these claims must be true. The question is, which one? My point is that a definition will not suffice. What is needed is empirical evidence. Moses had his burning bush, Thomas had his confrontation 'with the risen Christ,' Aquinas starts with observations about the world, as does Paley. Even Hume admits that arguments that start with obervation are closer to the human heart then those that start with abstractons. Simply relying on the possiblity that x exists and trying to draw anything of human significance from this seems to me to be missing the point; we do not want to know if it is possible that God exist and has a certain nature and what is consistent if we happen to believe that this type of God exists, we want to know if such a God actually exists, or at a minimum, if the evidence supports a belief that he exists.

I am going to end this comment with a polemical statement. If we need the concept of possible worlds along with a possible world semantics in order to understand God, then so few people are going to understand him that it makes him directly unapproachable. It creates, and reinforces, the notion that there must be a body of people who have the expertise to understand God that we must approach and listen to in order to have a glimpse of what he is and how he wants us to live, if he even cares. If I were a God, I would make it a tad bit easier for the average person to have the knowledge that I exist, so people interested in knowing me would not have to rely simply on faith derived from and reinforced by the pronouncements of 'experts.'

John you write,

. . . We must remember that from 'it is possible that x exists,' it follows that 'it is possible that x does not exist."

Not so, John. It follows rather from 'it is a contingent fact that God exists' that 'it is possible that God does not exist'. But suppose as the Anselmians do that God is a necessary being (not a contingent one). In that case it follows that if God possibly exists then God necessarily exists. That is, if God possibly exists then it is not possible that he does not exist. This holds for all necessary objects, not just for theistic objects. For instance, if there exists a greatest prime number, then there necessarily exists a greatest prime; if there is a best possible world, then there necessarily is a best possible world; if there is a proposition that is identical to the set of all possible worlds, then there is necessarily such a proposition, and so on.

Mike
I take it that part of the meaning of 'it is possible that x' means that I do not know that x is the case. If I do not know that x is the case, then it is possible -x is the case.

Please get rid of the 'if'in your examples so I do not have to suppose anything. The examples that you give are necessarily true, but what does they tell us about the world we live in? Is there a largest prime number? Is there a best possible world? etc. Even if I grant that there is a necessary being, call it God as Aquinas says we do, what does this prove about it's nature; only that it necessarily exists.

Thought of a question. Why can't the God we are discussing be part of a line of contigent beings who have the ability to create worlds of other contingent things? For each possible world there is a God who created it. Would we be stuck in an infinite regress?

I take it that part of the meaning of 'it is possible that x' means that I do not know that x is the case. If I do not know that x is the case, then it is possible -x is the case.

That's not part of the 'meaning' of 'it is possible that' unless we're talking about epistemic possiblity. But we've been about metaphysical possibility and necessity. To say that God necessarily exists, is to say that God exists as a matter of metaphysical necessity.

You add this,
Please get rid of the 'if'in your examples so I do not have to suppose anything. The examples that you give are necessarily true, but what does they tell us about the world we live in? Is there a largest prime number? Is there a best possible world? etc. Even if I grant that there is a necessary being, call it God as Aquinas says we do, what does this prove about it's nature; only that it necessarily exists.

I don't think you take the point. I was trying to show you that arguing from (1) to (2) is not idiosyncratic.

1. Possibly p
2. :. Not possible not p

Reasoning in this way is not peculiar to theistic arguments. It is characteristic of arguments regarding things that exists necessarily, if they exist at all. It was my effort to reassure you that there is no trick in the reasoning, that's all.
I understand that you want to know unconditionally whether God exists, whether there is a best possible world, whether there is a greatest prime. Who doesn't?

Part of the above discussion is about what it means for an argument to be a success. For those that haven't already read it, I'd recommend lecture III in van Inwagen's The Problem of Evil. And here is a related passage that I thought I'd throw out you for those who might be interested:

"Can one prove the existence of God?" seems to me less a clear question than a tangle of questions wwhich are rarely if ever sorted out and carefully distinguished. To most secular (that is, atheist) philosophers who consider the question at all, it seems clear and the answer seems all too obvious: the traditional ... 'proofs for the existence of God' are one and all fallacious, and the very idea of proving the existence of God is absurd. Moreover, the majority of these philosophers take it to be quite clear what a 'proof' is: a demonstration that something is the case using the standards (or supposed standards) of, if not science, then, let us say, analytic philosophy. In addition, it is supposed that a sound proof ought to be able to convince any rational person who sees is. (Why the arguments of analytic philosophers themselves--not even the philosophical, as opposed to technical logical, arguments of Frege, or Russell, or Quine, or Davidson or David Lewis--all fail to meet this test is not something that analytic philosophers discuss a great deal.)

Mike
Thanks for your responses (and your patience). I am talking about epistemic possibility. I accept your point about the difference between what I am doing and what you are doing. But even if you demonstrate that God metaphysically necessarily exists, which I am willing to grant, what else does this tell us about God? At no point in any comment on any thread posted on this blog, have I never denied that God exists. I have simply stated that I do not believe in God as defined by the theist because, for one reason, even if we grant that God necessarily exists this tells us nothing about his nature. Consequently God could be completely evil and fulfll the metaphysical necessity of existing. God's nature are epistemic issues, I think, about what we can know about what exists regrding God. Your comment regarding my desire to unconditionally know, etc. seems to miss the point that I am making. As philosophers (if I may be called one) should we not try to settle these epistemic issues? Simply discussing metaphyical implications leaves the discussion as an 'if-then.' If we cannot answer these epistmic questions then let us admit that philosophy's chief merit is that it can discuss implications of what it is to believe in x, but that it cannot provide any reason for believing in x.

John, Anselm's ontological argument isn't supposed to show that some otherwise vacuous necessary being exists. It's supposed to show that a perfect being exists, i.e. a being than whom no greater being can even be conceived. Take your pick of whatever way something can be good. This being is as good as possible in that respect and any other you can come up with.

So if his argument is sound, you can indeed derive quite a bit about God.

I looked up the word 'possible' and found that it means, "that may be or may become true, opposed to actual; said of a thing, an event, or a statement." Websters') The definition also includes the idea of contingency. So that is how I am using it, and I suspect that most, non-philosophers would use it the same way. Draw your own conclusions as to how it applies to the discussions above, but I stand by my usage until someone explains why I should not accept this definition as applying to standard usage.

Jeremy
I think I understand what Anselm's argument is trying to do. The problem is that there appears to be an implied equivalence between 'perfection' and 'completely good' which I think is mistaken. Descartes makes the same mistake when he argues that deception is an imperfection therefore God being perfect would not deceive. Deception would be and evil for a completely good being but not for a compeltely evil one. A being who is completely evil needs nothing else to be more evil so it is perfectly evil. A perfectly evil being is consistent with Anslem's idea that God is "a being whom no greater being can be conceived.".

John, the word 'possible' is a technical term in philosophy.
"X is possible" is defined as "not necessary that not-X", and "X is necessary" is defined reversely as "not possible that not-X". An easy way to think of the three terms:

necessary -- true in all possible worlds
possible -- true in at least one possible world
contingent -- false in at least one possible world

Given these definitions, it's easy to show that anything necessary is also possible, and therefore possibility implies neither necessity nor contingency.

This isn't about what the words mean in English. It's a definition that philosophers have agreed as a group to use because it allows for clarity of discussion to have agreed-upon definitions for the sake of precision and consistency across various publications. So dictionaries may point out broader meanings of these terms and so on, but it doesn't affect what the technical terms mean in philosophy.

John,

The philosophical uses of modal terms like 'possible' and 'necessary' have become very precise and are used in very technical ways. Webster's is not going to get you the precision you need here. For instance, if Webster's is conflating the contingent with the possible, it's really not serving you well.

Mike wrote:
I think there could be very good arguments that convince no one.

I agree. A bit of empirical evidence, a touch of self-deception, and now I believe that most of my arguments have those qualities.

I think we're still talking past each other, because I still find myself agreeing with nearly everything you say.

When I wrote:
However, the assertion that they might be successful arguments even if no reasonable atheist or agnostic could be moved by them seems too low a bar.

I had in mind a situation in which prior to reflecting on argument A, we had a somewhat idealized rational atheist or agnostic such that A couldn't offer them reasons to revise their attitudes by decreasing confidence in atheism or increasing confidence in theism. The claim was (roughly) that it's quite odd to say:

(1) There is a sense in which "A is a successful argument for theism" is true.
(2) There is no possible rational agent who is either atheist or agnostic at t who could at t2 rationally decrease confidence in atheism or increase confidence in theism having considered A.

If the rules of rational belief revision and updating couldn't lead someone to change their confidence levels in any way having exposed them to A, it seems (1) is false. Anyway, I think this is how I'd answer Christian Lee's question as to why I wouldn't say an argument is successful if sound. The argument I offered in the first comment on this thread is sound if the theist is right, but it's not successful as an argument in any way. And I think I'm in agreement with Tim's point that the supposed virtues of such arguments really have to do with the way reflecting on the considerations constituting the argument help us grasp certain connections between concepts, considerations, ideas, etc... and not due to anything positive about the argument qua argument.

Anyway, I'm off. I need to try to prove that knowledge isn't the norm of belief. Apparently that sort of thing won't just prove itself.

Mike and Jeremy
I have granted you the use of your terms re the metaphysical realm. What I am interested in, and something no one has addressed, is, given that God metaphysically necessarily exists what does this tell us about the nature of God. Frankly, what is the significance of proving that a necessary bieng exists if it cannot lead us to a greater understanding as to the nature of that being? And it is that part that is essentially epistemological and not metaphysical; what can we know about God.

(1) There is a sense in which "A is a successful argument for theism" is true.
(2) There is no possible rational agent who is either atheist or agnostic at t who could at t2 rationally decrease confidence in atheism or increase confidence in theism having considered A.

I agree, Clayton. That would be strange. But I don't think Alex wanted to make a claim about every possible rational agent of the sort you describe.

John,

"what is the significance of proving that a necessary bieng exists if it cannot lead us to a greater understanding as to the nature of that being?"

Well, if God exists then, metaphysical naturalism doesn't obtain. So, there is value in that. Also, if metaphysical naturalism doesn't obtain, we ought not be so sure that methodological naturalism is the best way to pursue all scientific inquiry. Which means our the structure of our public schools system ought to be changed. Also, if we have good reason to believe God's existence is metaphysically necessary then, maybe God plays a role in grounding morality--seeing as naturalism has so much trouble with this. Which may point us towards the character of this Necessary Being.

Furthermore, if this is the case then, Moral relativism might be a false view of our world. If necessarily God exists then, maybe the claims that miracles have occurred are worth further investigation--maybe we have written them off too quickly. If a miracle has ocurred at some time in history then, maybe this proves that this Necessary Being is personal and able to interact with creation in a meaningful way. Maybe the character of these miracles can further point us to the character of God.

If God's existence is metaphysically necessary then, religious beliefs in this Necessary Being are justified and maybe atheism and other religions that deny this Necessary Being are not.

Frankly, if we can prove God's existence, this matters greatly and we have a lot of work to do to ensure our world does not write this Necessary Being off any more than it has already done. Thus, God is not dead! Rejoice, John!

If the standard for a successful argument is that it increases the epistemic probability of the conclusion for some possible agent, then just about every sound argument is successful, unless it is question-begging in the strictest sense of having the conclusion literally among the premises. (There are probably some uninteresting counterexamples containing premises like "(4) No rational person believes premise (3).")

Take the following sound argument:
1. Horses exist and horses exist. (Premise.)
2. Therefore, horses exists.

Even this argument is successful by the above standard. (The example I am using is essentially due to somebody at Rutgers.) For consider a possible agent who was told by someone he knows to be very reliable: "Horses exist and horses exist." Right after having heard the sentence, the rational agent goes through the following reasoning:
a. Jones said: "Horses exist and horses exist."
b. Jones is reliable.
c. Therefore, probably, horses exist and horses exist.
d. Therefore, horses exist.

At the point of having arrived at step c, our agent is precisely in the position in which the argument (1)-(2) would increase her credence in the existence of horses.

So the standard is too weak.

--I still don't see why we're considering how rational people would respond to an argument. If prior probabilities are unconstrained, then rational people can disagree as whther rejecting the conclusion of an argument is is the correct way to respond to an argument, or whether, instead, denying some premise in the argument is the correct way to respond.

We constrain a reasonable reponse by conditionalization. So, the right response will be fixed by one's priors and rules for updating belief.

There's an independently interesting question as to which priors are admissible. It seems like 0 and 1 are both bad. Moreover, it seems that priors ought to be as fined grained as possible and should refer to fundamental facts.

There's also the question as to which arguments are interesting. But this will vary between people with different interests. That said, I'm having a difficult time seeing what the debate over an argument's success amounts to.

Kevin
1) Why presume that the necessary being is God?
2) Science has its own sphere of inquiry, religion has it's and it is best not to equate the two. The fact that utilizing the scientific method does not answer any religious questions is not a limitation on science, nor does it prove that religious ascertions are not true.
3) You seem to confuse political issues, the structure of our schools, with metaphysical/epistemological issues.
4) The fact that there are miracles does not prove there is a God, only that there are things we have not been able to explain using other certain established criteria.
5. You write; "Frankly, if we can prove God's existence, this matters greatly and we have a lot of work to do to ensure our world does not write this Necessary Being off any more than it has already done. Thus, God is not dead!" You conclusion does not follow from your premises.

In short, you have not answer my initial question andy more then Kevin or Jeremy have. You have not gotten rid of the 'if's' and 'maybe's.' You have simply asserted some ideas without argument.

I would rejoice if just once someone on this blog would answer my straightforward question; what can we know (epistemologically) about God's nature if a necessary being exists. I have granted that a necessary being exists, it is up to those who believe in theism to do the rest.

Dear John:

Questions 3-23 of the First Part of the Summa Theologiae are an argument that the First Cause, if it exists, is God and has the attributes traditionally attributed to him. I do not know how successful these arguments are: one day I want to make a careful study of them, but I have not done so yet. I suspect, however, that if these arguments work, one could argue if there is a necessary being, then there is a necessary being that has the attributes traditionally attributed to God.

There's also the question as to which arguments are interesting. But this will vary between people with different interests. That said, I'm having a difficult time seeing what the debate over an argument's success amounts to.

Christian,

If we take some Bayesian approach to belief revision as you seem to be suggesting (setting aside the problem of establishing priors) then I suppose the question would move in the obvious direction of trying to provide an argument that would persuade an agnostic. We have the same problem. Agnostics will assign various credences to incoming evidence (how?, well, subjectively) plug them into their theorems and generate a posterior probablity. Can a theist provide an argument for God's existence that will have every agnostic update his credences positively? I'm doubtful. Does that mean none of these arguments is successful? I'm doubtful again.

"If prior probabilities are unconstrained, then rational people can disagree as whther rejecting the conclusion of an argument is is the correct way to respond to an argument, or whether, instead, denying some premise in the argument is the correct way to respond."

But why think that prior probabilities are unconstrained? Take Parmie who assigns probability 1-10-100 to the claim that he is identical with a giant undifferentiated ball that constitutes the whole cosmos, and who assigns probability less than 10-100 to the claim that his senses are reliable. Since he has no reason to think he needs to eat, Parmie starves to death.

Now, Parmie is crazy--if he really lived by his beliefs and if people got to him soon enough, he might be institutionalized and force fed. Yet if there are no constraints on priors, he might be quite rational.

I guess this doesn't recognize the superscript tag. "10-100" should have "-100" superscripted.

John,

you wrote: "1) Why presume that the necessary being is God?"

Come on man, who else could it be?

you wrote: "2) Science has its own sphere of inquiry, religion has it's and it is best not to equate the two. The fact that utilizing the scientific method does not answer any religious questions is not a limitation on science, nor does it prove that religious ascertions are not true."

See, but your still responding from a naturalistic perspective--we have already ruled that out, remember? How can we be so confident religion and science do not overlap? How can we be so confident that science cannot answer any religious questions? For a skeptic you sure do seem confident on this.

you wrote: "3) You seem to confuse political issues, the structure of our schools, with metaphysical/epistemological issues."

Do you really believe that what takes place in philosophy classroom has no bearing on our culture? Do you really believe that our culture's particular institutions are not already grounded on some metaphysical/epistemological foundation? Sure they are, come on.

you wrote: "4) The fact that there are miracles does not prove there is a God, only that there are things we have not been able to explain using other certain established criteria."

There you go with that naturalism of the gaps thinking. But, remember, we have already dismissed naturalism. We cannot, now, dismiss miracles with some naturalism of the gaps philosophy.

you write: "5. You write; "You conclusion does not follow from your premises."

What premises?

you wrote: "In short, you have not answer my initial question andy more then Kevin or Jeremy have. You have not gotten rid of the 'if's' and 'maybe's.' You have simply asserted some ideas without argument."

Why are you so confident that we have to get rid of the 'it's' and 'maybe's'?

you wrote: "I would rejoice if just once someone on this blog would answer my straightforward question; what can we know (epistemologically) about God's nature if a necessary being exists. I have granted that a necessary being exists, it is up to those who believe in theism to do the rest"

Now, I think you are getting back to Pruss's original point for his post. Simply because you are unwilling to accept any of God's characteristics, this does not mean that we cannot know any of them. Just because you are willing to grant that a necessary being exists but at the same time believe we can go no further, what does this imply for the theist? Nothing. The theists' arguments for the existence and character of God do not have to convince everyone, i.e. you, to be useful. If you refuse to believe we can know anything further about this necessary being, why should the theist have any responsibility to convince you otherwise? The skepticism of some does not halt the theist from moving forward.

Now, Parmie is crazy

I agree. But this points to other well-known worries aside from determining priors. Suppose you want to know the posterior, P(G/k) (the probablity that God exists on evidence k). You are urged to use (1).

1. P(G/k)=P(G) x P(k/G)/P(k)

So now you're wondering what the priors are on G and what the posterior is on k, or P(k/G). Naturally for the posterior P(k/G) you move to (2),

2. P(k/G)=P(k) x P(G/k)/P(G)

In order the know the value of P(G/k) you need to know the value of
P(k/G). But in order to know P(k/G) you have to know the value of P(G/k). That's a bit of an obstacle. So there's lots of room for disagreement here on probability assignments.

Kevin
1) I presumed that when you used the word 'God' you were referring to a being that has the characteristics commonly given to him by theists. Simply proving that there is a necessary being does not, in itself, warrant any belief about the characteristics that that being might have, other then it necessarily exists. Furthermore, Spinoza defined God as the necessary being, but he certainly was not a theist. I suspect that he would be what you refer to as a metaphysical naturalist, so it is conceivable that God exist but is identical to the processes found in nature.
2) I have never claimed to be skeptical. As far as the distinction is concerned I am simply following Descartes.
3) I do not think that Intelligent Design is a scientific theory; how would one test for it? However this does not mean that it should not be taught in school, just not in science classes.
4) Premises are "if we can prove God's existence, this matters greatly and we have a lot of work to do to ensure our world does not write this Necessary Being off any more than it has already done." Your conclusion that "God is not dead" does not follow because you have not removed the 'if.' Hence the importance of removing 'ifs' and 'maybes.'

I think that if we, as philosophers, are simply content to analyze the implications of 'if-then' scenarios, then we are engaged in mental masturbation and we ought to go get real jobs. If the best we can do as philosophers is to demonstrate that if God is omniscient, or whatever, then he can do x,y, and z, then I think that is not enough. We are doing God a great disservice if we cannot remove the 'if' and demonstrate the there is a God who has these characteristics. Or at least provide some rational justification for believing, based on the evidence that the best explanation for the way things are is that there is a God who has these characteristics.
5) Of course it is true that if a person is not convinced of x this does not mean that we cannot know x. If that is all Alexander is claiming then he is engaged in a really trivial endeavor. I think his endeavor is more worthy then simply making this point. I have also never claimed that we cannot go further. I have simply claimed that no one has gone further if they have not removed the 'if's and 'maybes.' Maybe, if you explain to me why you find the arguments for assigning certain characteristics to God convincing, then maybe it will be convinced. I am willing to follow the argument wherever it leads me, but so far no one will lead me.
6) You write, "If you refuse to believe we can know anything further about this necessary being, why should the theist have any responsibility to convince you otherwise? The skepticism of some does not halt the theist from moving forward." I think you have a moral responsibility to what your conceive to be the truth to argue with those you do not share your beliefs. You should not retreat, but should man the barricades and argue for the truth of your position. Is that not what the disciples and martyrs did? It seems to me that on this blog when anyone asks a really hard question, that people shy away from (ignore) it. Should I infer from this that people do not really have any rational basis for believing what they do? I am willing to put my beliefs to the test, why aren't you? One of the reasons I comment on this blog (and asked for a quest post) was that I know my beliefs are different from those on this blog. I want them tested.

Mike,

"If we take some Bayesian approach to belief revision as you seem to be suggesting (setting aside the problem of establishing priors) then I suppose the question would move in the obvious direction of trying to provide an argument that would persuade an agnostic."

I am sugesting the Bayesian approach. And I do agree that establishing priors is a problem. But I don't see why there is any special problems for persuading an agnostic.

If an agnostic assigns .5 probablity to (gods exists given their background knowledge) then adding some argument, say the design argument, to their background knowledge where they assign probability .9 to (God exists given the design argument) will raise the probability that God exists on their background evidence.

Alex,

Parmie is crazy!

"But why think that prior probabilities are unconstrained."

I don't. I think they are constrained in the sense that (A) we shouldn't assign 1 or 0 to priors, (B) we should apply the principle of indifference to a partion of priors that is fine grained, and (C) the priors should include only fundamental facts.

But my main point is just that someone who believes it's very likely that God exists, say .99 likely, then when they face, say, the argument from evil they will rationally lower their probabilities in some of its premises, rather than significantly lowering their probability in its conclusion, though they will need to lower the probability that God exists a little bit.

I'm not going to do the calculation. I have to pack. But Mike can probably do it in 5 seconds.

If an agnostic assigns .5 probablity to (gods exists given their background knowledge) then adding some argument, say the design argument, to their background knowledge where they assign probability .9 to (God exists given the design argument) will raise the probability that God exists on their background evidence.

Yes, of course it will. But now we get around to the problem. There is nothing that precludes one agnostic from finding the design argument silly and non-evidential and another finding it fairly credible. Obviously this can happen, since this is exactly what does happen. But now we have some people claiming that therefore the design argument is unsuccessful and others denying that therefore it is unsuccessful.
So this debate about the successfulness of theistic arguments is not resolved by going Bayesian. You seemed to be suggesting that it is so resolved.

Yeah, I'm suggesting it's resolved on the conditions that we go Bayesian and constrain our priors. Maybe I'm just wrong. But I take it that convergence theorems together with independent constrainst on priors show "something like this" is true.

"There is nothing that precludes one agnostic from finding the design argument silly and non-evidential and another finding it fairly credible."

Absolutely. But one should only find it silly if the conclusion of the argument has a low posterior probability when it's premises are included with one's background evidence. One should find it persuasive if the conclusion is high when the premises are included with their background knowledge. And that's just my point, that whether an argument should move us one way or the other depends upon one's background evidence.

In another way, when we face an inconsistent set of claims given our background evidence the probability that those claims are jointly true is 0. So, we must redistribute our credences in those claims by conditionalizing on our evidence and even an agnostic must do this. Perhaps an agnostic will remain an agnostic and rationally so, or change belief one way or the other, and raionally so, but what he should do is just be consistent and be Bayesian. We then can fault him only for being inconsistent or updating incorrectly or having nutty priors.

So, I'm deny that a successful argument entails that it will rationally lead "anyone" to raise one's credence in the conclusion. One could, instead, rationally respond by lowering one's credences in the premises and as far as I can tell, this is just "successful" as pushing someone towards the conclusion.

But one should only find it silly if the conclusion of the argument has a low posterior probability when it's premises are included with one's background evidence. One should find it persuasive if the conclusion is high when the premises are included with their background knowledge

Right. Look, suppose my evidence (premises) are E and I add that to k (my background evidence for/against God existing). Let my priors be as follows,

1. P(G/k) = .5

Suppose add E to k to get posteriors in (2),

2. P(G/k & E) =.5 x P(E/G & k)/P(E/k)

Suppose an agnostic does not doubt at all P(E/k) = 1. In that case (3) is true.

3. P(G /k & E) is guaranteed to be less than its priors.

But obviously agnostics could argue over the value of P(E/k), of course, since it is equivalent to,

4. P(E/k)=[.5 x P(E/G&k)]+ [.5 x P(E/~G&k)].

I might think that the chances of my having evidence E on the assumption that God does not exist is not so high. Suppose I put my credence at .6. In that case the highest the value of P(E/k) could have is .8. In that case my posterior for G, P(G /k & E) is greater than its priors P(G/k) (its over .6, if I got the numbers right)

But what's the point here? The point is that there is not much constraining my assignment of credences to P(E/k) or P(E/G&k) or P(E/~G&k). So it is very easy for agnostics to have there belief in God confirmed or their belief in God disconfirmed, given any theistic argument. If theistic arguments are successful only if they confirm and agnostic in his belief that God exists, then very few arguments are going to meet that standard. That has been the point under discussion all along. Going Bayesian adds a few constraints on consistent probability asisgments, but it just doesn't solve the philosophical problem.

Mike,

I'll think more about this. I'm taking off for awhile, but here are my thoughts.

"But obviously agnostics could argue over the value of P(E/k), of course, since it is equivalent to,

4. P(E/k)=[.5 x P(E/G&k)]+ [.5 x P(E/~G&k)]."

But P(E/k) = P(E)P(k)/P(E) by conditional probability.

So, agnostics cannot argue over P(E/k) without arguing over either P(E) or P(k). But to do this requires arguing over the prior that is P(E). They will not argue about P(k) since it's just assumed as background knowledge.

"The point is that there is not much constraining my assignment of credences to P(E/k) or P(E/G&k) or P(E/~G&k)."

I'm suggesting one's priors constrain P(E) and so P(~E) and so P(E/k).

"So it is very easy for agnostics to have there belief in God confirmed or their belief in God disconfirmed, given any theistic argument."

Yes, sort of. But only if they are consistent and that requires lowering their probabilities in those premises that jointly entail conclusions that have low probabilities on their background knowledge.

"If theistic arguments are successful only if they confirm and agnostic in his belief that God exists, then very few arguments are going to meet that standard."

I agree. I don't think that's a good standard. I think a good argument, roughly, will show incompatibilities between propositions such that, for example, a theist will have to deny some claim they would have assigned a high credence to, or in the case of an atheist, they will have to lower their credence in some proposition they would have assigned a high prior to. A really good argument will move one to reject a claim they otherwise might have accepted, where this just means they will significantly change their credence is some claim or other that had a high prior.

So, I'm denying that a successful argument for P must raise the probability that P on one's background knowledge. A successful argument for P can do this "or" force one to lower one's credence is some premise that one would otherwise have assigned a different probability to. In short, successful arguments move from otherwise pplausible premises to implausible conclusions where plausibility is just a rough way of saying that "most people" would have high credences in the premises and low credence in the conclusion.

Christian:

In the principle of indifference, what do you with normalization? Consider the hypotheses H_n: "There are exactly n photons in the world." What probability do we assign to each H_n?

But P(E/k) = P(E)P(k)/P(E) by conditional probability.

I don't think that can be right. You mean, Pr(E/k)=Pr(E & k)/Pr(k).

So, agnostics cannot argue over P(E/k) without arguing over either P(E) or P(k). But to do this requires arguing over the prior that is P(E). They will not argue about P(k) since it's just assumed as background knowledge.

If you assume that k is certain in the right side of your definition (presumably, because it is background knowledge in P(E/k)) then you're going to reduce the definition to P(E/k)= P(E). That follows from your assumption that k is certain or P(k)=1. You don't want that assumption. So, no doubt, agnostics are going to argue about the value of P(k) and the value of P(E). And depending on your choices for credences, you'll have the same evidence confirming and disconfirming theism. But apart from this, lots of people take conditional probabilities as primitive or undefined. The only advantage to adding these definitions is that you find it easier to work with unconditional probabilities, as some people do. But adding them certainly does not solve the problem under discussion.

Jeremy and Mike
In our earlier discussion on my use of the word 'possible' you both claimed that I was not using the word the way philosophers are using it. You claim that my use lacks the necessary precision. Jeremy wrote: "This isn't about what the words mean in English. It's a definition that philosophers have agreed as a group to use because it allows for clarity of discussion to have agreed-upon definitions for the sake of precision and consistency across various publications. So dictionaries may point out broader meanings of these terms and so on, but it doesn't affect what the technical terms mean in philosophy."

Correct me if I am wrong, but is there the unanimity among philsophers that you seem to be claiming regarding how we define modal terms, or even the importance of modal logic? Have not some suggested that it might be the case the when philosophers redefine words in order to get more precision what is actually occuring (unintended)is that they are refocusing the situation so that the results they are looking for can be found. Certainly if we define 'possible' as you have then it follows that if it is possible that x exists it is also impossible that x does not exist (something I denied). But why should we accept that definition. It seems to simply define the outcome in favor of those you want to seperate 'contingency' and 'possible.' But if we follow Humes who maintains that contingency and possibility are closely linked why should we make this move? It seems to me that the philosopher who would move away from standard usage needs to justify the move and that the move cannot be justifed simply because it allows us to reach an outcome we desire. Furthermore, does not making such a move create a serious disconnect between philosophers and non-philosophers who are not using terms the way philosophers are?

Mike and Jeremy

Here are two arguments utilizing yours and my definitions of 'possible.' I will assume that God is a possible being.

Yours:
1) A possible being is one that exists in at least one possible world
2) God is a possible being
2) Therefore God exists in at least one possible so world.

Mine:
1) A possible being is one who may or may not exist.
2) God is a possible being
3) Therefore God may or may not exist.

Because we are both relying on how we define 'possible' it is important that we have an argument that supports our usage. My argument in favor of my definition is that it is the one that is consistent with standard usage. I am simply analyzing the implications of how the word is commonly used. Your definition deviates from standard usage so I think you need to justify this deviance in such a way as not to beg the question. To reiterate a point I raised earlier, I think that relying on possible worlds and possible worlds semantics in order to understand God is very problematic.

It is probably clear that I disagree with Alexander's claim state in the last paragraph of the post. There he wrote: "What is particularly interesting here is that the arguments do not need to be successful in the full sense of being such that they would convince every reasonable person in order to fulfill the theological role of showing the way God is intimately involved with everything. Of course the arguments need to be sound: the premises must be true and the conclusion must follow from them. But to fulfill this theological role they need not convince the atheists or agnostic. It suffices that the theist see that they are sound and why they are sound."

There is a difference between 1) accepting a premises as being true, 2) agreeing that a premise is true, and 3) demonstrating that a premise is true. If we accept either 1 or 2 as the criterion that must be met in order for a premise to be true then a person can fail to accept presmies based ont these criteria and still be reasonable. However, if we hold out for #3 as the standard that must be met in order for a premise to be true and we meet this standard then it would not be reasonable for someone to not accept it.
I suspect that it comes down to the difference between being a person who accepts a coherence theory of truth( more consistent with 1 and 2) versus one that accepts a correspondence theory of truth (more consistent with 3).

Yours:
1) A possible being is one that exists in at least one possible world
2) God is a possible being
2) Therefore God exists in at least one possible so world.
Mine:
1) A possible being is one who may or may not exist.
2) God is a possible being
3) Therefore God may or may not exist.
Because we are both relying on how we define 'possible' it is important that we have an argument that supports our usage.

There is no conflict between these arguments. The conclusions do not conflict and the premises do not conflict. So where's the problem? You say you want to use 'possibly' for epistemic possibility. That's fine, but your argument does not use the notion of epistemic possibility. This one does,