Is It Impossible to Prove Atheism?

(Originally published on 1 September 2023; revised version published on 7 September 2023; revised again on 10 September 2023.)

1: Introduction

“Performative" or "pragmatic inconsistency arguments” attempt to show that some action with respect to some proposition is inconsistent with the proposition. You will sometimes hear people call such propositions “self-defeating,” but that isn’t technically accurate. It isn’t the proposition itself which is self-defeating; it is taking some action with respect to the proposition -- like asserting, believing, or defending it – which is self-defeating. For example, consider the proposition, “It is impossible to know anything.” It is logically possible that the proposition is true, but it is logically inconsistent to know that proposition is true. If you somehow knew that it was true, then you would know that it is false. That’s an example of a performative or pragmatic inconsistency argument.

Here are some other examples of propositions which lend themselves nicely to performative or pragmatic inconsistency arguments:

• "I do not exist."
• "I am dead."
• "I cannot assert or communicate anything."
• "I cannot talk."
• "I don't intend to be understood at all."

All of the above examples involve a first-person ("I") statement, but I'm not sure if a performative or pragmatic inconsistency must be a first-person statement. I'll leave that aside.

2: Exposition

Does atheism involve a pragmatic inconsistency? Before I can discuss this, I first need to define my terms. The word "atheism" is polysemous: it has multiple legitimate meanings. By "metaphysical atheism," I mean the proposition, "God does not exist." By "psychological atheism," I mean the psychological state in which a person lacks the belief that God exists. This post is about metaphysical, not psychological, atheism. For brevity, in the remainder of the post, I will use "atheism" to mean only metaphysical atheism.

According to some thinkers, the action of asserting the truth of atheism is a pragmatic inconsistency. For example, the late Christian philosopher Greg Bahnsen once said:

It's standard fare these days to point out that a universal negative cannot  be proven in an a posteriori fashion.  On the basis of experience, how could I prove universally there is no God? [...] God might be on the back side of the fourth moon of Jupiter.  Since I haven't examined the back side of the fourth moon of Jupiter, it's at least possible that God exists and I haven't encountered him yet.  So to say that there is no God is a very strong claim.  Atheists usually get hammered about that, that they're claiming too much.  In fact, in order to say 'there is no God,' ... you would have to be omniscient.  But if you were omniscient, you would be God. Therefore, the only person who could say 'there is no God' is God.  God  would not deny Himself.  So no one can say there is no God.[1]

This suggests the following argument, which I call the "unprovability of universal negatives argument."

(1) The only way to prove a universal negative in an a posteriori fashion is to exhaustively search for all possible sources of evidence of a thing and find no evidence.

(2) But it is impossible for finite, mortal beings to conduct such a search.

Therefore:

(3) It is impossible for finite, mortal beings to prove a universal negative in an a posteriori fashion.

Defenders of the "unprovability of atheism argument" add two additional steps:

(4) The proposition, "God does not exist," is a universal negative.

From (3) and (4), it follows that:

(5) It is impossible for finite, mortal beings to prove, "God does not exist," in an a posteriori fashion.

In the passage quoted above, however, Bahnsen suggests another argument which adds onto the first five steps. Atheism entails that there is no God, but the action of claiming to know that atheism is true presupposes that the person making that assertion has God-like knowledge. So we have:

(6) The action of claiming to know, "God does not exist," presupposes that the speaker is God. 

From (5) and (6), it follows that:

(7) If a finite, moral being claims to know, "God does not exist," then it is true that God does not exist and it is false that God does not exist.

From (7) and the Law of Noncontradiction, it follows that: 

(8) Any claims made by a finite, mortal being to know, "God does not exist," on an a posteriori basis should not be accepted.

I shall call the inference from (5) - (7) the "pragmatic inconsistency of knowing atheism" argument.  

3: Assessment

I shall argue that both the unprovability of atheism argument and the related pragmatic inconsistency of atheism arguments are unsound.

3.1: The Unprovability of Atheism Argument

Here, again, is the unprovability of atheism argument.

(3) It is impossible for finite, mortal beings to prove a universal negative in an a posteriori fashion.

(4) The proposition, "God does not exist," is a universal negative.

Therefore:

(5) It is impossible for finite, mortal beings to prove, "God does not exist," in an a posteriori fashion.

One immediate problem with this argument is its name. From the fact, if it is a fact, that it is impossible for finite, mortal beings to prove atheism in an a posteriori fashion, it does not follow that it is impossible for such beings to prove atheism in a priori fashion. One or more internal inconsistency arguments against God's existence might be sound. But let that pass.

Some philosophers deny the truth of the second premise, (4). For example, in his famous debate with Frank Zindler, William Lane Craig argued as follows.

[...] the statement that God does not exist is not a universal negative; it’s a singular negative statement. Certainly you can prove negative singular statements such as there is no planet between Venus and the Earth. You could provide arguments to show that a singular negative statement is true [...].[2]

Strictly speaking, the first quoted sentence is correct. Categorical propositions are propositions which assert or deny relationships between classes; universal negatives are categorical propositions which have the form, "No S are P." Atheism is not a universal negative in this strict sense. But the statement, "God does not exist," is the same as saying, "No (existing) thing is God." That restatement of atheism is a universal generalization; it is has the form, "There are no S's."[3] Therefore, a proponent of the unprovability of atheism argument might replace "universal negative" with "universal generalization" yielding the following revised version of the "unprovability of universal negatives argument," which gets renamed as the "unprovability of universal generalizations argument."

(1') The only way to prove a universal generalization in an a posteriori fashion is to exhaustively search for all possible sources of evidence of a thing and find no evidence.

(2) But it is impossible for finite, mortal beings to conduct such a search.

Therefore:

(3') It is impossible for finite, mortal beings to prove a universal generalization in an a posteriori fashion.

The "unprovability of atheism argument" would then become:

(4') The proposition, "God does not exist," is a universal generalization.

Therefore:

(5) It is impossible for finite, mortal beings to prove, "God does not exist," in an a posteriori fashion.

The meaning of the word "prove" is ambiguous, but the last two steps of the pragmatic inconsistency argument (steps 6 and 7) provide a clue. That argument targets not atheism, but what is sometimes called "gnostic atheism," the claim to know that God does not exist. This suggests that, at least for Bahnsen, to "prove" a universal generalization means to succeed in demonstrating that belief in the universal generalization can rise to the level of knowledge. So let's revise the first two arguments one more time.

The "unprovability of universal generalizations argument" gets renamed as the "unknowability of universal generalizations argument."

(1'') The only way to reasonably believe a negative universal generalization on an a posteriori basis is to exhaustively search for all possible sources of evidence of a thing and find no evidence.

(2) But it is impossible for finite, mortal beings to conduct such a search.

Therefore:

(3'') It is impossible for finite, mortal beings to reasonably believe a negative universal generalization in an a posteriori fashion.

Likewise, the "unprovability of atheism argument" becomes the "unknowability of atheism argument."

(4') The proposition, "God does not exist," is a negative universal generalization.

Therefore:

(5') It is impossible for finite, mortal beings to reasonably believe, "God does not exist," on an a posteriori basis.

Now consider the first two premises (steps 1'' and 2) in light of this clarification of "prove." 

The second premise (step 2) is false. Purely on the basis of experience, one can know that at least some negative universal generalizations are true. Some objects or beings are defined in such a way that, if they exist, they exist at a specific location and at a specific time. For example, consider the title, "Seattle Seahawks Superbowl 60 Most Valuable Player (MVP)." Let "NO SEAHAWKS MVP" refer to the negative universal generalization, "The Seattle Seahawks Superbowl 60 MVP does not exist." If the Seattle Seahawks do not advance to Superbowl 60 (or if they advance but do not win), then NO SEAHAWKS MVP is true: the person described by that title does not exist. In that case, an "exhaustive search" doesn't require superhuman or supernatural abilities. One could reasonably believe NO SEAHAWKS MVP by simply watching Superbowl 60 or reading about it after the fact.

One might object that, even in my highly specified example, the inference to NO SEAHAWKS MVP isn't deductively valid; it is at least possible that there is a global media conspiracy to spread misinformation about the outcome of Superbowl 60, one's sensory organs are defective, or one is literally delusional and imagining sense data not received by one's sensory organs. This objection fails because one does not need absolute certainty to be reasonable in holding a belief. Even if one did not know NO SEAHAWKS MVP, one would still be reasonable in believing NO SEAHAWKS MVP. Contrary to what Bahnsen implies, a person can reasonably believe something they don't claim to know. For the same reason, Bahnsen has not given us a good reason to believe that the first premise (step 1'') is true. 

A proponent of the unknowability of atheism argument might grant the above objections and concede that humans can reasonably believe at least some negative universal generalizations, while at the same time insist that there is something special about God such that humans cannot reasonably believe atheism. Here I think it is useful to follow Jeanine Diller in making a distinction between global and local versions of atheism.[4] "Global atheism" is the proposition, "God does not exist," for all concepts of God, including concepts of God which no one has yet even considered. In contrast, "local atheism" is the proposition, "God does not exist," for a specific concept of God. This distinction suggests two further versions of the "unknowability of atheism" argument.

I will call the first of these two versions the "unknowability of global atheism" argument.

(1*) The only way to reasonably believe global atheism on an a posteriori basis is to exhaustively search for all possible sources of evidence of God, for every possible concept of God, and find no evidence.

(2) But it is impossible for finite, mortal beings to conduct such a search.

Therefore: 

(5*) It is impossible for finite, mortal beings to reasonably believe global atheism on an a posteriori basis.

The second version, of course, is the "unknowability of local atheism" argument.

(1**) The only way to reasonably believe local atheism on an a posteriori basis is to exhaustively search for all possible sources of evidence of God, for a single concept of God, and find no evidence.

(2) But it is impossible for finite, mortal beings to conduct such a search.

Therefore: 

(5**) It is impossible for finite, mortal beings to reasonably believe local atheism on an a posteriori basis.

The first argument is, I believe, much more promising than the second. In light of the fact that there are surely legitimate concepts of "God" which no human has ever conceived, it is difficult to imagine how, on the basis of experience, one could be confident in the belief that all local theisms are false. In other words, in the context of global atheism, the second premise (step 2) seems to be true. Of course, the fact that I cannot conceive of such a justification does not rule out the possibility that one exists. But it seems unlikely that anyone in the year 2023 could reasonably believe global atheism is true. As John Schellenberg has persuasively argued, there are numerous signs of 

human intellectual immaturity at the macro level, and we've seen enough evidence of it to say that maturity--the realization (or near realization) of a deep and wide understanding of reality--may yet be a long way off.[5]

Further support for Schellenberg's immaturity hypothesis comes from the work of Paul Draper. In a recent article, Draper explores the conjunction of "merotheism" (God is a proper part of nature) and "monopsychism" (there is exactly one mind or subject of consciousness). He calls this novel version of theism "panpsychotheism" (all-minds-[are]-God-ism).[6] The important point to notice here is that, although some of the greatest minds in history have thought about God, there is no indication that anyone else had thought of this view. All by itself, this single example is evidence that there are other conceptions of God which no one has previously conceived. 

In contrast, I think the second argument fails and for essentially the same reason given earlier. At least some God concepts entail or make probable certain states of affairs which be confirmed or disconfirmed by experience. For example, suppose that, in honor of Richard Swinburne, we define a "Swinburne God" as

a being with most of the following properties: being a person without a body (that is, a spirit), present everywhere (that is, omnipresent), the creator of the universe, perfectly free, able to do anything (that is, omnipotent), knowing all things (that is, omniscient), perfectly good, a source of moral obligation, eternal, a necessary being, holy, and worthy of worship. [7]

Let "Swinburnian theism" be the proposition, "The Swinburne God exists." Swinburnian theism clearly entails other propositions; one does not need to know that one of those entailments is false in order to reasonably believe that one of those entailments (and, by extension, Swinburnian theism itself) is false.[8] This is why I think the second premise (step 2) of the "unknowability of local atheism" argument fails.

3.2: The Pragmatic Inconsistency of Believing Atheism Argument

In the last section, we saw that the unprovability of atheism argument is not successful, but can be at least partially repaired by transforming it into the unknowability of atheism argument. The conclusion of that argument is:

(5') It is impossible for finite, mortal beings to reasonably believe, "God does not exist," on an a posteriori basis.

Defenders of the pragmatic inconsistency argument should then revise the last three steps as follows.

(6') The action of reasonably believing, "God does not exist," presupposes that the speaker is God. 

From (5') and (6'), it follows that: 

(7') If a finite, mortal being reasonably believes, "God does not exist," on an a posteriori basis then it is true that God does not exist and it is false that God does not exist.

From (7') and the Law of Noncontradiction, it follows that: 

(8') It is impossible for a finite, mortal being to reasonably believe, "God does not exist," on an a posteriori basis. 

This argument fails for two reasons. First, its second premise, (6'), is false. Even if it is impossible for finite, mortal beings to reasonably believe atheism on an a posteriori basis, it does not follow that it is impossible for such beings to reasonably believe atheism on an a priori basis. One does not have to be omniscient in order to know, much less reasonably believe, that there is a successful a priori argument for local atheism.

For example, here is one schema for what I call an "internal inconsistency argument" for local atheism.

(9) If God existed, then He would have property P1.

(10) If God existed, then He would have property P2.

(11) Any being with property P1 would have some further property P3.

(12) Any being with property P2 would not have property P3.

(13) Therefore, it is impossible for any being to have properties P1 and P2. (from 11 and 12)

(14) Hence, it is impossible for God to exist. (from 9, 10 and 13)

The basic idea is to show that if God existed, then he would have two properties P1 and P2. The next step is to show that property P1 entails some third property P3, while property P2 entails the denial of property P3. Since it is impossible for any being both to have and not have P3, it follows that that being cannot exist. While it might or might not be the case that there are no successful arguments of this type for local atheism, it really doesn't matter one way or the other for the purpose of this post. The speaker does not need to be omniscient in order to know the truth of (9) - (12).

Second, the argument's conclusion, (8'), does not follow from (5') - (7'). (6') refers to the action of claiming to reasonably believe, "God does not exist," but does not specify that that knowledge is based upon an a posteriori basis.

As with the unknowability of atheism argument, let us now consider two versions of the pragmatic inconsistency argument based on the distinction between global and local atheism. The global version can be summarized as follows:

(5*) It is impossible for finite, mortal beings to reasonably believe global atheism on an a posteriori basis.

(6*) The action of reasonably believing global atheism presupposes that the speaker is God. 

Therefore:

(7*) If a finite, mortal being reasonably believes global atheism on an a posteriori basis, then global atheism is true and global atheism is false.

Therefore:

(8*) It is impossible for a finite, mortal being to reasonably believe global atheism on an a posteriori basis. 

There are two problems with this argument. The first is that it is invalid: (6*) refers to the action of claiming to reasonably believe global atheism but does not specify that that knowledge is based upon an a posteriori basis. It seems to me that this objection could be avoided by revising the argument again as follows:

(5**) It is impossible for finite, mortal beings to reasonably believe global atheism.

(6**) The action of reasonably believing global atheism presupposes that the speaker is God. 

Therefore:

(7**) If a finite, mortal being reasonably believes global atheism, then global atheism is true and global atheism is false.

Therefore:

(8**) It is impossible for a finite, mortal being to reasonably believe global atheism.

That solves the problem identified by the first objection, but brings into sharper focus the second objection. Even if we assume that premise (5**) can be adequately defended, premise (6**) does not seem to be true. If premise (5**) is true, then it seems that the action of claiming to reasonably believe global atheism presupposes that the speaker has knowledge which greatly exceeds that of any human being in history. But "has knowledge which greatly exceeds that of any human being in history" is not a sufficient condition for being God. First, it's far from obvious that such knowledge requires omniscience, and many conceptions of God posit an omniscient being, not just a really smart being. Second, even if the speaker's actions did presuppose that the speaker is omniscient, being omniscient is not a sufficient condition for being God. It is at least conceivable that there exists an omniscient being which does not have any of the other attributes typically associated with God, such as omnipotence, moral perfection, eternity, aseity, sovereignty, immutability, necessity, etc.  

I don't know how Bahnsen would respond to this objection. Perhaps Bahnsen could show that the speaker's actions require such a superhuman degree of knowledge that the speaker's actions presuppose that the speaker is not necessarily God, but at least a supernatural being. Since metaphysical naturalism rules out the existence of any supernatural beings, not just God, the argument could then be transformed into a pragmatic inconsistency argument against believing metaphysical naturalism. 

Now consider the local version of the pragmatic inconsistency argument, which can be summarized as follows:

(5#) It is impossible for finite, mortal beings to reasonably believe local atheism on an a posteriori basis.

(6#) The action of reasonably believing local atheism presupposes that the speaker is God. 

Therefore:

(7#) If a finite, mortal being reasonably believes global atheism on an a posteriori basis, then global atheism is true and global atheism is false.

Therefore:

(8#) It is impossible for a finite, mortal being to reasonably believe local atheism on an a posteriori basis.  

This argument suffers from the same defects as its global counterpart. Unlike the global pragmatic inconsistency argument, however, I am doubtful that the second premise, in this case (6#), can be repaired by replacing "God" with "supernatural being." For ordinary human knowledge is all that is needed to reasonably believe that some local theisms are false. Thus, the action of claiming to reasonably believe local atheism does not appear to be a pragmatic inconsistency, at least not in any sense related to the issues discussed in this post. 

4: Conclusion

In this post, I have considered two arguments: the unprovability of atheism argument and the pragmatic inconsistency of knowing atheism argument. The unprovability of atheism argument is best and most charitably understood as the unknowability of atheism argument, which has two versions. The unknowability of global atheism argument seems promising, but the unknowability of local atheism argument fails. For at least some concepts of "God," one can reasonably believe that local atheism is true. Finally, the pragmatic inconsistency argument fails because one does not need to be superhuman, much less omniscient, in order to reasonably believe the truth of the premises of an internal inconsistency argument. Perhaps there is a successful global pragmatic inconsistency argument (against the action of believing naturalism), but the local pragmatic inconsistency argument seems hopeless.[9]

Notes

[1] “Michael Martin Under the Microscope” audiocassette, tape 3.

[2] William Lane Craig, “Second Rebuttal Speech” Atheism vs. Christianity (January 1993), https://www.reasonablefaith.org/media/debates/atheism-vs.-christianity

[3] Thanks to Paul Draper for helping me to see this clearly.

[4] Diller, J. "Global and local atheisms." Int J Philos Relig 79, 7–18 (2016). https://doi.org/10.1007/s11153-015-9550-1

[5] J.L. Schellenberg, Progressive Atheism: How Moral Evolution Changes the God Debate (New York: Bloombsury, 2019), p. 50.

[6] Paul Draper, "Panpsychotheism." In Current Controversies in Philosophy of Religion (New York: Routledge, 1990), pp. 160-177.

[7] Richard Swinburne, The Coherence of Theism (2nd ed., New York: Oxford University Press, 2016), p. 1.

[8] For an example of such an argument, see Paul Draper's formulation and defense of his "Decisive Evidence Argument" in Draper, Paul, "Atheism and Agnosticism", The Stanford Encyclopedia of Philosophy (Summer 2022 Edition), Edward N. Zalta (ed.), URL = <https://plato.stanford.edu/archives/sum2022/entries/atheism-agnosticism/>.

[9] Thanks to Paul Draper for some very helpful conversations which significantly improved the quality of this post. Any errors are, of course, solely my responsibility.

Empathy and Apathy: An Evidential Problem for Theists

Serial killer Ted Bundy

Image Source: Florida Memory Project; License: Public Domain

1. Definitions

Following Paul Draper, I'm going to define my terms as follows.[1]

By "naturalism," I mean the view that the physical exists and, if the mental exists, the physical explains why the mental exists.  If naturalism is true, then there are no purely mental beings which can exist apart from a physical body and so there is no God or any person or being much like God.

By "supernaturalism," I mean the view that the mental exists and, if the physical exists, the mental explains why anything physical exists.  If supernaturalism is true, then there is no purely physical matter which can exist without some sort of ultimate mental creator. "Personal supernaturalism" is a type of supernaturalism; it adds on the claims that one or more personal mental entities exist and, if a physical world exists, it or they produced the physical world for a purpose. "Theism" is a type of personal supernaturalism; it adds on the claim that there is just one mental entity, God, who is all-powerful, all-knowing, and morally perfect. 

"Otherism" is a catch-all category. It says that both naturalism and supernaturalism are false. 

Podcast Interview with Justin Schieber of Real Atheology

Justin Schieber

A few weeks ago, Justin Schieber interviewed me for an episode of the Real Atheology podcast. That episode just dropped. 

RA040: Jeffery Jay Lowder on Philosophy of Religion

"In this episode, Justin Schieber sits down with Jeffery Jay Lowder for a wide-ranging interview. Fans of the show will no doubt be familiar with Jeff's work. From founding Infidels.org and the Secular Outpost to his contributions to The Empty Tomb and his several public debates, Jeff has earned his reputation as a fair-minded and philosophically informed atheist. We discuss the origin of infidels.org, Paul Draper's famous argument from Pain and Pleasure and everything in between."

LINK

Some Very Rough and Incomplete Thoughts on the Prior Improbability of the Alleged Resurrection of Jesus

Bayes Theorem

Source: Matt Buck; License: CC BY-SA 3.0

DISCLAIMER (5 May 2023): I consider this a partial blog post on the prior improbability of the alleged resurrection of Jesus. I use the word "partial" because I'm not presently motivated to write what I would consider to be the entire blog post. If my level of motivation changes in the future, I will update the blog post.

Reply to a (Sunni) Islamic Theodicy for the Argument from Pain and Pleasure

In reply to Paul Draper's evidential argument from pain and pleasure, a self-identified Sunni Muslim proposed the following theodicy:

T*: God exists, and God uses human pain to purify the soul.

As he explained it to me:

"In even a prick of the finger, the Muslim has sins wiped away from him and in such pain as if he is made anew if he has actually taken the Hadith of the Prophet any further than his own lips. So while it may not fully fulfill the purpose of detection of threats or illnesses, it does its job in relation to the Islamic framework assuming its truth."

In order to evaluate this theodicy, we should apply Draper's Weighted Average Principle (WAP). I will assume that readers are already familiar with this principle; readers who need an introduction may consult my explanation in section 3.3.2 here. I will argue that this Islamic theodicy does not significantly raise Pr(O | theism) for two reasons. First, I will argue that Pr(T* | T) is not high. Second, I will argue that Pr(O | T*) is not high. 

The Holocaust is Strong Evidence Against Theism

(Revised 27 April 2022)

(Revised 1 March 2022)

(Revised 14 February 2022)

(Revised 7 February 2022)

The following tweet recently appeared on my Twitter feed: 

Primo Levi Auschwitz survivor
"The notions that our world comes from God and is governed by God are hard to sustain if one looks seriously at evil. What egocentric presumption to think that God listens for our repentance and acts accordingly!" pic.twitter.com/J2gNxTEdNt

— Dave of Letters (@liverational22) August 6, 2021

In this blog post, I intend to defend a weaker version of Primo Levi's claim. I do not claim that Auschwitz makes God's existence impossible. Rather, I intend to argue that the Holocaust is extremely strong evidence against theism and, other evidence held equal, makes God's existence improbable.

Baggett and Walls on 'The' Moral Argument vs. 'The' Argument from Evil

David Baggett	Jerry Walls

The idea that there is some sort of connection between moral arguments for theism, on the one hand and arguments from evil against theism, on the other hand, is not new. In this blog post, I want to comment on a very cryptic statement made by Baggett and Walls about this connection. In their book God and Cosmos, they write:

... Second, it should be said in such a case that the probability of theism has increased (by much or a little) relative to morality; in theory the probability of atheism could increase or decrease relative to other phenomena. (However, the success of the moral argument would decisively undercut the problem of evil, which tends to be counted as the best evidence against theism.) (20, boldface mine)

Is Evil Evidence for God?

(Another backlog item)
 
Over at The Secular Outpost, I recently blogged about a 'turnaround' argument used by some critics of the argument from evil (AE). According to this turnaround argument, not only is evil not evidence against God's existence, but evil is evidence for God's existence. In this post, I want to examine one specific attempt to use this strategy, by Melinda Penner at Stand to Reason.

My basic plan will be as follows:

1. Provide an overview of a specific version of the argument from evil.
2. Provide an overview of Melinda's use of the turnaround strategy.
3. Assess the evidential value of Melinda's turnaround strategy against the argument from evil.

Victor Reppert on the Argument from Evil as a Reductio

[This post was originally published on The Secular Outpost on June 12, 2012 and then republished on February 22, 2016. It was republished here on November 7, 2021 with the date manually adjusted to reflect its original publication date. The link was also updated to its current location.]

(Redated post originally published on 12 June 2012)

Can atheist who rejects moral realism press the argument from evil? Many theists, including William Lane Craig and Ravi Zacharias, have argued that the answer is "no." In my old critique of one of Zacharias's books, I wrote the following.

Zacharias presents two objections to AE. First, he suggests that it is incoherent for atheists to appeal to evil as evidence of the nonexistence of God since objective moral evil could not exist if there is no God (p. 48). Yet, as I argued above, objective moral values are logically possible even if there is no God. And the atheist need not postulate the existence of objective moral values in order to use evil as evidence for atheism. An argument from evil might not contain any normative premises; the atheist could appeal to God's loving nature rather than God's moral nature.[13] For example:

(1) If a perfectly loving God were to exist, then he would not permit the occurrence of any unjustifiable suffering.
(2) But unjustifiable suffering does occur.
(3) Therefore, a perfectly loving God does not exist.

To emphasize the point (though I consider this unnecessary), the atheist might even change the name of the argument from "the Argument from Evil" to "the Argument from Unjustifiable Suffering."[14]

Most importantly, AE may be understood as a challenge to the internal coherence of a theistic worldview. An AE can be understood as saying something like the following:

Look. You theists believe that X, Y, and Z are evil. You theists believe that God is good. You theists believe that good persons are opposed to evil. So you theists need to explain why a god who is good (in your sense of 'good') would allow so much apparently pointless evil (in your sense of 'evil'). If you can't explain it, then that is a problem for the internal coherence of your worldview.[15]

When AE is understood in this way, it doesn't presuppose that there are objective moral values.

In an old post, Victor Reppert says that he is aware of running the argument from evil as a reductio ad absurdum argument, but he is unconvinced. He writes:

I still maintain that important phenomena that give rise to the problem of evil are themselves deeply problematic for naturalism, including consciousness, objective moral values (and yes I know all about running it as a reductio without presupposing objective moral values--I still think subjectivism undercuts the argument).

Here's my question to Reppert: why? How does subjectivism undercut the argument? For example, how does subjectivism undercut Draper's version of the evidential argument from evil (summarized here)?

Biological Evolution as Evidence against Theism

Many conservative Christians and lay atheists alike claim that if biological evolution is true, then God does not exist. Ironically, while many conservative Christians have attacked evolution because it supposedly entails atheism, no contemporary atheist philosopher has used evolution as evidence for atheism. Indeed, the only philosopher who has formulated an argument for the claim that evolution is evidence against theism and for metaphysical naturalism is agnostic philosopher Paul Draper.

Draper defends an evidential argument from evolution for naturalism. In other words, Draper's argument does not claim that evolution is logically inconsistent with the existence of God. Rather, it claims that known facts about evolution that are consistent with theism nevertheless provide evidence against it.

This argument is focused on God in general, not necessarily the Christian God. Draper argues that, all other things held equal, known facts about the origin of complex life are prima facie evidence against theism.

Silver's Defense of Draper's Argument from the Biological Role of Pain and Pleasure

(Another item from my drafts folder.)

David Silver has written this interesting paper on the problem of evil, "Religious Experience and the Evidential Argument from Evil," Religious Studies 38(3):339-353.

The Argument from the Biological Role of Pain and Pleasure

Informal Statement of the Argument

The naturalistic explanation for this is obvious. If animals are the products of evolution by natural selection, we would expect physical pain to aid survival. But not all physical pain and pleasure aids survival. For example, think of the horrible pain that inflicts many people with terminal illnesses. If naturalism is true, this is what we would expect: evolution by natural selection is not an intelligent process; there seems to be no way for creatures to have evolved so that they only feel pain when it will aid survival. In contrast, if theism were true, God could "fine tune" humans so that they experience pain only when it is necessary for some greater good. If God did exist, what possible reason could he have for allowing people with terminal illnesses have to endure such agonizing pain until they finally die? The chances that such a reason would intersect with the biological goal of survival is pretty slim. Thus, the biological role of pain and pleasure is more likely on naturalism than on theism.

Some (Very Incomplete) Thoughts on Luke Muehlhauser's "How to Debate William Lane Craig"

[This post was originally published on The Secular Outpost on October 12, 2011. It was republished here on November 7, 2021 with the date manually adjusted to reflect its original publication date. The link was also updated to its current location.]

After writing a post about William Lane Craig and John Loftus debating, I remembered that Luke Muehlhauser (Common Sense Atheism) posted an article in April 2009 about debating William Lane Craig. (LINK) Here are some very incomplete thoughts about Luke's article.

1. I agree with Luke that many of Craig's debate opponents were unqualified, in the sense that they did not have both (a) the relevant knowledge (e.g., of philosophy of religion, metaethics, etc.); and (b) suitable debating experience.
2. I strongly disagree with Luke's assumption that Craig has 'won' literally "all" of his debates, but I do think he has 'won' most of them. Off the top of my head, I think the following opponents 'won' their debates with Craig:
   • Paul Draper
   • Doug Jesseph (the first debate, not the second debate which I consider a draw)
   • Shelly Kagan
   • Keith Parsons (both the debate on God and the debate on Christianity)
   • Michael Tooley
3. I agree with the five specific points in his section, "How to Win." On the other hand, in my experience, the people who would most benefit from following Luke's advice tend to be the same people least likely to follow it.

The Evidential Argument from Mind-Brain Dependence: A Reply to Bilbo

[This post was originally published on The Secular Outpost on November 26, 2006. It was republished here on November 7, 2021 with the date manually adjusted to reflect its original publication date. The link was also updated to its current location.]

In my response to Victor Reppert's anti-naturalistic argument from pain, I stated that a more specific fact (consciousness is dependent upon the physical brain) about consciousness is antecedently more likely on naturalism than on theism. Bilbo provided several feedback messages in response to this claim. As I read him, he thinks the argument is no threat to Christian theists. It is apparently his view that mind-brain dependence is at least as probable on Christian theism as on naturalism, since Christian theists believe in specific, sectarian doctrines that raise the probability of mind-brain dependence. Since I think Bilbo is wrong about this, I'll explain why.

First, here is the logical form of my evidential argument.

Let D = the mind is dependent upon the physical brain.
Let T = classical theism
Let N = metaphysical naturalism

(1) D is known to be true.
(2) T is not much more probable intrinsically than N.
(3) Pr(D/N) > Pr(D/T).
(4) Other evidence held equal, T is probably false.

Note the modest nature of this argument. It doesn't claim that theism is probably false; it merely claims that evidence (D), by itself, favors N over T. This distinction is crucial since it allows for the possibility that there could be other evidence that both favors T and outweighs the evidence for N provided by D. Thus, D should be understood as a prime facie reason for rejecting T.

Let me now turn to some of Bilbo's specific comments.

Bilbo: "My first problem is that it seems more accurate to say that minds *are* the physical brain on Naturalism (not merely dependent on them). And since this has not been demonstrated, it would beg the question to assert that the postulate that "minds are the physical brain" provides evidence for Naturalism."

That would indeed beg the question, which is why I never said that! My argument is an inductive argument. It would be illegitimate to state the evidence in such a way that entails the hypothesis to be proved. Therefore, I'll stick with my original formulation of the premise ("the mind is dependent upon the physical brain").

Bilbo: "But the Christian views I mentioned also make it antecedently probable that minds will be dependent on the brain. Let's use the structure of your argument with Christian non-reductive physicalism (hereafter NRP)..."

If theism is improbable given D, then so is Christian theism (or any other more specific belief system that entails theism). Christian theism entails theism; therefore it cannot be more probable than theism. Premise (4) entails that, other evidence held equal, Christian theism is probably false.

I don't deny the potential relevance of sectarian doctrines to the issue of whether my argument is sound. They could raise Pr(D/T) or lower Pr(D/N). In order to assess the evidential significance of such doctrines, we would need to apply a principle that Draper calls the "weighted average principle" (WAP). Let H represent some Christian doctrine. Then WAP can be represented as follows.

Pr(D/T) = Pr(H/T) x Pr(D/T&H;) + Pr(~H/T) x Pr(D/T&~H)

This formula is an average because Pr(H/T) + Pr(~H/T) = 1. It is not a simple straight average, however, since those two values may not equal 1/2.

Let us consider, then, Christian non-reductive physicalism (CNRP), which Bilbo proposes as a specific doctrine that he believes raises the probability of D given (Christian) theism. Bilbo defines CNRP as "the non-existence of all supernatural beings, except for those that are or were at some point divine (angels, demons, God, etc.)." Bilbo claims that CNRP "entails the nonexistence of disembodied human minds." Using WAP, we obtain the following.

Pr(D/T) = Pr(CNRP/T) x Pr(D/T&CNRP;) + Pr(~CNRP/T) x Pr(D/T&~CNRP)

In order to reject my evidential argument, therefore, Bilbo would need to show that CNRP raises Pr(D/T) so that it is greater than or equal to Pr(D/N) by using the above formula. Does CNRP do that? I shall leave the question as an exercise for the reader.

(Note: I have borrowed heavily from ideas in another paper by Paul Draper, "More Pain and Pleasure: A Reply to Otte." I am, of course, responsible for any errors in my post.)

Victor Reppert's Anti-Naturalistic Argument from Pain

[This post was originally published on The Secular Outpost on November 25, 2006 and then republished on February 11, 2016. It was republished here on November 7, 2021 with the date manually adjusted to reflect its original publication date. The link was also updated to its current location.]

There are a variety of approaches to formulating an argument from evil against theism. Two of the most influential versions of the evidential argument from evil were developed by atheist William Rowe and agnostic Paul Draper. Both involve appeals to pain. In a recent entry on his blog, Victor Reppert tries to turn the tables on proponents of arguments from evil (pain) by arguing that pain is a problem for atheists. Reppert formulates his argument against naturalism as follows.

1. If naturalism is true, then consciousness does not emerge.
2. If consciousness does not emerge, then pain does not exist.
3. Therefore, if naturalism is true, then pain does not exist.
4. Pain exists.
5. Therefore, naturalism is false.

Notice that premise (1) of the above argument entails that naturalism and consciousness are logically incompatible. Reppert's argument may be reasonably characterized, then, as a very specific example of a logical argument from consciousness against naturalism (as opposed to an evidential argument from consciousness as defended by Richard Swinburne). I believe this characterization is further supported by Reppert's statement, "the actual internally experienced state of pain is a huge, hard problem for atheistic naturalism, a problem that I personally consider to be logically impossible to solve."

Does Reppert literally mean what he says? As a so-called logical argument against naturalism, it isn't clear to me why this argument is in any way an improvement over logical arguments from evil against theism. In particular, Reppert has not yet presented a defense of (1) in its current form (i.e., lacking some sort of probabilistic hedge).

Perhaps, however, Reppert did not intend for his argument to be worded so strongly. Instead, he might say, we should understand his argument as an evidential argument against naturalism. If so, it seems to me that Reppert's argument from pain violates the Rule of Total Evidence. Allow me to explain. Let us assume, for the sake of argument that Reppert is correct that consciousness is antecedently more likely on theism than on naturalism. Furthermore, as Reppert points out, pain presupposes consciousness. It doesn't follow from these two assumptions, however, that the statement, "The kinds and distribution of pain we find in the world is epistemically more probable on naturalism than on theism," is false.

Here's why. Reppert has committed the fallacy of understated evidence. In general terms, we could say that Reppert has identified some general fact F (consciousness exists) about a topic X (consciousness) that is antecedently more likely on theism than on naturalism. It could still be the case that a more specific fact (S) about X is antecedently more likely on naturalism than on theism. As Paul Draper writes, "we know a lot more about phenomenal consciousness than just that it exists" ("Partisanship and Inquiry in the Philosophy of Religion," unpublished paper). Draper then gives an example of a more specific fact about consciousness, a fact that, given consciousness exists, is more probable on naturalism than on theism. As Draper explains, we also know from neuroscience that the mind is dependent upon the physical brain, a fact that is more likely on naturalism and consciousness than on theism and consciousness. Thus, as Draper concludes, "when the available evidence about consciousness is fully stated, it is far from clear that it significantly favors theism."

Sophisticated Critique of Many Worlds Explanation of Fine-Tuning

[This post was originally published on The Secular Outpost on February 18, 2006. It was republished here on November 7, 2021 with the date manually adjusted to reflect its original publication date. I have also updated the link to reflect the new location of the essay]

The following essay was recommended to me by Paul Draper. The paper is not a defense of the fine-tuning argument, but he regards it as one of the best critiques of the many worlds explanation of fine-tuning:

Roger White, "Fine-tuning and Multiple Universes," forthcoming in Nous
http://web.mit.edu/rog/www/papers/fine_tuning.pdf (PDF)

Here is some info about the author:

ROGER WHITE, (Ph.D., MIT), Assistant Professor of Philosophy, specializes in philosophy of science, epistemology, and metaphysics. He is currently focusing on epistemological issues in the philosophy of science, particularly those having to do with probability and explanation. His main interests in metaphysics concern matters of identity and essential properties. He is the author of "Fine-tuning and Multiple Universes" (Nous).