The Best Case for Atheism Online: A Reply

For Emerson Green

This post is inspired by Emerson Green’s case for atheism. I’m interested in Green’s case for three reasons. First, his case presents what I consider to be the best type of case for atheism—and his presentation is among the very best I’ve seen in the blogosphere.* Second, I like Emerson Green’s work, and I want more people to benefit from it. His work displays the qualities of honest truth-seeking, respectful engagement, and depth of insight; the more people who know about this work, the better the world will be. Third, it feels fun to share some reflections on Green’s case. (Related: the love language for philosophers is to argue with those they respect.)

*Another top representation of this type of case is Sebastian Montesino’s “Why I’m an Atheist,” which I also highly recommend.

I will offer a reflection. For efficiency, I’m utilizing Green’s presentation as a script. My reflection involves presenting parallel ideas on behalf of the theist.

Disclaimers:

1. My goal is not to refute Green’s ideas. Rather, I seek to offer a perspective of how certain things seem to me. Everything I say is fully compatible with everything Greens says, for what I say is merely a description of how things seem to me by my limited lights, and it is consistent with what Greens says that some things seem to me in certain ways that may differ from how they seem to Green.

2. This is not a complete analysis of my views, but a summary of some relevant ideas.

To begin, here is an outline of a reason one might be skeptical of atheism, which is a parallel of Green’s outline:

I. We don’t need atheism to explain anything.

II. There are few, if any, phenomena that are better explained by atheistic models than by theistic models.

III. There are many phenomena that are better explained by theistic models than by atheistic models.

IV. Atheism is more metaphysically profligate than theism as an explanation.

V. Atheism suffers from various internal problems.

I will consider each idea below.

I. We don’t need atheism to explain anything.

Here Green and I have a lot of agreement. I agree with Green that theists and atheists are both capable of accounting for what we know, however implausible those accounts may seem. For example, as Green points out, there’s no suffering so appalling that a theist couldn’t conjure up some logically possible explanation for why an all-good, all-powerful god would continue to behave in a way indistinguishable from non-existence.

In this respect, atheism and theism are symmetrical. If we don’t need theism to explain anything, we equally don’t need atheism to explain anything. There are ways proponents of both theories can always attempt to incorporate all the data they have.

Following Green, my analysis here is not deductive. While I think it is possible (at least in principle) for a person to deduce an answer to the question of God’s existence from other things they know, I do not think any proof (in any domain, not even in math or logic) proceeds from premises that everyone knows (or even should know given their current epistemic position). Non-deductive arguments also offer a more modest conclusion that could appeal to a wider audience. So, to parallel Green’s modesty, I will provide a non-deductive case. (My larger goal isn’t so much to show that my case succeeds in the end, whether objectively or subjectively, but to help display some principles that can serve all truth-seekers via parallel considerations.)

Finally, I also join Green in recognizing that there’s no such thing as a justified belief in the abstract. Justification, in my view, is person-based: something is justified to someone. What is evident to you depends on your total history of relevant experiences and considerations. For this reason, each of us occupies a unique position on the epistemic landscape (as I like to say)—or as Joe Schmid puts it—the grand epistemic landscape.

What I offer here, then, is my viewpoint from a certain position on the epistemic landscape. Things might appear to you differently, and that’s okay.

II. There are few, if any, phenomena that are better explained by atheistic models than by theistic models.

Green suggests that atheistic models better explain a lot of things, but I want to suggest a way in which the opposite could seem so.

First, atheism itself doesn’t predict or explain much of anything. Consider, for example, the proposition that there is no mind behind the universe or any other god. What follows from that? Does it follow that human beings evolved by natural processes? No. The proposition that no god exists is about what does not exist. From a proposition about what does not exist, it does not follow that there exists human beings, evolution, or even a universe! Hardly anything follows from atheism on its own.

In response, one might argue that a full atheistic model could have more explanatory power. After all, atheism does not exist in a vacuum on its own. Atheism fits into a larger set of background beliefs, including beliefs that theists and atheists usually agree on. For example, theists and atheists agree that there is a universe with humans in it. If there is a universe with humans in it, and if atheism is true, then that would seem to predict that humans evolved without divine intervention. So atheistic models can help explain lots of things.

Fair enough.

However, we need to be careful to compare apples with apples. If atheism picks up predictive (and explanatory) power when added to background beliefs, then so does theism. For example, if we conjoin theism with our background information about evolutionary biology, then we can successfully predict that humans likely evolved from previous ancestors via natural selection. Of course, this prediction does not on its own favor theism over atheism (unless there is something about theism on its own that raises my expectation of an evolutionary means of development). The point is this: if we compare apples to apples, we need to either compare atheism on its own with theism on its own, or we need to add one’s background beliefs to both.

Adding background beliefs is tricky. How do you determine which background beliefs to include? The challenge is to select background beliefs that are neutral between atheism and theism. That’s not easy.

The problem here isn’t only what Draper calls “the fallacy of understated evidence”—i.e., understating evidence against one’s view by pointing to general features (consciousness) in support of one’s view, while neglecting more specific evidence (painful torture of babies to death) against one’s view. That’s one problem.

There is another problem—what I’ll call “the fallacy of overstated evidence.” This fallacy occurs when one’s background information tucks evidence against one’s thesis. For example, the existence of blank spaces on Emerson Green’s birth certificate is evidence that Emerson Green was never born, since that would make those blank spaces more likely. But of course those blank spaces would only exist on his birth certificate if his birth certificate exists, which is excellent evidence that for an opposite conclusion! Here the background information—about the birth certificate—is not neutral with respect to Green’s birth.

The challenge of selecting neutral background beliefs affects our analysis of atheistic models. Consider that many atheistic models pick up explanatory power precisely by including information that is (arguably) more probable on theistic models. For example, an atheist model of morality could be thought to pick up explanatory power by taking for granted that there are principles of rationality, moral agents, and a context for moral interaction. But one could argue—and it has been argued—that the existence of principles of rationality, moral agents, and a context for moral interaction are themselves more expected on classical forms of theism (e.g., if fundamental reality is a rational moral agent whose nature grounds or includes principles of rationality).

Or consider an atheistic model that helps explain certain gratuitous suffering by building in knowledge of natural regularities that independently predict a high probability of gratuitous suffering. One could argue—and it has been argued—that the existence of natural regularities that facilitate environments for certain experiences in certain states of limitation (like in games) is what one would expect on a minimal form of theism (e.g., a mind-first theory of fundamental reality). So the existence of natural regularities is not neutral, which complicates our analysis of the evidential implications of gratuitous suffering. (For other ways atheistic models might inadvertently embed non-neutral background beliefs, see “A Problem of Evil for Atheism.”)

To be clear, I’m not saying there couldn’t be atheistic models that pick up predictions using neutral background beliefs. I just think it is tricky to identify such models, since there tends to be controversy over whether any background beliefs are neutral.

To sail on smoother waters (and to avoid being swallowed up by the fallacies of understated and overstated evidence), I suggest that we include no background beliefs in our models. Instead, we work with minimal versions of atheism and theism. We will still utilize background information, but we will use it as evidence of the data to be explained (and of the respective conditional probabilities), rather than tucking them into the models themselves. In this way, we can ensure we are comparing apples with apples, rather than comparing apples with oranges (stuffed with non-neutral background beliefs).

Once we work with minimal versions of the respective theories, I think it is fairly easy to see that minimal atheism does not predict or explain much on its own. Minimal atheism is a proposition that a certain type of being does not exist. Not much follows from that. In particular, it does not follow that any conscious, moral agents are likely to exist.

By contrast, minimal theism does seem to raise the probability of many things. For example, on the hypothesis that fundamental reality is itself a supreme conscious reality (per theism), one might find it more expected that there would be conscious, moral agents in a world meeting fine-tuned conditions to allow for the development of lifeforms of various kinds. Whether you agree with any particular prediction, my larger point is that it’s not clear to me that atheism on its own better predicts any known phenomenon.

Of course, this analysis is far too terse, but it’s a beginning (of something great!).

III. There are many phenomena that are better explained by theistic models than by atheistic models.

If atheism were true, what would I expect the world to look like? Would there even be a world?

Here’s a list of some lines of evidence that are more surprising to me on minimal atheism than on minimal theism:

– consciousness

– games

– plot devices (that develop sentient beings along various lines).

– the ability to reason

– principles of reason (e.g., logical truths, principles of rationality, etc.).

– moral sensations (of good, bad, progress, etc.).

– historical ironies that display profound truths (e.g., God’s light of reason within people sets them free from religion so they can see the greater truths, beauty, and goodness in the meadows of reality… to discover what was obscured by the dogmas of humans.)

– plots that include opportunities for valuable experiences in large contexts that individuals can choose to enter into for various reasons, and which are very, very unlikely to emerge out of arbitrary piles of mindless debris (on minimal atheism), such as these: plots involving anti-religious experience, psychopaths, tragedies & gratuitous suffering, divine silence during tragedies, biological roles of pain and pleasure, natural evil, teleological evil, evolutionary suffering, animal suffering, etc.—each of which is not an unlikely episode in any sufficiently large drama of great beings.

In light of these considerations, it seems to me that theism, in conjunction with natural explanations (themselves made more likely given mind-based orderly arenas), shines more light on more things than mere atheistic explanations.

I turn to an objection. Green writes:

Plate tectonic theory explains why there is a natural tendency for earthquakes to occur, and atheism explains why there is no one with the power and inclination to prevent earthquakes. So atheism, in conjunction with plate tectonic theory, predicts that earthquakes will occur from time to time and cause pointless suffering. Theists cannot help themselves to this explanation, since on theism, a being with both the power and inclination to prevent earthquakes and their devastating effects exists.

You can probably anticipate my response. My concern is with the neutrality of the background beliefs in play. Consider an analogy:

The rules of Minecraft explain why there is a natural tendency for zombies to come out at night, and the hypothesis that no good person made Minecraft explains why there is no one with the power and inclination to prevent those zombies. So this “no creator” hypothesis, in conjunction with Minecraft-zombie theory, predicts that you will get killed by zombies from time to time, which results in pointless suffering, both for the kids playing Minecrafe and their parents who hear them whining in the background. Those who think Minecraft was created by someone good cannot help themselves to this explanation, since on the creator hypothesis, someone with both the power and inclination to prevent zombies and their devastating effects exists.

Please know that this analogy is not intended to explain evil and suffering—it is not a good enough analogy for that.

Rather my point is to draw attention to the non-neutrality of certain background information. It is true that the rules of Minecraft can on their own explain the zombie attacks. No reference to a designer of the game is necessary. But of course, the Minecraft rules themselves are unlikely without some sort of designer. One could even think the rules of the game are evidence of a good designer aiming to create a context for good experiences—experiences good enough, at least, for many people to voluntarily play the game, despite the risk of psychological distress over the zombies. The point is that the context in which suffering is possible is not itself neutral with respect to the theories in question. So caution is required.

This reply does not imply that Green has not identified real challenges for theists to consider. I think he has. Indeed, Green’s challenges present valuable opportunities for us all to grow —and cooperate in growing—in our understanding of reality, which is itself a great good!

IV. Atheism is more metaphysically profligate than theism as an explanation.

I agree with Green that the simpler theory is more likely to be true.

As Green points out that Graham Oppy says,

“[T]he naturalist does not have beliefs in anything over and above the things the theist believes in. From the standpoint of the naturalist, the theistic beliefs of the theist are pure addition; and, from the standpoint of the theist, the naturalistic beliefs of the naturalist are pure subtraction. In short, naturalism is a simpler theory than theism. A central premise of my argument in support of atheism is the Principle of Parsimony. This general principle states that if there are two competing theories and one is simpler than the other, then, unless the more complex theory provides a better explanation of something than the simpler theory, one should endorse the simpler theory.”

For the fun of parody, I would like to propose a parallel of Green’s idea that theism is redundant, except I say that atheism is redundant. The natural world works without atheism; we have no need to appeal to atheism to explain the natural world. It all works without that hypothesis.

Let’s take a few examples. Atheists sometimes suppose that there is a natural order, and that this is best explained without god. But can theists provide an account of the natural order? Yes, they can. There’s no incompatibility between the existence of a natural order and theism. Indeed, as Felipe Leon helped me see (in our dialogue), one could think of God as a natural foundation of the natural reality. So one could be a theist and a naturalist. Hurray!

Moreover, theism itself could raise one’s expectation of a natural order. A natural order is a rule-governed arena, and one could expect rule-governed arenas to emerge from rational mind, especially if those arenas provide contexts for interesting experiences, like Zombie attacks. So not only isn’t a natural order something that requires atheism, one could even argue that a natural order is even more likely on theism.

I would also argue that theism allows us to simplify our theory of fundamental reality. The theist can say that fundamental reality is purely perfect—i.e., that it has no mixture or positive and negative properties, and nor does it have arbitrary limits in any basic respect, whether in power, awareness, or goodness. The pure perfection hypothesis is a relatively simple hypothesis from which we can derive many other aspects of fundamental reality. If instead fundamental reality has a mix of limits, then greater conceptual complexity is required to specify its basic nature. For example, it takes greater complexity to specify that fundamental reality has causal powers to make 10^90 particles and no powers to make thoughts than to say that it has the powers of a purely perfect foundation. Etc.

Bonus: contrary to many atheistic models, a theist can also decline to posit the general category of mind-independent phenomena. (For more on the simplicity of theism vs. naturalism, see “On the Simplicity of Theism vs. Naturalism.”)

I itch to add this: maybe nature is all there is. If we posit, in addition to the posit of a natural order, that there is no god (equivalently: that fundamental reality has some arbitrary limit in causal power [e.g., it cannot know anything] or imperfection), then we complicate our theory beyond necessity. To quote Green’s quoting of Laplace, “We have no need for that hypothesis.”

In summary, atheism adds unnecessary complexity to our theory of reality. If we don’t need god to explain anything, we equally don’t need no-god to explain anything—which, if you recall, was also Emerson Green’s first point! (Recall the point: theists and atheists are both capable of accounting for what we know.) If anything, theism allows us to simplify our theory of the nature of fundamental reality.

V. Atheism suffers from various internal problems.

My primary point here is that I agree with Green that versions of both atheism and theism are immune from many internal problems. The usual internal problems usually target specific versions of theism or atheism, not minimal versions of these theories.

I think Green will also agree that just as there are internal problems for certain theistic models, there are internal problems for certain atheistic models. For example, consider an atheistic materialist model on which the basic building blocks of reality are mindless, valueless spatial things (particles or fields). This model bumbs up against what seem at least to me to be 4 impossibilities:

1. Deriving mind from mindless debris.

2. Deriving unified consciousness from prior, disparate things.

3. Deriving value from purely valueless states.

4. Deriving rational thought (and its effects!) from reasonless motions.

More specific problems arise from more specific models. For example, a flying spaghetti monster god theory faces certain additional problems for the theist. Similarly, the theory that reality is fundamentally made of yogurt would face the additional problem of explaining how you could be made of yogurt. (Of course, no one holds these theories, but you get the point.)

Conclusion

I conclude with a parallel summary: since (i) we don’t need atheism to explain anything, (ii) there are few, if any, phenomena that are better explained by atheism itself than by theism, (iii) there are many phenomena that are better explained by minimal theism, (iv) atheism is more metaphysically profligate than theistic naturalism as an explanation, and (v) common atheistic models suffer from various internal problems, one could conclude that some form of theism is more probable.