What are theistic responses to Graham Oppy's argument for atheism from naturalism?

Question

Short version

I'm specifically referring to Graham Oppy's paper An Argument for Atheism from Naturalism:

Abstract This paper outlines an argument for atheism from naturalism that I have developed in more detail elsewhere (in particular, in The Best Argument against God). The overall shape of the argument is as follows: first, naturalism is simpler than theism; second, there is no data that naturalism does not explain at least as well as theism; and, third, naturalism entails atheism; so we have good reason to prefer atheism to theism. Note that this statement of the shape of the argument is NOT a statement of the argument itself.

In short, Oppy argues that naturalism is simpler than theism, and that, all else being equal, we should always rationally prefer a simpler explanation of the data.

What are theistic responses to Graham Oppy's argument?

Longer version

A few relevant quotes from the paper:

Theists differ in the ways that they depart from naturalism. Some theists believe in a God who created our universe ex nihilo. Some theists believe in a God whose actions preserve our universe in existence. Some theists believe in a God who inhabits an eternal realm that has no spatiotemporal relation to our universe. Some theists believe in an intelligent and active God who is neither a natural organism nor an artificial intelligence created by natural organisms. Some theists believe in a God that is a non-personal supernatural power or supernatural force that exerts influence on our universe. Some theists believe that the universe possesses the non-natural property of being divine, or that the non-natural property of being divine ‘permeates’ the universe. And so on.

Although theists differ in the ways in which they depart from naturalism, there is a common feature to theistic departures from naturalism. In every case, theists differ from naturalists by believing in something additional: either believing in one or more additional intelligent agents, or believing in one or more additional forces or powers, or believing in one or more additional non-natural properties of the universe.

Suppose that we are comparing a particular version of theism with a particular version of naturalism. Suppose, further, that these versions of theism and naturalism agree in their beliefs about which natural entities, and natural powers, and natural forces, and natural properties, and natural laws there are. In this case, it’s not just that the theist has beliefs in something over and above the things the atheist believes in; it’s also the case that the 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 this case, if all else is no better than equal, then there is clear reason to prefer naturalism to theism. For, if all else is no better than equal, then there is no reason to have the additional theistic beliefs. Hence, in this case, in order to decide between theism and naturalism, we just need to determine whether all else is no better than equal.

...

The burden of the rest of this chapter is to argue that there are no features of the natural universe that have a better explanation on theism than they do on naturalism. Of course, I won’t be able to examine every feature of the natural universe that might be thought to have a better explanation on theism than it does on naturalism. However, I shall try to examine all of the most prominent features of the natural universe that have been widely supposed to have a better explanation on theism than on naturalism. Given the treatment of the cases that I do discuss, it should be obvious how to extend the discussion to features of the natural universe that I do not examine here.

He then goes on to explain how 8 features of the world commonly used to argue for theism can be better accounted for under naturalism. Namely:

• Existence
• Causation
• Fine-Tuning
• Morality
• Consciousness
• Miracles
• Religious Experiences
• Meaning and Purpose

9. Conclusion
As I mentioned at the outset, I cannot claim to have considered all of the data that bears on the decision between theism and naturalism (and not can I claim to have given a fully adequate assessment of any of the data that I have considered). However, I hope that I have done enough to indicate how my argument for naturalism would look if it were set out in full and complete detail. (I give a fuller—but still incomplete—exposition of the argument in The Best Argument against God, Palgrave-Macmillan, 2013.)

10. Note about Evil
Of course, there is data that at least some theists suppose favours naturalism over theism—e.g. data about horrendous suffering, data about non-belief, and data about the scale of our universe. Some naturalists think that data about horrendous suffering is logically inconsistent with theism. As Epicurus argued long ago:

Is God willing to prevent evil but not able? Then he is not omnipotent. Is he able but not willing? Then he is malevolent. Is he both able and willing? Then whence cometh evil? Is he neither able nor willing? Then why call him God?

Other naturalists think that data about horrendous suffering renders theism highly improbable: given the major horrors of the twentieth century alone, isn’t it incredible to suppose that our universe is the work of an omnipotent, omniscient and perfectly good being? I have focussed on data that many theists suppose favour theism over naturalism because my argument requires only that, on any piece of data, naturalism does at least as well as theism in explaining that data. Even if it is true, for example, that naturalism affords a better explanation of horrendous suffering in our universe than is given by theism, that truth makes no contribution to the argument that I have been advancing here.

---

NOTE: Graham Oppy's formulation of the argument is arguably one of the strongest available in the literature, given Oppy's reputation as one of the most respected contemporary atheist philosophers. For instance, William Lane Craig once said about Oppy's book Arguing about Gods:

Oppy's book is not merely recommended but essential reading for anyone interested in natural theology today. No one can pretend to a successful theistic argument unless he has dealt with Oppy's criticisms first. (source)

However, the claim that naturalism is "simpler" than theism is not unique to Oppy, as it is commonly asserted by atheists in general. For instance, take a look at some of the answers to Could Occam's Razor ever favor theism?.

Answer 1

Great answers already! Will add my two cents.

The main answer on the other site (referenced in the comments) shows what I think is the "bedrock" impasse between Oppy's naturalistic theory and a theistic one (specifically Kalaam-style cosmological arguments, emphasis mine):

Fine-Tuning: Oppy makes a fair point here. The naturalist must assert some brute facts, why not simply assert the fine-tuned universe as one of them? That's logically sound, but I can't imagine finding that a more satisfying answer than asserting a wise creator. And again, here we see a fundamental paradox for the atheist: Fine-tuning requires the atheist to either believe in some unimaginably unlikely coincidences or that brute facts of the universe to themselves be complex and interconnected, in which case Occam's razor clearly favors the theist, who makes only one assertion of brute fact: God is.

Having seen Oppy discuss this with philosophers such as William Lane Craig (and others, see bottom of answer), this is the core issue/impasse that motivates other arguments against naturalism being superior.

Just as when we were kids, "Because I said so" or "It just is" is not an intellectually and emotionally satisfying answer - we feel that there is a purposeful obfuscation of underlying motives or rationale taking place. This feeling is underwritten by many examples from our lives where investigation is (prematurely) halted by declaring something as "just is" (because they don't want to reveal some motive, or they don't have a good one because to them it just seems self evidently true).

For example, it used to be thought (by Aristotle in his Metaphysics) that the orbits of the planets were caused by being embedded in spheres composed of a special substance and each sphere was moved by its own deity, who was the "unmoved mover" of each sphere. A (teeny, tiny little bit later ;-), the discoveries and theories of Newton/Kepler/Copernicus posited a new mechanism for the celestial motions (and a new place for earth relative to them!). It did away with the personal gods and replaced it with an impersonal forces -- but not by fiat; one could have easily said that each sphere was simply moved by an eternal, impersonal force and that is that. What made this more satisfying is that the explanation involved a specific causal mechanism (gravitational force) and it's properties (Newton's Law of Gravitation) that allowed for a much richer and accurate description of celestial mechanics.

Had Newton simply said "In my theory celestial mechanics, I take Kepler's laws of planetary motion as necessary facts", his theory wouldn't be very interesting. But the fact that he could account for Kepler's Laws in terms of a more powerful theory (in the sense of explaining strictly more) meant it was demonstrably better, so long as one is willing to pay the cost of taking the Law of Gravitation (plus Newtonian Mechanics) as necessary facts. Almost everyone agreed it was worth it ;-)

Circling back to the cosmological version of this we see that my analogy isn't perfect. Whereas we could aim to corroborate Newton's laws by appeal to further observations in the universe, we aren't afforded that luxury when talking about explanations of the universe. Given our history with causation as humans living on Earth, we have a strong desire to drive mysteries to ground and a presupposition that everything has a cause or some deeper explanation. Of course, to avoid infinite regress, explanations have to stop somewhere (as I frequently remind my 4 year old ;-), especially when we are talking about the most expansive causal arena possible.

When we imagine rewinding the clock and seeing things converge towards one another into an "initial singularity", where time itself stops (T=0), we bump up into a whole slew of intuitions and presuppositions that practically compel us to demand a reason any of this even started.

We say "It can't possibly be just because! It feels so arbitrary, so random (contingent even!). We need something more orderly so we see why it had to be that way."

This is where we hit bedrock. The naturalist can logically say that they take the initial state of the universe as necessary (and in a sense timeless since it defines the point at which "before" makes no sense). The theist will think back to all the instances where we have successfully found further explanations and balk at this "cop out", insisting that a freely-acting God (read: agent) provides a deeper explanation for T=0.

The naturalist can always reply that "Sure, science can advance and we will have a more complete theory that explains why the universe was in the initial state it was, but all that means is that I can move my necessary facts to whatever is assumed necessary in that theory, and my current necessary facts become contingent on that broader theory, which will itself have necessary features."

A theist will say that the infinite potential regress implied by the naturalist's move shows that ultimate first cause/explanation must be personal/agent based so as to provide a necessary fact that itself doesn't presuppose any "free variables". What is doing the metaphysical work here is the appeal to agent-based causation (libertarian free will). By appealing to a non-deterministic but non-random agent with intentions, many people find this a satisfactory explanation, as we can relate to this cosmic agent as beings who also feel like we have freedom to make choices in light of our desires.

Of course, it is debatable if "God is" is truly the sum total of the theist's necessary commitments, since this outwardly simple statement must entail God has the power and intelligence (read: embedded information) to create the initial state that led to our world, a state that, while embodying much intelligence and power, is still not as powerful or intelligent as God (per Anslem). Deny this would be to claim that God is not the most powerful entity, counter to His nature. We can claim God is "simple" in the sense of not being composite, but it is hard to see how god cannot be a complicated being in the sense of possible modes of action and thought -- the only (weak) analog we have is the human brain, whose "information architecture" incredibly complex.

I won't go any further here, as it will take the post into other topics (free will, determinism, randomness, evolutionary psychology, etc) but we can see that if libertarian free will is false or a least is only possible in material things like brains, then appealing to agency is not a viable explanatory extension to naturalism. Even IF libertarian free will were true, we are left asking "Why do I want what I want?" and we are back at the same problem: infinite regress or brute contingency. Libertarian free will supporters will contest that I am missing "agent causation" as a third way out of the above dilemma, but then they need to explain how we differentiate agent causation from randomness without appeal to reasons and motives.

If you liked the above, I highly recommend this a 2+ hour discussion between Drs. Andrew Loke and Graham Oppy on Kalam Cosmological Argument, with a good portion of the time spent on discussing the "satisfactoriness" of taking the initial state at T=0 as necessary. These are two very smart guys and it's very fun to see how philosophy is done by professionals.

A briefer response from Oppy on Kalam to a youtube interviewer.

Answer 2

It is true that simplicity can sometimes be subjective but I think Oppy’s point in this case is that it may in fact be undoubtedly simpler.

Let me illustrate this by way of analogy. People might disagree as to whether or not Newton’s theories may be simpler than Einstein’s theories. However, almost no one would think that the idea that there is a world full of humans is more complex than the idea that there is a world full of humans and animals. This is because the latter includes the former.

Similarly, atleast from an ontology perspective, the universe exists in both theism and naturalism. But God and the “supernatural” only exist in one of these perspectives. So unless you can successfully argue that God is necessary to explain the universe, then atleast from a simplicity standpoint, I don’t see how one can argue against Oppy’s point.

To play devil’s advocate, I suppose as a theist, one could argue that ontology is not the only thing that makes a theory more complex. However, when deciding between two theories and you don’t have any independent evidence to pick between the two, then I’m not sure what else you can go by except this.

Answer 3

I'll only respond to the part of his argument that is at least partly spelled out in the question: the part about naturalism being simpler than theism. He claims that theists believe all of the things that naturalists believe, but that theists also believe something additional as well. However, it is not the case that theists believe everything that naturalists believe. Naturalists need some principle to explain why laws of nature work. Why should a changing electric field cause a magnetic field? Why should there be universal gravitation? Why should light have the speed it does? Naturalists can have no ultimate explanation for any of this. All they can do is point to other laws of nature, but those laws have no explanation until they come up with some other laws of nature to explain those laws of nature. It is an endless regress. Whatever naturalists use to solve this problem, theists don't need that thing.

I'll also add that due to my knowledge of these issues, I feel some extreme skepticism on whether he actually has naturalistic explanations for the things in the list that are better than theistic explanations or that he has any naturalistic explanation at all for existence; however, he at least seems to be a serious scholar worth listening to (as opposed to Dawkins, Dennett, Hitchens, and the like).

Answer 4

On my more theistically-inclined days (they vary), I don't know that I'm often believing in order to explain the laws of physics, say; or the history of the world; or even what I ought to do. There is something almost disrespectful about the idea, as if I believed that someone exists for the sake of explaining other things when I have a different, personal relationship with them anyway. Setting that impiety aside, though, we should note that we can at least ask questions like, "How did God create the world?" and, as we shall see, look for some kind of answers to those, too. So it is not as though there is nothing to explain about God; believing in God does not put an end to the depths of explanation. For example, that God, even God, would by some ultramiracle be redeemed from some dystheism-theoretic sin is something that would call for an explanation; or differentiating between creation ex nihilo and ex possibilio can be explanatory; or describing how one divine being can exist as multiple persons "in full," and in that fullness assume full humanity unto Itself as well, means explaining some things in terms of other things; and so on and on. When Aquinas has the divine persons as, predication-theoretically, relations instead of properties of God, doesn't this in some way "explain how" the concept of the Trinity is possibly logically possible?

Suppose, then, that one took the higher-order properties or relations of the divine nature as axiomatized and then questioned divine incomprehensibility just as well as divine simplicity has been, or infinity (modulo Scotus) might be, etc. Firstly, however, the concept of divine incomprehensibility will be refined into at least a weak and a strong form:

• Weak incomprehensibiity. Some, but not most, factoids about the divine nature are beyond our comprehension, even such comprehension as would emerge from our own natures being transformed as perfectly as possible by God.

• Strong incomprehensibility. Most, if not all, factoids about the divine nature are beyond our comprehension.

That there is such a distinction already in the field can be indicated just by the theory and practice of theodicy. —Now, there is a serious tension between the doctrine of incomprehensibility and that of simplicity. What challenges our intellects are things like having to distinguish substances from accidents or extensions from intensions or so on and on; by contrast, if something was so simple that understanding It would not require those distinctions, then in fact, understanding This would be the simplest thing of all. Our intellect, with its urge towards abstract, absolute unity, with all its talk of ens realissima and the actus purus, would even find the appearance of divine simplicity entirely welcome, as fulfilling its (the intellect's) desire for unified necessity. So the extreme forms of the simplicity doctrine seem possibly inconsistent with an "axiom of (strong) incomprehensibility" (if not a weak one, too).

So on the other hand, if we are navigating between the Scylla of simplicity and the Charybdis of strong incomprehensibility, let us say with hope that our ship lands on the shores of the idea that many of the awe-inspiring works of the divine nature can in some sense be explained. Anselm at his weirdest and the Catechism of the Catholic Church at its most esoteric both make claims that involve an at least partial, metaphysical explanation of the possibility of a concrete doctrine (the Incarnation relative to Anselm, the Resurrection relative to the Catechism). There has also been a slight debate on the philosophers' side of things over the idea that "God ought to exist" + "ought implies can" = "God is possible" = "God is possibly necessary" = "God is necessary" = "God is actual." That dialectic is related to some extent to theories like Rescher's "axiogenesis" thesis, which also involves explaining things like the foundations of the world, in terms of hyperfoundations if necessary.

For there is no need to think that explanations of things "in" God terminate, anymore than we need to think that explanations of things in the world terminate in, say, the initial expansion. We are free to imagine that the question of a time before the initial expansion is not sufficiently grounded; or we are free to find those grounds in the metaphysics of e.g. string theory. Likewise, we can trace our understanding of things on one level of divine representation down to a deeper level, and given that we can't help but understand things coherentistically, then we will be called to keep tracing our understanding downwards regardless of whether we wish to.

Over on the ChristianitySE, one user asked about applying category theory to the doctrine of the Trinity. Let us abstract over this intramural concern and simply(!) ask what might be gleaned about the formulations of other, more abstract questions, by reformulating them in the language of category theory? As if we might take some category Creatrix and ask about initial and terminal objects therein (presumably God would be both initial and terminal, "the first and the last"), or degenerate and trivial cases of creation (creation-by-the-destruction-of-nonexistence and creation-by-default-emanation), etc. How much would this count as "explaining" the divine nature, then, after all?

---

There is also a way in which theoretical simplicity is possibly second-order self-defeating. This is because the topic of theoretical simplicity is, intuitively, "not simple": the SEP article on the topic does run through six major sections (one of which has three major subsections), for instance, and the subtopics themselves involve things as convoluted as the contrary principle of plenitude. Or so to say: the theory of theoretical simplicity is not theoretically simple ("by its own lights"). —And I do wonder what God would think about people debating Its existence on the basis of separate questions about simpler (or not) explanations of other things.

Answer 5

There are two obvious counter-arguments. One is that simpler is not necessarily correct in itself (consider the many well-known examples of simpler theories being supplanted by more complicated ones). The other is that usually, when we are asked to explain why things happen, we never content ourselves with the explanation 'they just did' on the grounds that it is simpler. Vast amounts of human intellectual effort is directed at determining why things are the way they are- we expect there to be reasons. So you might argue that there is an incongruity between that and the view we should assume that the Universe 'just is' because it is simpler than assuming there to have been a cause.

That said, it seems to me that assuming 'god' to be the cause of the Universe is simply using 'god' as a label to represent our ignorance. The truth is that we don't know what, if anything, caused the Universe to be what it is.

Answer 6

This could well be nothing more than a case of poor wording in the abstract, but the claim that "there is no data that naturalism does not explain at least as well as theism" is false.

The related, defensible claim is that the totality of the available data is explained better by a model without a God than by a model with a God.

However, specific data can be better explained by theistic models, even if the theistic models are false. In fact, specific data can be better explained by models which we deliberately design to be false.

Take the model which I just thought of: 1-6-ism. A 1-6-ist believes that dice rolls of 6 are caused by preceding dice rolls of 1.

If I select for my data-set 3 pairs of dice rolls, all of which were 1 followed by 6, the 1-6 model explains the data set perfectly. The "dice rolls are basically random" model doesn't explain it very well at all. To get the true model to show its superiority over the false model, we have to include a larger data set. For a suitably large data set, the totality of available data will better conform to the truer model, even though specific data better conform to the false model.

---

Note - I'm not asserting that the theistic model is false, I'm pointing out that even if it was not just false but preposterous, the claim that "there is no data that naturalism does not explain at least as well as theism" would be false.