What’s Wrong with the World

The men signed of the cross of Christ go gaily in the dark.


What’s Wrong with the World is dedicated to the defense of what remains of Christendom, the civilization made by the men of the Cross of Christ. Athwart two hostile Powers we stand: the Jihad and Liberalism...read more

Faux simplicity arguments against the existence of God

This interview between Esteemed Husband and Dale Tuggy on the Trinities podcast contains a lot of good stuff. (Yes, I know that Tuggy is not a trinitarian. Yes, I acknowledge the irony and weirdness of having a podcast named "Trinities" hosted by a non-trinitarian. No, that isn't relevant to this post nor to the content of this particular podcast.)

I have to admit up front that I have not yet listened to the entire podcast myself but only to the first twenty minutes of it. It takes me a while to listen all the way through podcasts, for some reason.

But a really interesting issue comes up in those first twenty minutes (indeed, in the first seven minutes) that is worth highlighting in itself.

Naturalists will often try to do a kind of crude "subtraction" simplicity argument against the existence of God and for naturalism, and it just stinks as an argument. To be clear, I am saying nothing against simplicity as a criterion of theory choice. It is ubiquitous and important. It is what shows us that conspiracy theories are ridiculous. We tacitly use simplicity considerations constantly.

What Esteemed Husband is speaking out against here is a double abuse of simplicity.

1) Reduce simplicity considerations in theory choice to a crude "subtraction of entities" notion, where "entities" does not include things like thoughts, intentions, actions, and coincidences.

2) Place an irrational amount and type of weight on those crudely defined "simplicity" considerations, so that one treats it as argumentatively legitimate to dismiss or explain away positive evidence for the existence of an entity on the grounds that one's resulting "ontology" is "simpler."

Tim's Abraham Lincoln example here (in the early minutes of the interview) is particularly instructive. Suppose that I take my "ontology of entities" and simply subtract Abraham Lincoln from those entities that I believe exist. But I still have the same evidence for the existence of Abraham Lincoln. Is this reasonable to do? Can I justify the resulting bizarre lengths to which I will have to go to explain away the evidence for the existence of Abraham Lincoln on the grounds that there is one entity the fewer in the world described by these conspiracy theories and hence that they are more probable than the theory that Lincoln existed? Obviously not. "One entity the fewer" doesn't translate into "simpler overall" and certainly doesn't translate into "more probable on the evidence that I have."

In fact, any theory that tries to explain the evidence I have for the existence of Lincoln while denying his existence will be far more complex than the theory that Lincoln existed. In the place of the one man, Lincoln, I will have to hypothesize a large number of intentions to deceive, coincidental confusions and mistakes, etc., on the part of innumerable other people, all converging to appear erroneously to support the existence of Lincoln. A-Lincolnism is by no means a simpler theory, and it certainly is not more probable than Lincolnism! But, if I attribute these deceptions and errors all to persons who really existed (other than Lincoln), and if I don't consider thoughts and intentions to be separate entities in an ontological list, then the "list of ontological entities" is shorter by the absence of one entity--Lincoln. Whoop-de-do.

This is very much what the naturalist is doing. The naturalist/atheist makes a great song and dance about the "greater simplicity" of his atheism, but that is really quite meaningless and pointless in the presence of evidence for the existence of God.

Simplicity considerations are always relevant to our actual knowledge of what exists only by being comparative considerations applied to candidate explanations of existing evidence. None of us sits around in a literal evidential vacuum counting up entities.

The question of whether God exists arises because there is some putative evidence for the existence of God, just as the question of whether Lincoln exists would never have arisen in the first place if there weren't (taken to be) some reason to think that Lincoln exists. (The same argument can be applied to the existence of your mother or your best friend.)

Hence, simplicity should be thought of in actual practice as a three-place (at least) relation: Theory A is simpler than Theory B (or theories B-Z) as an explanation of evidence E, which I possess.

As Tim points out (a very important point), there is no general rule in probability theory that, relative to some body of evidence, (A & B) has lower probability than (A & ~B). People are sometimes confused about this from the fact that (A) alone is strictly more probable than (A & B). But it simply does not follow that (A & ~B) is more probable than (A & B). Asserting the negation of B is itself a theoretical commitment that takes on its own theoretical risk and may be improbable relative to the evidence at hand. (And to cover agnosticism, asserting the metalevel claim "The probability of B is no greater than some ceiling n" is also a commitment that may be wrong, given the actual evidence. It is not the same thing as merely asserting A.)

Agnostics and atheists harp on a faux simplicity argument against the existence of God, and sometimes Christians are thrown by it. I think that they may be thrown because they think that the correct way to proceed is to go back to some hypothetical pre-evidential state (that no real person is ever in), to "award points" to an hypothesis that has fewer entities than another hypothesis, then to move forward to take the evidence on board while always continuing to award these extra points (and they are usually supposed to be large amounts of points) to the hypothesis with the fewer "entities." (Again, with a very restrictive notion of "entities.")

But we would never do this in the case of the existence of Lincoln, your mother, zebras, or any other actual entity. We would never sit around in angst over how many points we should award to the hypothesis that there are no zebras, which would in turn allow us to explain away large swathes of positive evidence for zebras. We would never do that for the existence of Lincoln, either. Any consideration of additional ontological "weight" that we take on with the existence of Lincoln or zebras is obviously wholly swamped by the consideration that we would have to compass land and sea (hence radically violating simplicity norms) in order to do without Lincoln, zebras, or Mom in our ontology while still explaining the evidence.

Whether or not the positive evidence for the existence of God actually works to support God's existence is, of course, a separate matter that has to be examined by actually looking at the evidence. (A point Tim emphasizes repeatedly in the interview.) In the course of that evaluation, naturalists should not be given a free pass to keep running back to some unreal, pre-evidential situation and awarding their atheism special points for not having God in its ontology. Evidence evaluation just doesn't work that way.

Comments (27)

Yes, well said. Simplicity is not equal to rational validity.

"People are sometimes confused about this from the fact that (A) alone is strictly more probable than (A & B). But it simply does not follow that (A & ~B) is more probable than (A & B)."

I didn't know that before, but when I heard it, it made sense to me.

Have you listened to the previous two in Tuggy's podcasts? I've only listened to Paul Moser's and I found his points confusing. He seems too skeptical of natural theology and first cause arguments. To me it seemed odd that he complained that some first cause arguments don't get you to the God worthy of worship, but in regards to most of them, I don't think they are supposed to do that. (I suppose a thomistic first cause argument would, since God would be shown to be pure intellect, pure actuality, etc.) I was just wondering what you thought of his interview, if you've had the time to listen to it.

No, to be honest, I generally avoid listening to podcasts. I find that they are too slow a way to assimilate information. I would be more likely to read transcripts, because I can take in the information so much faster that way. (When I'm interested in one of WLC's Defender's podcasts I almost always wait for the transcript to come out.)

However, I've read the essays by Moser and Oliphint in the 4 Views book (as well as Tim's, of course). It goes without saying that I strongly disagree with Moser. He really has no good argument against natural theology. He is just hostile to it. (And to evidentialism as well.) To be quite frank, his notion of an argument against natural theology is a pulpit-pounding insistence upon a highly controversial (to put it mildly) interpretation of certain biblical texts. That's pretty much it. It's some kind of idiosyncratic, "because I said that St. Paul says so," neo-Barthianism. If you don't think that the Bible says that natural theology is baaad, or you don't consider it self-evident that doing natural theology is prideful, etc., etc., then you aren't going to get anything better than that from Moser right now, as far as I have seen.

Sean, I'm glad you got the point of that section of the post about A & B despite the fact that I had a typo. It is now fixed. What I meant to say was that there is no general rule in probability theory that, relative to some body of evidence, (A & B) has lower probability than (A & ~B). I originally had "higher" instead of "lower" in that sentence. But you seem to have not been thrown by that, fortunately. The following sentence (which you quote) was right.

He did have some good points, I think, about the motivations of a person considering whether God exists or not and about the attitude of one defending God's existence. It's hard to see any benefit coming out of such an investigation if it is just to add another proposition to view of reality. But, yeah, I don't see how this counts against natural theology anymore than it counts against using the Old Testament or historical evidence to show that Jesus is the Christ, which the apostle Paul did.

I didn't see the typo when I read it, but just read it as I recalled your husband put it. This point seems key to me: "It is not the same thing as merely asserting A."

I think your husband mentioned it in the interview, and I've read a number of your posts about unintended coincidences - what is the title of your book on that subject going to be? A friend of mine awhile back gave me a pdf copy of Blunt's book, and I'd be interested in purchasing your's when it comes out. Can you pre-order it on amazon or anywhere else?

I don't see how this counts against natural theology anymore than it counts against using the Old Testament or historical evidence to show that Jesus is the Christ, which the apostle Paul did.


I think most reasoning creatures, those capable of even thinking about the matter, are capable of seeing that if a God exists who created us, this may very well have implications for our lives. If Moser's problem were *just* with the alleged difficulty of getting a "God worthy of worship" out of classical natural theology *alone*, then he should be very interested in the prospects of historical, evidentialist apologetics to defend the existence of the Judeo-Christian God specifically. And there it is even *more* obvious that the conclusion in question would not *merely* add another proposition to a view of reality. Any more than, "You have a loving Father who cares for you and to whom you should be deeply grateful" is *just* another proposition in reality. It's a very special kind of proposition. But unfortunately, Moser is no more welcoming of rational, evidential defenses of *that* richer proposition than of the natural theology defenses of bare theism. So this is a much deeper issue on his part with rational defense in Christian thought generally.

The title of the forthcoming book is Hidden in Plain View: Undesigned Coincidences in the Gospels and Acts. It is not yet available anywhere for pre-order, but I am *hoping* that it will become available for pre-order for the publisher's own site within a week or two. (They are working on that.) Projected release date of the book itself is "spring, 2017." Nothing more precise known right now as to the date.

Would this be a simple and clear example of the same fallacy?

I observe X happen. In "accounting" for X, I can look for some possible cause, or I can posit that it had no cause at all. Because the "no cause" is the simpler hypothesis, that is the "better" account (at least until direct and positive proof X had a cause is demonstrated.)

Interesting question, Tony. Someone *might* try that. What that would amount to is a denial of the Principle of Sufficient Reason, which is knowable to be true a priori. (Or so I believe, anyway.) Such a person is not just trying to avoid postulating an entity. He's trying to say that he doesn't need to postulate any causal explanation *at all*.

The examples I was thinking of involved an attempt to provide an explanation (evolutionary explanation for life, reductive explanations of consciousness, etc.) while keeping God or minds or other things naturalists don't like off one's "entity list" and then triumphantly claiming that one's explanation is "simpler."

I haven't studied the literature concerning the attempted bold rejection of the PSR altogether to see whether simplicity is brought up as an alleged advantage.

I have seen people who try to take issue with the knowability of the PSR, but I don't recall whether I saw them using this method.

My underlying point was that the "simpler hypothesis" is a really difficult to use properly in an argument. In difficult matters, there is almost always disagreement on the exact extent of the matters to be included for which the hypothesis must have some accounting, and disagreement as to whether A + BBBBBbbbbb + C is simpler than Aa' +BBBbbb +Ccc' as they are not directly commensurable. The hypothesis that God causes everything, directly, is the simplest hypothesis possible, (allowing for causality), but THAT's not how those who argue with the principle want it understood.

Alternatively: if you and I see a cardboard box suddenly move 2 inches to the side with nothing outside it moving it, we could posit all sorts of explanations. But whatever explanation we posit (everything from 'little green men' to psychokinesis to the thesis "there is a puppy in there"), all of these are more complicated than the thesis than "the air molecules by random chance all happened to line up with a push on the side of the box". For we know there is air in the box (it is cardboard), and we have no evidence for any of the other hypotheses". Introducing ANY other actor into the equation (before you have more evidence than initially) is forbidden under the "simplest hypothesis"...unless you can also take into account OTHER possible scenarios that haven't been mentioned, such as "my roommate likes to play jokes on me" to "Mom was talking about getting a puppy". But once you can take all those other things into account, there is NO END to the things to consider that affect whether some other hypothesis is "more" or "less" probable, and thus there is no end to conditions that might factor into the relative probability of ANY hypothesis, including the "air only" one. Not to mention the fact that most factors like this cannot be stated to have specific value 0.34 probability or something like that, and even in the rare cases where it does, WE are not usually in a position to calculate it when we use it.

Yes, background evidence always winnows down hypotheses. And from a Bayesian perspective, my perfectly rational probability for some H right now should be the probability of that H on *all* my evidence, including background evidence other than the fact that the box is moving. I do think that there will be simplicity considerations that come in quite ubiquitously, even when taking background evidence into account. For example, if there is a high probability that everyone who has causal access to this space is a rabid dog hater, that obviously means I'll have to add an epicycle or two to the puppy hypothesis in order to account for all the evidence within that theory.

We absolutely cannot do without simplicity. It is in some ways foundational to empirical knowledge. But that doesn't mean that it's easy to use. Moreover, one of the worst things we can do is to reduce simplicity considerations to the "list all the entities in my ontology" version that the naturalist is trying to get so much mileage out of. Indeed, to place that much weight on *that* version would license all sorts of ludicrous conspiracy theories with "fewer entities."

Epicycles! I like epicycles. Ptolemy's use of them was a great save for his "circle" hypothesis. Now, if only he had paid more attention to ellipses in his geometry...

"Indeed, to place that much weight on *that* version would license all sorts of ludicrous conspiracy theories with "fewer entities."

Actually "properly" understood, employing this view of simplicity leads to no such thing. In fact, there can't be any conspiracies. There is only one particular in existence, me, who imagines others exists and that he is a human being who began to exist and also is contingent. Conspiracies, it goes without saying, require a plurality of persons to exist, which of course is not the case. Why adopt this ontology? (You can't, since there is, in truth, no you to adopt it, but ignore that for now.) This ontology explains everything: I'm a necessarily existent being that imagines the existence of an external world.

(If Lawrence Krauss was me, he might say, 'There is nothing, and it is only me! Time to pack up theists; natural theology is doomed. Where is your God? I am you God!')

Don't bother try to disprove it, since anything you can come up with is to complex for me to accept. And besides I'm not interested in learning new propositions about persons (like my friends or family) who might exist. I want something more personal; and that kind of impersonal knowledge isn't powerful enough.

Good question: Would a full, extreme, macho attempt to reduce entities to a minimum in one's ontology at any cost license solipsism?

Well, at least this gives the atheist some epistemological wiggle room to posit an unobservable "multiverse" in reply to the fine tuning argument.

Sage, it's an interesting double standard, isn't it? The positing of the multiverse in that context positively reeks of 5th-wheel ad hocness. Occam's Razor has been thrown out the window so that one can grow a ginormous hipster beard of ontological entities. But when it's a matter of the existence of God, suddenly one's ontology must be so clean-shaven as to risk taking off the chin with the beard.

I seem to recall that Walter Sinnot-Armstrong tried to give some argument as to why that makes sense. Something along the lines of 'well we already know that universes can exist . . .' And 'God is a new kind of entity, but we don't know that he exists.' I can't remember the rest, nor Craig's response, but it was in the first 20 or so pages of "God?" Which the published a few years ago.

what I'll say could be stupid, but in the past scientists made many extravagant and complicated solutions for various problems which turned out to have very simple solutions after involving new entities, but that's maybe not a perfect analogy since God's existence cannot be discovered empiricaly. But, if the multiverse could be inferred from the existence of our universe (we don't know for sure that other universes exist and as I know there are also certain problems with the theory), God could also be inferred from various reasons since his existence is not improbable per se.

As I see it, God would be a more simpler explanation than the multiverse hypothesis regardless being a new entity, but the multiverse has maybe a greater inductive value since we already know that an "entity" like the universe exists.

my question here is, what does simplicity entail? is it more in line with simplicity to entail a new entity if it is the simpler explanation, or proposing a complicated explanation involving just entities from a "closed" ontology? Or does it depends on the particular situation?

my conclusion is: I'm confused😊

Something along the lines of 'well we already know that universes can exist . . .' And 'God is a new kind of entity, but we don't know that he exists.'

If so, that's *extremely* lame. But in any event, postulating a multiverse *merely* to "explain" apparent fine-tuning is blatantly ad hoc. I have a paper on this in Phil.Christi. Oddly enough, though, we have a completely separate criticism of the fine-tuning argument. I also have a paper updating it and suggesting that some portions of the FTA can be construed as in-world events, which avoids our criticism. The whole multiverse-cum-selection-effect thing is extremely poor argumentation, and the literature is deadly dull to boot.

is it more in line with simplicity to entail a new entity if it is the simpler explanation, or proposing a complicated explanation involving just entities from a "closed" ontology?

I would say there is no particular virtue, and certainly no virtue of simplicity, in having a closed ontology or an ontology artificially limited to some particular types of entity. Indeed a closed ontology seems like a fast road to a closed mind.

Look at it this way: Suppose that I argued that it's "simpler" to postulate that there are no zebras in the world but only horses disguised by humans to look like zebras. In order to avoid postulating a new *type* of entity (zebras), I add many different layers of complexity to the behavior of the other types of entities--the humans, for example. I postulate a larger number of horses, plus massively deceptive behavior on the part of humans. It is highly artificial to say that I have thereby obtained a "simpler" answer because I've avoided expanding the types of entities in my ontology to include zebras.

Any idea of giving a free pass to multiplied natural entities (and their highly specific, coincidental properties and/or behavior) on the grounds that we thereby avoid postulating the activity of a supernatural agent and therefore keep the *types* of entities in our ontology to a smaller number seems to me subject to the same criticism. It just is a faux kind of simplicity.

Well, at least this gives the atheist some epistemological wiggle room to posit an unobservable "multiverse" in reply to the fine tuning argument.

Grrrr. The thing that irritates me to no end about multiverse (and many worlds) hypotheses - beyond the incredibly ad-hoc aspect of them, which Lydia is right about - is that they DON'T ACTUALLY ACCOUNT for anything. They are a way of saying, rather, that there isn't an accounting.

Take, for instance the many-worlds notion that at every random quantum junction which may produce A or B, instead of picking one, they say BOTH occur, A happens in one world, B happens in another. But what this amounts to is saying that not only are we refusing to try to account for the apparent randomness that WE actually observe, here in world A, but we insist that THERE IS NO SUCH THING as an account for why A happens in universe A, and B happens in universe B. They are both unaccountable brute facts. It's totally unscientific; it's just a pretty band-aid on being unable to understand what we are seeing here in front of us, a way of saying "I can stop looking for why now." The slacker's notion of science.

The "infinite" varieties of the multiverse, with "infinitely many Hubble universes" aren't any better. They usually posit (mathematically incorrectly, of course) that there will not only be another Hubble universe just like this, there will be infinitely many of them. And other universes for every other set of possible initial values of cosmic constants. But none of that accounts for THIS set of initial conditions of our own cosmic constants having the values they do. That some other universe has a set of values JUST LIKE OURS, or just like except for X value, doesn't say WHY these are present here. It is just a lazy man's cop out: "I would prefer not to be forced to say why, so I make up a fairy story that posits it as a brute fact, and then I don't have to say why because there is no why."

And then they don't like hypotheses with free will!!!!! Go figure!

But what this amounts to is saying that not only are we refusing to try to account for the apparent randomness that WE actually observe, here in world A, but we insist that THERE IS NO SUCH THING as an account for why A happens in universe A, and B happens in universe B.

Well, part of the boring multiverse literature (really, I can't stand the stuff) concerning fine-tuning (I really kind of hate to get into this) is the claim of something called a "selection effect." Namely, that the reason that *we* observe that our universe is life-permitting is that no one could have observed any non-life-permitting universe.

This has many problems. If one is any sort of dualist or believes in the possibility of any non-physical beings, one can deny it outright. Non-embodied angels could have observed a universe that was a single black hole, for example.

Moreover, as I and others have discussed in the literature, one can conditionalize on the fact that we find the life-permitting conditions to be so difficult and narrow *rather than* on the fact that the universe is life-permitting, and this order of conditionalization simply dispenses with the so-called selection effect altogether. It just doesn't arise. That the universe is life-permitting may be something we couldn't have observed (depending on how one defines "we"), but we could *easily* have observed that the conditions for life were less demanding than they are.

All of this just to say that there is a response to this objection when the multiverse advocacy in question is in the context of the fine-tuning argument. But it still isn't very good.

To say nothing, Tony, of the problem that in order to add some plausibility to the hypothesis they would need to specify some set of laws according to which universes were "thrown up" by the multiverse itself. Those laws would, of course, have to be of a sort precisely configured so as to make possible the formation universe we actually observe. So it just pushes the problem back a stage any way you slice it, even if we had the theoretical means for formulating such laws ready to hand, which we don't.

Or the "universe generator" that spits out the component parts. But the *theory* is that it just randomly spits out all different universes, and this is the one we happen to be in. Not a helpful theory. (See above comments.) Also, "universe generator" is science fiction,, not science.

If there is a multiverse, I don't see naturalism and the mutliverse as being the most likely true. Life could be permitted in some universe's, but they'd be in the smallest minority of them. However, if God existed and uses the mutkiverse, you'd expect more life permitting universe's, thus what you find - that we exist in a life permitting universe - is more probable given God and the mutliverse, and should be preferred. Or so goes in abbreviated form what I think is a good, at least interesting, argument by Michael Rota.

Universe generators are science fiction? If diesel generators or real, why not universe generators? You theists have no imagination.

Thank you much for the helpfull clarification.Indeed if we have good reasons (confirmed or not by a kind of evidence) to postulate a new entity then why not? It is at least as plausible as an other "closed ontology" hypothesis, even more. Of course we have no reasons to postulate fairies or something like that over other inductive "closed ontology" solutions, but we do not postulate God without having reasons to do so.

ps. I also recommend Rota's interesting essay.

Life could be permitted in some universe's, but they'd be in the smallest minority of them.

Yes, I made this point in my Phil. Christi article on the topic. The multiverse postulated along with naturalism and the selected effect has this definite property: Life-permitting universes within the set as a whole are very rare. Indeed, that's part of the point. It is supposed to be analogous to my throwing dice over and over and waking someone up only when I finally throw a double six. If he discovered that a double six had been thrown every time and that some other person had been awakened for every other throw, there would still remain a question as to how so many double sixes showed up in the set of throws as a whole. So not just any old multiverse will do. But we certainly do not know (in any independent, non-question-begging way) the properties that a multiverse would have (e.g., rare life) if it existed. But the original move could be *repeated* if we did somehow find out that there were more life-permitting universes out there--that is, simply make the hypothetical size of the multiverse larger than previously thought so that even these are rare. The repeatability of the original inflationary move to "explain" any evidence about properties of other universes makes its *essentially* ad hoc nature quite evident.

Elliott Sober distinguishes between Ockham's Razor of Silence, and Ockham's Razor of Denial. While there is no rule saying P(A&B)https://philreligion.nd.edu/assets/44795/1011lecture.pdf)

Post a comment

Bold Italic Underline Quote

Note: In order to limit duplicate comments, please submit a comment only once. A comment may take a few minutes to appear beneath the article.

Although this site does not actively hold comments for moderation, some comments are automatically held by the blog system. For best results, limit the number of links (including links in your signature line to your own website) to under 3 per comment as all comments with a large number of links will be automatically held. If your comment is held for any reason, please be patient and an author or administrator will approve it. Do not resubmit the same comment as subsequent submissions of the same comment will be held as well.