Home » Inference to the best explanation » From Skeptical Theism to Skeptical IBEism

From Skeptical Theism to Skeptical IBEism

Skeptical theism is, roughly, a strategy that employs would-be facts about our cognitive limitations and applies them to various atheological arguments from evil against the existence of God.

Inference to the best explanation (IBE hereafter) is, roughly, the type of inference in which one derives the conclusion that explains the available evidence best.

Skepticism about IBE, is, roughly, the view that the above type of inference is not trustworthy to lead us to truth.

I wonder if accepting skeptical theism puts any pressure on the one who accepts it to also accept skepticism about IBE. That is, should the skeptical theist become a skeptical IBEist? In this blog I’ll try to sketch some flat-footed reasons for thinking so. The literature on both topics is large and complicated, and so I’m really wondering if the below argument warrants further inspection, that is, whether there is even a prima facie push for the skeptical theist to become a skeptical IBEist. Of course, some have argued that skeptical theism implies something like Cartesian skepticism. If that’s true, then skepticism about IBE follows quickly enough. But that’s a strong claim, I’m going for something far more modest—though, as I will suggest, if my worry is real, there will be unpleasant enough consequences, at least for some Skeptical theists.

Skeptical Theism

An atheist says that “If God is good, then he wouldn’t allow any evils to take place that didn’t he have a moral justification for so allowing; but,” she says, “take a look at the fawn caught in forest fire, and who suffers a slow, agonizing death. When we perform a search for a good or a reason that would justify allowing such an evil, we come up blank. Therefore, there is no such justification. Therefore, a good God probably doesn’t exist.” (Call justifications of evils, E-justifications.)

A rejoinder by a skeptical theist might go like this: “You say you’ve searched high and low for a possible justification for evil E. You conclude that, since you didn’t see a justification for E, then there (probably) isn’t one. Normally, this would be a safe inference to make. You take a look in your bedroom and you don’t see an elephant, ‘thus,’ you say, ‘there isn’t an elephant in the room’. However, there are other cases when such an inference is highly problematic. You stand in one spot in the Serengeti, and, after a quick turn of your head in both directions, you say, ‘I don’t see any termite mounds, therefore, there aren’t any termite mounds in the Serengeti at all’. That’s obviously a problem. In general, the problem with the second inference might be put like this: claims like, ‘I can’t see an x’ justifies believing ‘there is no x’ only if we have no good reason to be in doubt about whether we would very likely see an x, if there were one. We could state this principle in various ways, but the basic idea is that if the set of possible justifications God could have for allowing evil E far outstrips the possible justifications you have come up with (and perhaps could come up with) for E, it’s unsafe to move from ‘I don’t see an x’ to ‘there is no x.’ This is a bad ‘noseeum’ inference.”

Basically, the skeptical theist is saying that the atheological argument cannot be rationally compelling unless we assume privilege. Privilege says that for some reason or other we are predisposed to hit reliably upon the right E-justifications and include them in the range under consideration. However, given that God’s intellectual capacities far exceed our own, the range of possible E-justifications that lay within his ken (namely, all of them) would be exceedingly larger than those that lay within our ken. Here’s a picture of the situation (not to scale):


Inference to the Best Explanation

The basic idea behind inference to the best explanation seems simple enough. We have some phenomenon P that we want to explain. So we offer various theories that explain P. There may be several non-identical theories. But the one that best explains P is the one we go with. Moreover, those non-skeptics about IBE will want to claim that a theory’s best explaining P warrants us in accepting the conclusion that the theory is true.

We may note that examples of such reasoning are fairly ubiquitous in our everyday lives. Here’s an example: You wake up in the morning only to find that Tweety is missing from his birdcage. You notice blood and what appears to be a cat’s paw print in the blood. Normally, Delilah is locked up at night, but last night you forgot to put her in her crate. Moreover, when you find Delilah, she’s licking her lips and there are yellow feathers scattered all around her. You conclude Delilah ate Tweety. This inference was not a logical inference. You don’t deduce (deductively or inductively) that your cat ate Tweety. Rather, the theory that Delilah ate Tweety best explains the missing avian, the bloody paw prints, the feathers around Delilah, etc. Moreover, you claim this explanation is true, and what warrants that belief is that it is indeed the best explanation.

Note, the above is not the only theory that explains the evidence. Perhaps your brother, the mad veterinarian scientist, gave your dog, Brutus, a shot of the cognitive enhancing serum he has been working on. Now, it’s a well-known fact that Brutus hates Delilah. Perhaps, in effort to frame Delilah, cognitively enhanced Brutus first drugged Delilah with roofies, then took Tweety out of his cage, killed it, put Delilah’s paw prints in the blood, and spread it around the room. Then, Brutus spread the feathers around Delilah, and, just as you woke up, Delilah came out of her drugged state, yawned, and smacked her lips. This explains all the data too. But we don’t go with this theory, we go with the previous one. Supposing, for simplicity sake, that these two theories are the only ones on offer. We say that the first theory best explains the data. How it does so is beyond the scope of this post. Generally, the idea is that the first theory has certain theoretical virtues that the previous one lacks, and hence isn’t the best explanation, though it is an explanation.

Skepticism about IBE

Despite its ubiquity in everyday life, and in arguments for the truth of scientific theories, some philosophers do not think such an inference is a good or safe one to make. Now, many of these philosophers do not want to dissuade you from reasoning in such a fashion. In fact, there may be rational reasons to select one theory over another based on IBE. The problem arises when we think that IBE warrants us in accepting a theory as true. There are many arguments for skepticism about IBE, below I will briefly present the argument of the bad lot.

Ladyman et al. provide a very succinct statement of the argument of the bad lot. The argument of the bad lot

purports to show that, even if it were in general the case that the best explanation of the evidence is true (or highly probable), that would not suffice by itself to make IBE acceptable as a rule of inference. For, evidently, the potential explanations between which we can choose are the ones we have actually come up with. So to conclude that the best of these is true an additional premise is required, viz., that none of the possible explanations we have failed to come up with is as good as the best of the ones we have. (Ladyman et al., “A Defense of Van Fraassen’s Critique of Abductive Inference: Reply to Psillos,” The Philosophical Quarterly, Vol. 47, No. 188, p.306)

The argument isn’t saying that it is more likely than not that our best theory is the best of a bad lot.  The claim is that our “best” theories may well be the best of a bad lot. The IBE only works in connecting our best (available) explanation to truth if it is more likely that the truth of the matter lies within the range of explanations we are considering. Hence, the argument of the bad lot ferrets out a suppressed premise needed for IBEs to move out of the realm of something like pragmatic justification to the realm of epistemic justification. We have seen the suppressed premise before. It’s privilege. “Hence IBE cannot be rationally compelling unless we assume privilege, that is, that for some reason or other we are predisposed to hit upon the right hypothesis and include it in the range under consideration” (ibid).

We can, again, picture the situation

From Skeptical Theism to Skeptical IBEism

Of course, the above has been much too cursory. Nonetheless, I trust we can see some prima facie similarity. Skeptical theists and skeptical IBEists (at least those who endorse the argument of the bad lot) both point out that the non-skeptics must invoke privilege to make their argument rationally compelling. Moreover, non-skeptics both seem to employ a “noseeum” inference. Atheists will reason from, “I don’t see a E-justification” to “therefore, there (probably) isn’t one;” and IBEists will assume that because they don’t see a better explanation, therefore, there probably isn’t one. That is, both “noseeum” inferences assume privilege.

So, why would a skeptical theist—or, one who accepts IBE as safe reasoning—claim, on the one hand, that the atheist who assumes we are in a place of privilege vis-à-vis E-justifications, is engaging in bad reasoning while, on the other hand, claiming that we are indeed in a place of privilege vis-à-vis explanations of evidence, and thus our reasoning here is truth-apt? Call the conjunction of these two claims privilege disparity or (PD) for short.

The tricky part for the skeptical theist will be to offer an argument for PD that won’t be subject to a “partners in crime” rejoinder? That is, their indictment of skeptical IBEism must not also allow for an indictment of skeptical theism.  We might say that they need to avoid unprivileged parity (UP). So, for example, to argue that not knowing the justification for some evil aids in the practice of faith invites a UP reply, viz., not knowing the real explanation fosters faith. Or, to claim that a large percentage of evils might have justifications so metaphysically detailed that we couldn’t understand them seems to invite a UP response.

It is interesting that Peter van Inwagen (a skeptical theist) is skeptical of the utility IBEs have, at least employed for some philosophical conclusion. In “Impotence and Collateral Damage: One Charge in Van Fraassen’s Indictment of Analytic Metaphysics,” van Inwagen states that he

heartily applaud[s] all that van Fraassen says against those philosophers who ape the practice of scientists—or what they take to be the practice of scientists—by appealing to “the method of inference to the best explanation.” If I had ever thought that there was a method called “inference to the best explanation” that could be used as instrument of philosophical discovery (or which could be used to validate a philosophical theory however it had been discovered), van Fraassen would have convinced me otherwise. But thank God I never have! (Philosophical Topics, Vol. 35, Nos. 1 & 2, Sprint/Fall 2007, p.68)

Here van Inwagen seems consistent with his skeptical theism. And here we might find an argument for PD. van Inwagen’s skeptical theism is, in part, a consequence of what has been called his “modal skepticism.” In his SEP article on skeptical theism, Trent Dougherty explains van Inwagen’s modal skepticism thus,

Modal skepticism can be brought to bear on the argument from evil in the following way.

  1. Our modal intuitions not connected to ordinary life are unreliable.
  2. The inference from inscrutable evils to pointless evils is justified only if certain modal intuitions not related to ordinary life are reliable.
  3. Thus the judgment that there are pointless evils is unjustified.

The argument here applies modal skepticism to our modal intuitions “not connected to ordinary life.” But, it might be thought, our scientific explanations are connected to “ordinary life,” and so they’re reliable. This may explain why van Inwagen thinks that IBEs in philosophical contexts are unsafe inferences, while IBEs in scientific contexts, i.e., “ordinary life” are safe.

However, it is unclear to me how this argument would support PD and avoid UP rejoinders. In support of premise one in the modal skepticism argument above, Dougherty says van Inwagen has pointed out that “our cognitive faculties evolved in a milieu which may underwrite their reliability in ordinary life but which does not extend outside that domain.” But why think the “milieu” in which “our cognitive faculties evolved” selected for scientific realism rather than something like constructive empiricism? Perhaps God, who van Inwagen would say had some hand in the evolutionary process, only thought it worth while for us to select for empirically adequate theories, which doesn’t entail selecting for true theories? To respond, “God is concerned with us knowing the truth,” seems to invite a UP reply vis-á-vis E-justifications. But a more pressing worry is this: a large percentage of scientific theories don’t seem to bear on “ordinary life.” For example, discovering the goings on of the inside of distant, and now non-existent stars, or finding the unobserved causes of obscure phenomena, may certainly be enlightening and interesting, but it seems like van Inwagen’s point would apply here as well: was this epistemic ability evolutionary advantageous? Speculative responses about why God would want us to know the inner workings of stars would seem to be subject to UP responses, and so must take care to avoid them.

Conclusion and Problems and Prospects

In this post, I wondered if a skeptical theist should also be a skeptical IBEist. It’s not clear, presently, that they should—though I hope to have raised a prima facie case that this is a worry—if it is indeed a worry—to look out for. One problem, though, would be that if such a slide from the one skepticism to the other is forced, then skeptical theists would also have to give up on many of the natural theology arguments that make use of IBE. But, such a move has its prospects too. If we think IBE is indeed an unsafe inference, this would also undercut atheological arguments that employ them. In any event, it seems interesting issues lie within the intersection of these problems. Further work may be warranted.


  1. James A Gibson says:

    Suppose the skeptical theist argue that Rowe’s inference is a bad one because we have independent grounds for thinking that there is a morally sufficient reason, even if we cannot cite what that reason is. (Compare: Holmes knows that scene is clearly an instance of murder, although he does not yet know what the murder weapon is nor who the murderer is.) Would your argument be a problem for a skeptical theist like this?

  2. Paul says:

    Hey James,

    I’m not sure. It seems that it should be granted that we have independent grounds for thinking there are an indefinite number of counter theories, T—each of which are mutually incompatible with each other and empirically adequate—to what we currently cite as the ‘best’ theory. They IBEist just says, “Yeah, but we’re in the range of the best of the good lot.” I’m also not sure how the skeptical theist is supposed to argue from your point (which I agree with), to the conclusion: “That’s why we don’t have privilege with respect to E-justifications but we do with respect to mundane or scientific explanations of evidence.”


    1. I think skeptical theists should say something like what you say.

    2. I don’t happen to thinking skeptical IBEism is a problem for skeptical theists :)

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

March 2014
« Feb   Apr »


Follow Analytic Theology, et cetera. on WordPress.com

Get every new post delivered to your Inbox.

Join 33 other followers

%d bloggers like this: