In 2021, Andrew Moon published a philosophical study, “Circular and Question-Begging Responses to Religious Disagreement and Debunking Arguments,” in Philosophical Studies 178, pp. 785–809, in which he attempts to build on the Christian epistemological ideas of Alvin Plantinga to defend “circular arguments” in defense of certain kinds of (particularly Christian) belief. In my own personal opinion, Plantinga is a truly terrible philosopher and we probably shouldn’t be trying to build frameworks on top of anything he built. But Moon is far better at it and deserves attention. His study is careful and critical in all the ways it should be. It’s merely wrong. But it’s wrong only in the sense that its intended conclusions are unusable in practice. On a charitable reading, his paper’s conclusion is trivially true. But as it is so extremely convoluted (like many an apologetical device ends up being), it can easily be misunderstood or misused as if it said something that, in fact, it did not. So a careful reply is warranted.

The Sitz im Leben

I’ve noted recently how bad most peer-reviewed philosophy is (in Is Science Impossible without God? The Argument of Tomas Bogardus). Moon’s paper is an example of philosophy that is good, in the sense that it warrants passing peer review for discussion and consideration. It is merely wrong. But being wrong is not of itself a basis for denying publication. In all academic fields (even the most rigorous sciences), the job of peer reviewers is not to agree with the submitting authors, but merely to ensure an author is engaging within sufficient standards of argument—in particular, in the field of philosophy, that should mean no obvious fallacies should ever pass, and no arguments dependent on a premise already refuted or contradicted by well-established science (including history) should ever pass.

Moon’s study meets these conditions. Indeed, in footnotes he often breaks the fourth wall by mentioning where changes to his paper were made in light of his peer review reports, so you can see examples of how this process makes a paper not only better, but worthy of publication. Moon’s paper meets another condition such studies are best doing (when applicable): he abundantly cites relevant scholarship, on all sides of the points he makes or relies on, acknowledging where important disputes exist, making his paper invaluable even just for its citations and bibliography. Philosophers (like academics in any field) prize papers that, even though they may find them wanting, nevertheless have use as stepping stones to further discussion of the issues raised through their citations and bibliography. Moon’s paper does a definite service here. It is not a required service, but it is a valuable one.

Moon’s thesis is, in a nutshell, that if (emphasis on if) a Christian (or Muslim or Buddhist or Taoist or Hindu or Osirian…) has formed a Christian (vel al.) belief in a way that is prima facie justified (as in, so far as they know, their belief thus formed is justified), and is then presented with merely a certain kind of so-called de jure objection to their belief (rather than a de facto objection, a distinction I’ll explain shortly, as it is the lynchpin distinction of Moon’s entire argument and where most readers will err in understanding Moon’s argument the most), then they are justified in dismissing that objection as inapplicable to them. The reason is that, if Christianity is true (at least when formulated in certain ways, of course, but Moon means in ways commonly formulated—he isn’t trading on any special pleading), then there will necessarily exist warranted pathways toward justified belief (because, by definition, Christianity, as a hypothesis, entails that God would ensure this; and this is not ad hoc but really is entailed). Therefore, once one comes to believe in Christianity by any way fitting the expectations of the hypothesis (for example, by long-established traditional modes of inspiration by the Holy Spirit), then only de facto objections can warrant abandoning that belief—apart from certain obvious exceptions, like proving the presence of a genuine logical fallacy underlying one’s entire reason to believe, which Moon rules out by stipulation, as he narrows his argument only to cases where such conditions don’t exist (where the Christian is in a state of justified belief prior to encountering a defeater).

Moon’s entire argument trades on a distinction wholly obscure to almost everyone not an expert in modern religious epistemology: de jure vs. de facto “defeaters,” or what colloquially we might call “objections,” reasons to doubt or reject a belief. In light of that distinction, Moon’s conclusion is, I think, wholly trivial. It amounts to saying “once I justifiably get the posterior probability of my belief to be high, I can only reduce it by encountering new evidence.” Which is trivially true. It’s simply a description of evidence-based reasoning, which is to almost no ordinary person controversial.

This is why to understand Moon’s argument correctly, you need to understand what a de jure objection to a belief is. Because that is the only kind of objection that he is saying we can legitimately reject by circular or question-begging arguments (and only in certain limited cases). And the fact of the matter is, in the real world, pretty much no such objections to religious belief really exist—once we limit consideration to those specific circumstances Moon limits his case to. Insofar as there even are any, I suspect they find existence only in the hypothetical fictions of ivory tower musings. Down on the ground, where real beliefs come into contact with real objections, they are almost always (if not always) de facto. Moon might not want to grant this. So this is what I will mainly aim to explain here.

A Quick Summary

Moon describes his thesis, introduces the contextual machinery (“introduction”), sets up some examples (“background”), argues for “a de jure objection to Christian belief that evades Plantinga’s response(s)” to that category, and then proposes a circular argument that “deflects” that specific set of defeaters, and defends that against a slew of objections (as best I could tell, all the best objections one could muster; one of the best merits of Moon’s paper), some of which gets into the eristics of debating “reliability objections” to various methods of forming beliefs (where the question, “Did you form your belief in a reliable way?” is central, including analogous discussions of why we trust our memories and cognitive faculties on what is largely a circular basis, and what would result if what one might call a sensus divinitatis were among those), and then closes with an extended discussion of what to do with religious “disagreements” (such as when a Muslim uses Moon’s entire apparatus to defend his religion to a Christian doing the same), ultimately ending with “standoff” cases (when both parties have to just agree to disagree). All in all, his discussion is pretty thorough, and any objection you might be thinking of as you read it, he probably gets to mentioning himself and answering later on. So if you want to see what a “best defense” of a position like Moon’s looks like, this is it.

I won’t be analyzing most of this because it isn’t really all that relevant to the central question at hand. For example, Moon’s analysis of “disagreement” cases is fine, once you grant what he asked you to in the first half of the paper: that he is only talking about cases where there are no de facto objections being raised but only de jure objections, and only certain specific kinds. That extremely bizarre and rarefied scenario, if it ever even happens (I am skeptical), may indeed play out the way he imagines it should. We just, none of us, are in that scenario (see Misunderstanding the Burden of Proof and Justin Brierley and the Folly of Christianity). But someone who read only Moon’s second half, unaware that he bracketed that to this very narrow and strange circumstance, might mistake him there for arguing that his disagreement protocol applies to any arguments or disagreements with religious believers. Alas, it does not.

The only part of Moon’s thesis that really interests me here is his assumption (never really defended) that what he calls “Freud-Marx” and “CSR” (i.e. Cognitive Science of Religion) “defeaters” are proper de jure objections as he defines it. Because that is the only substantively “real-world relevant” part of his paper. Given that people really do make these arguments against Christian (and Muslim etc.) belief, his conclusion would be a remarkable one and important to epistemology and religious debate. If his argument applies to those in reality, then his conclusion would no longer be trivial. So, does it?

The only definition Moon gives of the “Freud-Marx” argument (as he calls it) is that “religious belief is the result of mere wishful thinking or cognitive dysfunction [and] is thereby unwarranted,” and he never gives an example of this argument actually being made. Which is a common source of error in philosophy: to stay in the clouds of the abstract, rather than (as you should) bringing the question down to earth with a specific (and preferably actual) example; you then build your abstraction out of the particulars, rather than skipping that step and jumping straight into the abstract. This case is illustrative.

To begin with, neither Marx nor Freud used this argument. They never argued religion is necessarily false or should be abandoned. Marx simply described the political reality of religion (e.g. its use by Capitalists to exploit people); he did not say this meant religion was false. Many even a devout Christian believer today says exactly the same things of Christianity that Marx said. And Freud only presented his own psychoanalytic theory of how religion came to exist; that religions were also false (in the forms often presented) is more the fact he uses this model to explain, than a conclusion he used this model to prove.

One can find later authors using the observations of Marx or Freud (or improved, more scientifically updated versions of them) to argue for the falsity of religion, but if you want to analyze them, you need to actually present them. Moon never gives us an example to analyze. The same goes for his CSR examples (and in a footnote, some OTF, or Outsider Test for Faith, examples, though he doesn’t describe them that way, but rather as arguments “according to which one’s religious belief is due merely to factors such as one’s schooling, culture, or upbringing” and is therefore “defective”). He doesn’t, for instance, try to analyze and characterize the arguments of Daniel Dennett (Breaking the Spell: Religion as a Natural Phenomenon (Penguin 2006)) or John Loftus (The Outsider Test for Faith: How to Know Which Religion Is True (Prometheus 2013)), just to pick two top examples.

This is a problem, because no one I know uses these arguments in the form Moon presents. Everyone I know who argues that religious belief is untrustworthy because it is formed unreliably makes an argument to the improbability of that belief. They are therefore not saying, “Regardless of whether the belief is true, it was formed unreliably.” They are saying, “Your belief is not likely to be true, because it was formed unreliably.” This is a de facto defeater. Not a de jure one.

Why Moon’s Solution Is of No Use

Moon opens with his thesis: “I will argue that some religious believers can permissibly wield circular or question-begging arguments in response to certain debunking arguments and religious disagreements.” As I already noted, the real question is not whether such hypothetical circumstances can exist, but whether they ever really do. We’d do better to reword his thesis as “I will argue that in some rare and possibly entirely hypothetical cases, religious believers can permissibly” wield the arguments he describes. And the reason I say this is that he argues at length that his thesis does not apply in response to de facto arguments; it only applies to de jure arguments, and only in a particular circumstance: when someone “is [not] [nor] should be seriously questioning or doubting the trustworthiness” of their belief. Since that describes, so far as I have encountered, zero Christians, it is hard to grasp what use his thesis is. Moon either misunderstands “reliability” objections to belief (whether “Freud-Marx” as he calls it, or CSR; or OTF), or is describing a subvariant of them so rarely voiced as to be trivial to address.

As Moon summarizes the core distinction:

A de facto objection to Christianity concludes that it is false. One could make such an objection without making any reference to a Christian believer or belief. De facto objections are not Plantinga’s focus. A de jure objection concludes that some Christian belief is not justified or warranted. Here, the belief or the believer is the target. Someone who wields a de jure objection but not a de facto objection might say, “Regardless of whether Christianity is true, my Christian friend is not justified in holding her belief!”

This distinction is Plantinga’s brainchild, and it’s somewhat awkwardly snatched from unrelated terms at law. Ironically, contra Moon, Plantinga argued as I do: that de jure objections mostly don’t even exist. Nearly all are really just disguised de facto objections (although “disguised” is the wrong word, because no one made this distinction by which to disguise it but Plantinga; Christians do like to invent problems to harangue). Plantinga got to that conclusion by a slightly different route than I, and reached the wrong conclusions from it (see discussion in the Internet Encyclopedia of Philosophy; and for my response to Plantinga’s “reformed epistemology” generally, see Epistemological End Game). But admittedly there is something odd about Moon trying to defend Christians from arguments even Plantinga thinks don’t exist.

The existence of “true” de jure arguments against religious belief (apart from direct defeaters like “your conclusion was reached by a demonstrable fallacy of logic,” which Moon effectively excludes from his analysis) remains much debated in the field and I have no opinion on that. Whether there are conceptually such arguments is moot in my view, because really all that matters is what’s happening on the ground—what is the actual epistemic environment actual Christians are clashing with. It does not matter if in a massive haystack of de facto objections to Christian belief we can pick out maybe one rare needle of an obscure de jure objection somewhere. I’m fine with getting rid of the needles. I have no need of them. Because once removed, you still have the massive haystack crushing your belief. Who cares about barely-perceived needles?

So at most what we can take away from Moon’s analysis is that (again, apart from direct challenges of logic) we should always formulate objections to religious belief as de facto rather than de jure arguments, including what Moon calls “Freud-Marx” and “CSR” defeaters: rather than say, for example, “you should doubt your religious belief because religious belief can be the result of mere wishful thinking or cognitive dysfunction,” you must say (and demonstrate) that “you should doubt your religious belief because religious belief is usually the result of mere wishful thinking or cognitive dysfunction,” which is a frequency argument—indeed, an argument to a low prior probability. If it is (as a fact) more probable that the lauded “method” of coming to believe in a religion (as Moon describes it) results in a false belief, then P(Christianity|ThatWay) must be low; because more often, by far, “that way” produces false beliefs (Islam, Taoism, etc.). Hence, by contrast, P(~Christianity|ThatWay) is essentially 100% (given that the ratio of all other religions arrived at that way, to 1, is quite large; even more so when competing Christian sects are counted). The resulting likelihood ratio tanks Christian belief (see my Bayesian Argument from Religious Experience). This is simply evidence against Christianity.

Likewise any CSR explanation, and so on: we should formulate every objection as a factual, frequency-indicating argument, not as a hypothetical possibility. Because possibility is not an effective argument; whereas probability always requires appeal to evidence.

I should note that Moon’s argument, even in its only rarefied application, has been challenged in print already by Imran Aijaz, in “The ‘Diderot Objection’ to Plantinga’s Reformed Epistemology,” Religious Studies (2023), pp. 1-24. Though he defers “a complete evaluation of it for some other time,” he argues that his own running counter-example, of a Christian and Muslim arguing and taking Moon’s position against each other, undoes Moon’s conclusion. And his two arguments against Moon are similar to mine (pp. 17–18): first, in the real world, pretty much no one lacks defeaters casting their belief in doubt, so the circumstance Moon needs for epistemically circular argumentation to work doesn’t really exist anywhere; and second, because the exact same conditions “can indeed be met by a reflective Muslim,” which entails anyone in Moon’s suggested “posture seems unable to respond to concerns that his defeater-deflector strategy is meretricious and dogmatic.” Aijaz acknowledges Moon ends his paper with an attempt to get around this problem, but the end result is that he can only do so by continuing to maintain that in such “disagreements” there are (as he has stipulated) no de facto objections on the table, whereas, in fact, there simply always are. So that an epistemically circular argument can legitimately be used to evade only de jure arguments, and only when no other arguments are proffered, is moot. As one might put it: “Who cares?”

You Need to Get Back to Probabilities

A lot of what’s going wrong here is that Moon is messing around with syllogistic reasoning rather than asking more directly the question people actually need to ask: “What is the probability that Christianity is true, given the information being presented?” In other words: hypothetico-deductive reasoning, not deductive syllogistic reasoning (a common trap philosophers fall into: see Why Syllogisms Usually Suck). Once you frame the whole debate as around what the probability is, all these semantic rabbit holes spun out of obscure distinctions like de jure and de facto vanish, and we’re left with the simple fact that things like the “Freud-Marx” and “CSR” objections are actually presenting evidence against Christian belief, and not presenting a mere de jure “defeater.”

This becomes clear when you ask: “How and why does the information presented in this objection change the probability of Christianity being true?” If the objection is framed as “it’s possible that” the Christian has been misled by the processes presented, the Christian has nothing to fear anyway, as possibly does not get you to probably. Moon is thus correct. That’s simply a badly framed argument. But if the objection is framed as “it is probable that” the Christian has been misled by those processes (and frequency evidence is presented establishing that probability), then the Christian is in trouble. And Moon’s argument cannot help them.

Not only can Christians not use Moon’s argument to “circularly” defend themselves against evidentiary arguments (which almost all, if not all, real-world arguments are, even “reliability” arguments), they also can’t use Moon’s argument to bypass fundamental principles of critical thinking, like that you can never be justified in a belief until you try genuinely hard, in some sense and at some stage, to disprove it first and fail (see The Scary Truth about Critical Thinking), because that is the only way to establish a belief has a high probability of being true (for example, this is, essentially, what childhood development accomplishes for our memory and other cognitive faculties). Moon admits his thesis does not apply when a “belief was already unjustified (or unwarranted) to begin with,” and all beliefs that fail at this first principle are already unjustified (or unwarranted) to begin with.

Of course, there is a source-reliability argument one can make here. Most things that we justifiably believe we believe not because we fact-checked them but because we fact-checked the source of them, and we are extending the tested probability of the source’s reliability to the claim’s reliability (at least for the kind of fact-claim in question; since we know the reliability of different sources varies by type of claim). If you didn’t meet the first principle here—if you trust sources you haven’t vetted the expected reliability of—your belief is not justified. Because the frequency of unvetted sources being reliable is low, not high. And that low probability commutes to a source’s claims.

Even “casual” vetting is better than none, e.g. fact-checking weakly rather than strongly: that can justify a belief, so long as the risk of that “casual” approach is being commuted to your confidence in that belief. If your mom, whose reliability in such matters you have vetted (by a long history of living with and observing her), tells you your uncle is reliable, but you haven’t checked, you can believe your uncle is reliable, but you should make contingency plans in case he doesn’t turn out to be so reliable. Low-confidence belief is one of the most fruitful consequences of “proportioning belief to the evidence.” Likewise, risk theory. We have limited resources (time, money, favors, goods), and so we can’t vet everything. So you should devote more of your limited resources to vetting beliefs for which the cost of being wrong about them is high; because you can affordably assign lower confidence to beliefs for which the cost of being wrong about them is low. Religion is always presented as a high-cost belief (even if only in opportunity cost; there is no need of threats of hell here); though that cost diminishes precipitously with its probability: once you establish the claimed supernatural “opportunity cost” of one religion is almost certainly false, you have no remaining reason to believe the claimed supernatural “opportunity cost” of any other religion is true. Once you catch one grifter, you know to doubt them all (see Pascal’s Wager and the Ad Baculum Fallacy).

But across all of this is the same unavoidable principle: the only way to vet the reliability of a source is with evidence. Contrary to Moon’s claim that a Christian “has enough reason” to trust the Holy Spirit once they are convinced it proved Christianity’s truth to them, this bears no analogy to sources generally. I know the Washington Post is reliable by having tested its reliability quite a number of times. Likewise any other source I trust (weakly or strongly, as that’s just a matter of how much testing I did and how effective that testing was, and thus how high my confidence can justifiably be). If we could test the Holy Spirit’s reliability, with enough Bayesian runs to get a high probability it could be trusted in such specific and bizarre ways as the Christian requires, and it tested out as reliable, then a Christian could trust it as much as mom or the Washington Post. But Christians have no such tests, much less have conducted any. Trusting the Holy Spirit (as purported to be experienced) is thus like trusting a dubious stranger in a dodgy neighborhood.

Which is the point of “Freud-Marx”/CSR/OTF arguments: the evidence indicates “Holy Spirits” actually have a very bad track record. The evidence establishes that “Holy Spirit”-style persuaders are very unreliable, not at all reliable. It’s just another Nigerian Prince. This is in fact a failed prediction of the Christian hypothesis. Moon’s argument hangs a lot on the conditional logic that “if” Christianity is true, “then” belief in it is likely to be reliably formed (because God would then make sure of it). This is similar to arguing “if” the Washington Post is a reliable source, “then” I can trust what it said happened yesterday; so when I have no immediate evidence what it said was false, then I do not have to fact-check that to believe it (see A Vital Primer on Media Literacy). In Moon’s case, he argues we can circularly argue that we have a reason to believe our Christian belief was formed reliably once we have a reason to believe our Christian belief is true (which he correctly points out is only epistemic circularity, not logical circularity). But this isn’t true.

We can only base a commutable confidence in a source on evidence of that source’s tested reliability. But lo, contrary to what the Christian hypothesis predicts, the “Holy Spirit” (in its relevantly identical phenomenology) is convincing people to believe in Allah, Buddha, Osiris, the Tao, and indeed spent most of human history convincing people only of other religions, and even still today is busy convincing half the world of “other religions.” It’s 1.9 billion Muslims, 1.2 billion Hindus, and 0.9 billion or so Buddhists and Pagans, for 4 billion “other religions,” to a mere 1.3 billion Catholics, 1 billion Protestants, and 0.3 billion Orthodoxists, for 2.6 billion “Christians,” and I’m glossing over here the head-bashing disagreements across thousands of Christian sects over even rudimentary facts of the faith (just cage-match a Jehovah’s Witness against a Mormon to see what I mean; don’t even get me started on the Rastafarians and Shakers). That evinces the “Holy Spirit” is extremely unreliable and should not be trusted.

Had Christianity been true, there should have been a far more culturally consistent message coming through that spiritual radio across all tribes, peoples, nations, and—most importantly—historical eras. The English should have landed in Australia to find Aborigines, despite no contact with any other peoples for tens of thousands of years, already preaching the Gospel (or something near enough). After all, the Holy Spirit can surely reach Australia. God does not need ships. Had that indeed happened—if the Gospel was sent to all, uniform across all eras and peoples, with only occasional or otherwise minor disagreements—that would indeed have been very good evidence for it being true (and the Spirit a reliable way to know it). In Bayesian terms, absent any better explanation for it, this being the case would increase the probability of Christianity being true a very great deal. But here’s the thing about probability: if that’s being the case would increase the probability of Christianity being true a very great deal, it’s not being the case necessarily must reduce the probability of Christianity being true by exactly that same amount—in other words, a very great deal.

Thus what we see is the failure of the Christian hypothesis, not its success. When Moon says Christianity entails Christianity*, which is the belief that “Christian belief was formed reliably,” he is inadvertently framing the hypothesis of Christianity with a testable predictor: that God would ensure Christian belief was communicated, and in a manner ensuring its justification. Christian belief should therefore be observed to be formed reliably, and in justifiable fashion. As God and his Holy Spirit are global and eternal, there should have been no global constraints on when and where Christian belief arose. That this is not what we observe—that what we observe is what “Freud-Marx”/CSR/OTF predict instead—is evidence against Moon’s Christian hypothesis being true.

This is a de facto defeater of Moon’s religion. It’s just all the worse that it isn’t the only one. Indeed, peculiarly, in a dozen different ways God had to have made our world look exactly like a world with no God in it, which entails the opposite motive for God than the one Moon requires. But the important point here is that this argument is not de jure and thus cannot be escaped by Moon’s device. And I think this can be shown for every claim that religious belief is “irrational” in the sense Moon is exploring (see Theism & Atheism: Doxastic Foundations): these are not de jure objections at all, but de facto ones.

Demonstration

When we say a religious conclusion was reached by logical fallacy, we present evidence of that being the case (if we don’t, we are just saying it’s possible it was, and that is ineffectual as an objection). Moon would agree, if thereby successful, such a defeater cannot be overcome by his proposal. Likewise when we say a religious conclusion was reached by an unreliable method, we present evidence of that being the case (if we don’t, we are just saying it’s possible it was, and that is ineffectual as an objection). The evidence in each case drastically lowers the probability of that religious conclusion from where it was before that evidence was presented, because it removes all the gain in probability that was attributed to the irrational move to begin with (whether that’s a fallacy or dependence on an unreliable method). Moon’s argument has no effect here. His circular approach can only work when these effects of a defeater don’t obtain, like when someone says “your belief is irrational because it could just be emotion/culture/historical happenstance.” Absent any evidence (facts) that this is the case (and not merely “could be”), then, yes, Moon is correct. But that just isn’t what anyone is being challenged with—and insofar as they are, once we remove that challenge, they still have all those other challenges left over, leaving them in no better position than they started.

I do think a better grasp of Bayesian epistemology would have corrected this. Because semantically, when someone says a belief should be abandoned because it was arrived at by an unreliable method (whether fallacy or approach), it does look like saying “regardless of whether that belief is true.” But that isn’t the actual intent of such assertions. Because, apart from exception-cases like disingenuous political rhetoric, no one would say this if everyone agrees the background-established prior probability of the belief is already high. “You should doubt that Richard Carrier is human because he could be a lizard person from another planet” just doesn’t come up in regular discourse for a reason. Even the cranks who actually believe in lizard people give reasons to believe someone is a lizard person. Even lunatics, therefore, understand these kinds of de jure arguments are weak tea—hence they try to invent de facto ones (albeit fallaciously; they are, after all, insane).

In more rational discourse, when someone says a belief should be abandoned because it was arrived at by an unreliable method, they usually take it as understood that, but for reliance on that method, the remaining (prior) probability of the belief being true would be low. Yes, “technically,” they mean this is the case “regardless” of whether the belief is true or false; but they also mean for you to recognize that the remaining background evidence renders that belief probably false. In the absence of any reliable means of arguing Richard Carrier is a lizard person he probably is not a lizard person. Ergo, in the absence of any reliable means of arguing the bizarre theological system of Christianity is true it probably is not. They are not proposing some “all things equal” scenario whereby, once the criticized method is abandoned, the truth of Christianity still remains likely or even just even-steven. That is not their point. By attending to the Bayesian structure of an argument—by asking, “How is this argument meant to lower the probability of the targeted belief?”—you’d catch this feature of it. But if you attend only to surface semantics (as I suspect Moon has done), you’ll miss the entire point of these arguments.

Conclusion

In this analysis I did not touch on every claim, argument, or example in Moon’s paper. There are many there that I consider worth a look; and most are quite sound, even some not fully built-out. So if any readers do read his study and have questions about specific claims or arguments in it that I didn’t address, do mention those in comments (please provide a quote from his paper so I can more easily find what you are asking about; or a timestamp if you view the very useful interview of Moon on this paper that I also watched on The Analytic Christian to make sure I was reading him right, though there Moon makes some claims I think are more dubious). But the overall thesis of Moon’s paper is that any prima facie justified belief is immune to most kinds of de jure objections as long as the “belief” in question contains or entails its own would-be warranter. I agree. As stated, this conclusion is true. It’s just not very exciting. Since all realistic objections to any “otherwise justified” religious belief are de facto, getting to dodge these kinds of de jure objections is not very helpful. It certainly is no rescue for religion.

The general reliability of our memories and faculties is an example Moon returns to a lot: there is a sense in which mere de jure objections to belief in their general reliability (and of course that does not mean absolute reliability, but rather, a “most of the time” conditional reliability) carry no weight, because there is a sense in which we can legitimately appeal to memory and our faculties in an epistemically circular way (which is not a logically circular way): if our memories/faculties were generally reliable then we would be justified in believing they are reliable (in most cases etc.), and so once we get ourselves into any justified belief in their reliability, merely re-citing the de jure objection can carry no effect. In a sense this is just a restating of a simpler principle: we should never believe things without a reason to believe them. Saying “but maybe you were just inserted into the Matrix and all your memories are fake” is not a defeater because you have no reason to believe that has happened.

Moon’s point is that this same reasoning, per Plantinga (and in a sense, IMO, Moon is trying to “fix” Platinga here), applies to belief in Christianity as well. Because if Christianity is true then God would ensure reliable pathways toward justifiably believing it. So merely re-stating the “possibility” that you’ve been deceived in this carries no weight. The only way to dislodge belief like this (whether in one’s memory, faculties, or religion) is by presenting facts that render the belief less probable than you thought before. In other words, you need evidence to dislodge belief. Since hardly anyone really objects to that conclusion, Moon’s conclusion is hardly of any use to anyone.

I think where Moon goes wrong is, indeed, in mis-interpreting (by error or stipulative intent I can’t tell) so-called “error theory” objections to religious beliefs as simply de jure defeaters. Because, actually, they are de facto defeaters. When I say a Christian who relies on a certain historico-geographically contingent psychological process to form a belief in Christianity has a reason to doubt their belief “because” that same set-up causes an abundance of what Christianity entails are false beliefs (i.e. the same historico-geographically contingent psychological process causes belief in quite many a false religion as well), I am not making a de jure objection. I am presenting evidence against the reliability of Christian belief itself. Because Christianity entails “the Holy Spirit” shouldn’t be convincing people to believe in Allah or Osiris or the Tao. Yet that this is what the same process causes is a de facto defeater. Christianity necessarily has to be as likely to be false as every other false religion made this way.

This is an argument made from evidence, evidence that is counter-predicted by Christianity, and that therefore necessarily reduces its posterior probability. Since Moon is not providing any defense against that argument, the argument in that form stands. And so on down the line.

-:-

All my articles here are funded by my highly-appreciated patrons at Patreon, who enjoy various perks for making this content available worldwide for free. But this article in particular has been one of many resulting as well from targeted patron-funded research. With such projects I don’t offer to come to any conclusion, but I do offer to take, on contract, a serious, time-consuming look at peer-reviewed studies and analyze their merits on my blog. So if you have something like that that you want to see me analyze here, contact me with the details and if it’s in my wheelhouse I’ll quote you a rate.