I just finished a draft of a paper with this title, and I could use the expert assistance of Prosblogion members to make it better!
Here's a brief summary of the paper: After describing the two main cognitive science of religion theories for why humans have religious beliefs and arguing that the two can be combined into a more powerful theory, I present an argument that the evidence for these theories gives reasons for thinking that belief in God is unjustified because this evidence shows that our religious belief-forming processes are unreliable. After defending this argument against a handful of objections, I then criticize this argument by arguing that these theories give an unrealistically abstract account of the belief-forming and sustaining processes of religious believers. A more detailed description of these processes reveals that we cannot argue that the belief-forming and sustaining processes of religious beliefs - even when the cognitive science findings are taken into account - are unreliable without previously showing that standard arguments for the existence of God fail. The cognitive science of religion thus presents no independent reason for thinking that belief in God is irrational.
Here is the paper:
CogSciandRelBel.doc
Feel free to post comments here or to email me with comments at joshua.thurow@mtmc.edu.
I think if we would conclude "that our religious belief-forming processes are unreliable", we would also have to conclude that a lot more of our belief-forming processes are unreliable, even if dualism is completely false.
I agree with the conclusion, but I was struck by this sentence:
"evidence shows that our religious belief-forming processes are unreliable"
I immediately thought "Yes, the evidence of the Old Testament, Christ's insistence that 'few' find the gate to salvation, and the early Christian condemnation of the religious practices of 'the world'". It is a basic truth of Christianity that most people (the world) will form unreliable judgments about religion. St. Thomas more or less gives this as a motive for giving revelation! He says that by ourselves, the truths we attain about God are known only to a few, after much time, and with much error. So you could include St. Thomas and Scripture, it seems, along with the cognitive scientists.
But I should read your paper now.
Joshua,
It's an interesting paper, but I had a general worry about the marketability and structure of the paper. It doesn't seem that anyone actually defends the arguments that you are addressing and as you've structured the paper as a response to an argument that no one mentioned in the paper would deny misuses the empirical research. It's difficult to see how it advances any actual discussion. I think with some minor changes to the paper's structure, you can deal with this problem. Either find someone who would defend the relevant arguments or just structure it as a sort of proof that no disturbing results follow from these developments in cognitive science. I take it that it is this second thing that you really are trying to do and it's not as if the paper does a bad job of that.
Joshua, thanks for posting the paper. I think its time that philosophers of religion started to consider the impact of Cognitive Science of Religion. However, as Clayton says, the paper simply demonstrates the failure of an argument that nobody has offered, and that's of limited interest.
I also agree with Clayton that, with the right change of emphasis and structure, the paper would be much more interesting.
One thought that occurs to me is that your paper might be used to demonstrate that reformed epistemology is not enough without a good dose of traditional apologetics - natural theology, historical study, reliance on reports of miracles and exceptional experiences and so on. The attraction of reformed epistemology, it seems to me, is that it purports to demonstrate that most believers, who lack sophistication in such matters, are nevertheless rational in their beliefs.
A supporter of CSR could claim that what they have shown is that the very processes on which Plantinga and Alston encourage the believer to rely are such that they would produce belief in the supernatural irrespective of whether any supernatural entities existed. Certainly, any religious beliefs based on good arguments and strong evidence would still be rational, but only for people who had access to the evidence, and had the time and level of sophistication to engage in such study.
This is an interesting conclusion, because initially, one might have thought that CSR could provide support for Reformed Epistemology. Indeed, I note that Barrett's book is recommended by Nicholas Wolterstorrf. Also, it seems to me, much of the interest in Reformed Epistemology is due to the claim that, if the arguments of Reformed Epistemology are correct, it is perfectly rational to believe in God without basing that belief on good arguments. Perhaps CSR undermines that claim.
Joshua,
I already sent you some comments. Here're some more.
Basically, I'm going to give another argument that the theories you mention give us no good reason to think that the relevant processes which produce theistic belief are unreliable.
Consider Plantinga's point that you can't separate the de jure question from the de facto question. I think that is the case in this discussion.
Suppose God in fact exists. Here is a quote of me summarizing Plantinga's argument from the post here: http://prosblogion.ektopos.com/archives/2008/08/reading-group-w-3.html#more (August 21 10:36am).
Suppose that God exists and is maximally good... Then it is likely that God desires us to love and know him, to be aware of his presence. This point seems plausible to me.
Next: "And if that is so, the natural thing to think is that he created us in such a way that we would come to hold such true beliefs as that there is such a person as God..." (WCB 188-189). That seems to follow from the previous point. Yeah?
Next: "And if that is so, then the natural thing to think is that the cognitive processes that do produce belief in God are aimed by their designer at producing that belief" (189). Check? Then, on Plantinga's account, the theistic beliefs that are present are likely to be warranted.
Okay, not only that, but it seems that if God did design those processes, then they'd be reliable. Furthermore, regardless of which of the three theories you mention which describe the processes produce belief in God, IF God exists, then either God used that process as a reliable means to produce belief in him, or the theory doesn't accurately describe the relevant process which we should be measuring for reliability.
(Clayton and Christian attack this argument in the above discussion I linked, and I defend it.)
So that's another way to attack the argument. I really don't know how much this will help your paper, since it really takes things a different direction.
(Side point. I think the generality problem is a horrible problem for reliabilism. And throughout your paper, I was seeing a number of ways in which the processes could be individuated more generally or not with differing results for reliability, and I could see no nonadhoc way to pick out the relevant process. Anyway, the process I think that matters is the one that the designer intended.)
Ben,
I didn't get to see your comment before I posted mine. But I don't see how the findings of CSR count against reformed epistemology, Plantinga's version anyway (see previous comment). See also below on Alston.
Joshua,
You write in the paper,
"For example, for perception to be reliable, it must be the case that when a certain physical object is not present in normal circumstances, for the most part, one will not believe that the object is present (setting aside other sources of knowledge, for the moment). Contrary to perception, if the target belief of PBP – i.e. some god exists – were false because no gods exist, one would still believe that a god exists."
Like I said in my comments to you, I think you are requiring sensitivity for reliability. But Nozick himself has a counterexample to this claim (as I mentioned in my comments). And I gave another counterexample, "(Also, my old colleague Justin McBrayer, has a paper coming out in Faith and Philosophy where he points out that sensitivity doesn’t work for our reliable, inductive reasoning. I believe that the sun will rise tomorrow. I am using induction based on my past experiences. If the belief were false, I would still believe that the sun will rise tomorrow. I seem to be using a reliable process to form my belief, but the belief is insensitive.)"
(we can suppose that the belief is false because a demon destroys the sun.)
So if God himself designed this process (according to one of the three theories you outline) by which we come to believe in God, and if we would still form belief in God by this process if God didn't exist, this still seems to me to be a genuine case of a reliably produced belief.
Here's another point.
As you pointed out, even if HADD and the processes mentioned by other theories are responsible for our belief in God, we might have other reasons available, and other sources of knowledge. How about religious perception? Alston's argued that this is so; perhaps theistic belief formed on the basis of HADD and the other processes are not sensitive. If the belief were false, we would still believe it.
But that doesn't show that theistic belief might be caused by way of perception (as Alston thinks), and those instances of theistic belief are sensitive because they are caused by God.
But then we don't even need arguments to justifiably believe in God. We need only genuine cases of experiencing or seeing God. Those beliefs would also be sensitive, so your charge that they are unreliably produced would fail.
Thanks for your comments, everyone! I appreciate your comment, Clayton, about the structure of the paper. I definitely intended to do the second thing that you mentioned, and your comment shows me that I need to be clearer that that was my goal. James's comment nicely complements the part of Andrew's comment about perceiving God in Alston. Their comments seem to me to detail two kinds of reasons that theists typically give for their beliefs that must somehow be accounted for in any CSR theory of theistic belief formation/sustainment, and so any argument that attempted to show that those processes are unreliable would have to show that these reasons do not reliably guide us to the truth. This is the thrust of the reason that I give in the paper for thinking that CSR findings do not show religious belief to be unjustified.
Andrew, your alternative route to showing that CSR findings do not show religious belief to be unreliable seems promising. I'll have to think about it a little more. If, after more thought, it seems to work, and I can fit it in the paper in an economic way, I'll be sure to cite you as the source!
Lastly, concerning Andrew's note about the generality problem: I'm not inclined to make much of the generality problem. It seems far more dialectically effective to argue that the argument fails even assuming that the process can be acceptably characterized. That way, my argument that CSR does not show that religious belief is unjustified will not depend upon assuming the controversial claim that the generality problem is a nasty problem.
Dear Joshua and Andrew,
The primary goal of my previous post was to prompt a re-think about the over-all direction of the paper.
Clearly, philosophers of religion need to respond to CRS, to consider how it makes a difference to our projects, in particular, to the project of defending the rationality of Christian belief. Which apologetic projects are now exposed to attack? I'm certainly not trying to advance a strong claim such as that after CRS, Reformed Epistemology is a failed project. I do think that, if Joshua's paper is correct, Reformed Epistemology, (specifically Plantinga's variety of Reformed Epistemology - although I didn't single out Plantinga in my previous post), is weakened more than other forms of apologetics. There are at least tools that the atheist can find in CSR and use against Plantinga.
Plantinga has always maintained that we have within us a sensus divinitatis that will induce us to believe in God when placed in favourable conditions. After all, many people, when placed in favourable conditions, do indeed come to believe in God. What else would cause this but the sensus divinitatis?
Well, after CRS one can advance the claim that what Plantinga calls the sensus divinitatis is simply a fancy Latin name for HADD - or perhaps for HADD operating in conjunction with other faculties that the cognitive scientist of religion will also point out. Furthermore, there is available a plausible naturalistic story about how we acquired HADD, and why it would be that, even if there were no God, HADD would keep throwing up false positives. Before CRS, I suggest, a naturalist could have said something like "Well, maybe one day we'll have a plausible evolutionary theory about why this tendency to believe in God would have developed, even if there weren't a God", now the naturalist can say that we have such a theory: that, I think, is to be regarded as a step forward for naturalism.
Any believer in God, including Plantinga, could respond by pointing to reasons for believing in God that simply cannot be identified with HADD etc: visions, miracles etc. Plantinga has never claimed that anyone's belief in God is purely the result of trusting the sensus divinitatis without any other kind of evidence being involved. Still, I think the interesting question to press is whether building up empirical information about HADD undermines Plantinga's claims about the sensus divinitatis. Focussing on other ways of defending the existence of God side-steps the more interesting question.
In other words, I'm advising narrowing the focus from
(i) Does Cognitive Science Show Belief In God To Be Irrational? to
(ii) Does Cognitive Science Undermine Plantinga's Reformed Epistemology?
I'm not saying that the answer to (ii) is "Yes", but that it is more likely that the answer to (ii) will be "Yes" than that the answer to (i) will be "Yes", therefore it would be a more useful question to pursue.
Hi Josh,
Thanks for the response. Here're some more comments, some of them off the cuff.
Keep in mind that Plantinga has two arguments for the conclusion that if theism is true, then it is likely to be warranted (and hence, reliably produced). The second one is on p. 189 (or maybe 190), and I think it's stronger.
I think an advantage of your argument over Plantinga's is that your response may show that theories of cognitive science provide no threat to religious belief regardless of how we answer the de facto question. I think Plantinga thinks that if God in fact doesn't exist, then belief in God is probably unwarranted for the reason that the belief wouldn't be formed by truth-aimed cognitive faculties. It might still be reliably produced, however, and it's interesting to explore whether that's the case. (Note, however, that on Bergmann's theory of justification, if the belief isn't formed by truth-aimed cognitive faculties, then it isn't justified.)
Let's explore sensitivity and its application to your paper. In Nozick's example, the grandmother sees her grandson visiting her. She believes that her grandson exists. If it were false that her grandson exists, then her grandson would not have visited, but her relatives would have told her that her grandson exists to keep her happy. So she would still believe that her grandson exists. But it seems that the grandmother’s belief that her grandson exists is reliably produced (or is a case of knowledge). Nozick tried to solve this problem by relativizing to methods of belief formation. Sensitivity*: if the belief were false and the person were to use the same method as in the actual world, then the person would not believe. This is supposed to be a necessary condition for the belief's being reliably produced (or knowledge). (I think this is how Nozick relativized it; I'm working off memory.)
Now suppose someone believes in God because of HADD. If it were false and someone were to use the same method as the one used in the actual world, then the person would still believe that God exists. So this belief fails to be sensitive*. So that counts against the claim that it is reliably produced.
But McBrayer's induction example seems to show that sensitivity* is not required for reliable belief formation. In fact, consider any belief that p that is not caused by the fact that p (e.g., my belief that the sun will rise tomorrow). If p were false, I would still believe that p, even if I use the same process of belief formation (say, reliable induction) in the counterfactual circumstance. So from the induction example, what we can learn is that sensitivity* is not a good test for the reliable formation of any belief that p which is not caused by the fact that p (as in induction). This makes sense. If my belief that p is not caused by the fact that p, then if p were false, I am likely to still have the belief that p. This is because there is no causal dependence of my belief that p on the fact that p.
But beliefs formed because of HADD seem to be precisely this sort of belief. The belief that God exists (which is produced by HADD) is not caused by the fact that God exists. And it also fails to be sensitive*. But this isn't reason to think that the belief is produced unreliably. Sensitivity* is a bad test for reliability when the belief that p is not caused by the fact that p.
Overall, I'd take a deeper look at Nozick's stuff and the standard counterexamples to his work.
Hi Ben,
I too am interested in whether the findings of cognitive science cast any doubt on Plantinga's religious epistemology. But from what I read in Josh's paper about the results of cognitive science, I couldn't see any challenge. Or I guess I couldn't see anything that cast into speculation two of Plantinga's main conclusions in Warranted Christian Belief: If God exists, then theistic belief is likely to be warranted. If God does not exist, then theistic belief is not likely to be warranted. Of course, these are both dependent on Plantinga's position that the de jure (epistemic status) question is separate from the de facto (truth) question. (See also above, my first response to Josh, where I outline Plantinga's argument for the first claim.)
Let me engage a little more. You said, "there is available a plausible naturalistic story about how we acquired HADD, and why it would be that, even if there were no God, HADD would keep throwing up false positives."
Suppose we take a positive answer to the de facto question. Suppose that God in fact exists. Then it is false that if there were no God, then HADD would keep throwing up false positives. If there were no God, we wouldn't exist because God wouldn't have created us; no creatures would exist. (Josh deals in his paper with the problem that such a counterfactual would have an impossible antecedent; still, if God exists, it is contingently true that he is the creator of all creatures.)
Suppose we could suppose a negative answer to the de facto question and God does not exist. Then it may in fact be correct that if there were no God, then the HADD would be throwing up false positives. I agree with that.
But this all only shows that the de jure question really depends on the de facto question. Just as Plantinga showed that the Freud-Marx objection all really assume a negative answer to the de facto question, it seems plausible that many of the arguments from cognitive science would do the same.
Andrew,
You say, "beliefs formed because of HADD seem to be precisely this sort of belief. The belief that God exists (which is produced by HADD) is not caused by the fact that God exists. And it also fails to be sensitive*. But this isn't reason to think that the belief is produced unreliably. Sensitivity* is a bad test for reliability when the belief that p is not caused by the fact that p."
The way HADD works is that it attributes agency to a host of specific events. So, if God really is the cause of those events, then God is a cause of my belief, via the operation of HADD, that God exists. This is quite different than in the induction case you mention.
Josh
Hi Joshua,
I'm a little worried about the evolutionary case of HADD, which is, more or less, a just so story, right? But, you say, ". . .humans possess what Barrett calls a hypersensitive agency detection device – HADD (2004, p. 32ff)." But that's a major metaphysical claim, so there really ought to be no better explanation available. You add ". . in virtue of HADD, “people seem to have a strong bias to interpret ambiguous evidence as caused by or being an agent” (2004, p. 31)". But isn't there a better and simpler explanation for the phenomenon of positing agential causes than positing the HAD device? For starters, it is rational (not hypersensitive) to posit agential causes to what appears anomalous. Falling pictures in the middle of the night certainly do not appear to be lawlike events, and neither do lots of other events. We don't need anything like HADD to explain why we find the best explanation for anomalous events to be agents. When things do not appear lawlike, the rational explanation is an agential one. It is actually the hypothesis that makes the best sense for such events. So instead of positing HADD, why not posit rational agents providing the best available explanation?
Mike,
That's a great question. I haven't read the research articles that argue for HADD; all I've read are summaries and brief descriptions of the reasons given in these arguments (which is what Barrett does in his book). So, I can't fully answer your question. However, here are two other pieces of evidence that Barrett mentions: 1) In one experiment, ball berings were made to move around using hidden magnets, and adults tended to attribute agency to the ball berings; 2) people often attribute agency to computers and reason about them as if they were agents. One might well doubt that people really believe that ball berings and computers are agents, but even so the fact that people attribute agency to these things suggests we have some disposition to regard them as agents even though we have lots of evidence to think they aren't agents. If we were rational agents providing the best explanation, as you suggest, it's hard to see why we would still have the disposition to attribute agency to these things. HADD can explain this, though. HADD produces the disposition to regard these things are agents, and this disposition is still present in us, although it is countered by the background evidence we have that these things are not agents, so the disposition doesn't result in full belief.
I don't think these experiments are enough to totally sink your suggestion, though. Maybe there's some other explanation of this data that is consistent with your suggestion. Here's one idea that comes to mind: perhaps people attribute agency to computers and ball berings because it is a useful fiction - it conveys important information about these objects to others (and enables one to remember important information about them, perhaps) in a convenient, harmlessly misleading, way. This is where a greater knowledge of the literature that purportedly supports HADD would be useful; perhaps it contains evidence to rule out this alternate explanation. I'm not sure.
I think, though, that there are good dialectical reasons to simply grant that there is good evidence for the existence of HADD. Then, I can argue that HADD (together with the rest of the story) still doesn't show religious belief to be unjustified. It follows, then, that the justifiability of religious belief isn't held hostage to whether or not the evidence ends up in favor of the existence of HADD.
Hi Andrew, I agree with you that CSR doesn't undermine Plantinga's main claim that if God exists then belief in God is rational. However, it can still strengthen the position of naturalism.
Before CRS, both naturalists and theists could agree that we have some psychological mechanism that induces us to believe in God. Theists had a theory that led them to expect such a mechanism to exist, naturalists did not. At best they could say "Well, the belief-in-God inducing mechanism must result somehow from evolution, perhaps as a by-product of other psychological mechanisms that have survival value." If HADD really is nothing more than a Just-So story, then it doesn't help the naturalists at all - but the claim is that HADD has strong empirical support.
The choice of conditional is important. Joshua addressed the issue of conditionals with impossible antecedents, and that's a topic I'm very interested in. Perhaps though the best conditionals to consider would be the following:
"If naturalism were true, then there would be no reason to expect widespread belief in God", and
"If naturalism were true, then there would be reason to expect widespread belief in God."
Now, if God exists, then it is true that if God did not exist, nothing would exist. But even if that is true, still if naturalism were true, then it would not be true that "If God did exist, then nothing would exist." What we need to be considering is not how things would be if God did not exist, or (perhaps I should say we should not take into consideration that if God did not exist, things would not be), but how things would be if naturalism were true.
Also, I think the claim that HADD often makes us think that there are agents even when no agency is involved. Sensitivity might not be a general requirement for reliability, but if the naturalist can show that sensus divinitatis=HADD, and that HADD is frequently unreliable, that does undermine the credibility of the sensus divinitatis to a certain extent (this doesn't do much harm to Plantinga's case of course: a reliable cognitive system need not be infallible, and often, to say that HADD was giving a false positive would merely beg the de facto question.
Joshua, I think you're right that, for dialectical purposes, it is better to accept claims made about HADD in this paper. I think the issue of whether, if such claims are correct, we can say that HADD=sensus divinitatis is worth pursuing.
Good luck with the paper.