Ed Yong (via Pharyngula) reports on a cool study conducted by psychologist Nicholas Epley:
Epley asked different groups of volunteers to rate their own beliefs about important issues such as abortion, same-sex marriage, affirmative action, the death penalty, the Iraq War, and the legalisation of marijuana. The volunteers also had to speculate about God’s take on these issues, as well as the stances of an “average American”, Bill Gates (a celebrity with relatively unknown beliefs) and George Bush (a celebrity whose positions are well-known).
The result: “In every case, [Epley] found that people’s own attitudes and beliefs matched those they suggested for God more precisely than those they suggested for the other humans.”
Ed says that Epley’s study shows that “relying on a deity to guide one’s decisions and judgments is little more than spiritual sockpuppetry.” (A sockpuppet is a “false identity through which a member of an Internet community speaks with or about himself or herself, pretending to be a different person, like a ventriloquist manipulating a hand puppet” — Wikipedia.)
I can think of at least one other plausible interpretation of this study.
If you believe in God, you probably think God is morally omniscient. That is: you believe that, if any given action X is wrong, then God knows that X is wrong — and conversely, if God believes that any given action X is wrong, then X really is wrong. (You might think God is morally omniscient because you are a theological voluntarist. But even if you deny voluntarism, as many believers do, you probably still think God is morally omniscient, if you believe in God.)
But if you think God is morally omniscient, then you would be irrational if you believe that, say, abortion is wrong (or permissible, or whatever) without thinking that God shares your belief. Given God’s omniscience, a given judgment is correct if and only if God agrees with it. So your endorsement of any given judgment has the immediate implication that God shares your view.
The result is that, if you believe God is morally omniscient, then your moral beliefs also serve as conjectures about God’s attitudes. Thus, in order to explain Epley’s results, we don’t need Yong’s “Sockpuppet Hypothesis,” as I’ll call it. Epley’s results are precisely what we should expect if religious believers consider God to be morally omniscient, regardless of whether religious believers treat God like a ventriloquist treats a dummy.
An analogy can help here. Suppose I think that Dr. Smith, a famous scientist, knows everything there is to know about biology. Then, if I believe that platypuses are not mammals, I should believe that Dr. Smith believes that platypuses are not mammals. (After all, if I believed that Dr. Smith considers platypuses to be mammals, and believed that Dr. Smith knows everything about biology, then I would be crazy to think platypuses are not mammals.) But this doesn’t mean Dr. Smith is my sockpuppet. If Dr. Smith were to tell me that platypuses are mammals, I’d believe him, even if I previously thought otherwise.
We can make similar points if we go deeper into Epley’s study (well — deeper into Ed’s summary of the study, which I consider trustworthy; I haven’t read the study itself). Ed writes:
Of course, correlation doesn’t imply causation – rather than people imprinting their beliefs onto God, it could be that people were using God’s beliefs as a guide to their own. Epley tried to control for that by asking his recruits to talk about their own beliefs first, and then presenting God and the others in a random order. And as better evidence of causality, Epley showed that he could change people’s views on God’s will by manipulating their own beliefs.
[Epley] showed some 145 volunteers a strong argument in favour of affirmative action (it counters workplace biases) and a weak argument opposing it (it raises uncomfortable issues). Others heard a strong argument against (reverse discrimination) and a weak argument for (Britney and Paris agree!). The recruits did concur that the allegedly stronger argument was indeed stronger. Those who read the overall positive propaganda were not only more supportive of affirmative action but more likely to think that God would be in the pro-camp too.
These results do not support the Sockpuppet Hypothesis. If God is morally omniscient, then a convincing argument for (or against) affirmative action is also a convincing argument for the view that God believes affirmative action is a good (or bad) thing. Thus, we should expect precisely these results even if we assume that the Sockpuppet Hypothesis is false.
Epley’s study does contain some evidence for the Sockpuppet Hypothesis. Apparently, when people are thinking about their own moral beliefs, their brains light up in a certain way (as shown by an fMRI scanner). When they are thinking about God’s moral beliefs, their brains light up in a similar pattern. But when they are thinking about other people’s moral beliefs, their brains exhibit a different pattern. This makes it look as if your brain is doing just the same thing whether it is forming moral judgments or speculating about God’s beliefs. This would be hard to explain without the Sockpuppet Hypothesis.
But I don’t think we should put much stock in the Sockpuppet Hypothesis on the basis of fMRI readings alone. I think we just don’t know enough about how brains work; and the inference “Pattern X occurs when doing mental task Y, and Pattern X occurs when doing mental task Z; therefore task Y and task Z are the same task” is obviously too crude. But this does look like an interesting first step toward evidence for the Sockpuppet Hypothesis.
Here are some ideas for further experiments that could yield more evidence about the Sockpuppet Hypothesis:
(1) Find out whether people think God’s beliefs are consistent with the collection of all their moral beliefs. Most people do not regard themselves as morally omniscient — if you ask any given person whether she has any mistaken moral beliefs, she’ll probably say yes (if she gives the question any serious thought). Given this, and given the assumption that God is morally omniscient, you’d expect people to believe that some of their moral beliefs are inconsistent with God’s beliefs. (Of course, they wouldn’t know precisely which of their moral beliefs are inconsistent with God’s beliefs — otherwise they’d abandon those beliefs.) However, if the Sockpuppet Hypothesis is correct, then people might think that their beliefs are fully consistent with God’s beliefs (in the same way that a ventriloquist would expect her dummy to agree with her about everything). So, if religious believers generally believe (and act as if) all of their beliefs are consistent with God’s beliefs, that would be pretty good evidence for the Sockpuppet Hypothesis.
(2) Ask people what the Ideal Observer thinks, and watch what their brains do in ye olde fMRI scanner. The Ideal Observer, like God, is (in some versions) morally omniscient, but the Ideal Observer is a purely hypothetical entity and has no religious role or overtone. Suppose that, when you’re thinking about the Ideal Observer’s beliefs, your brain lights up in a different pattern than when you’re thinking about God’s beliefs. That might be evidence that the assumption of God’s moral omniscience can’t fully explain what’s going on when believers think about God’s attitudes. In that case, you might need to go to something like the Sockpuppet Hypothesis.
Related articles:
Neuroethics: Ethics and the Sciences of the Mind
By Neil Levy, University of Melbourne (December 2008)
Philosophy Compass
Morality and Religion
By Tim Mawson, St Peter’s College (December 2009)
Philosophy Compass
Following God without Belief: Moral Objections to Agnostic Religious Commitment
By Samantha Corte (February 2008)
Philosophy Compass
The Philosopher's Eye 
Very good.
Couple of things. Firstly, a genuine question: How do you suppose that an Ideal Observer is *not* a sockpuppet? I can’t imagine a way of thinking about an ideal observer that doesn’t require thinking about my own beliefs. But that’s because it’s a hypothetical construct, so I don’t feel I’m mischaracterizing anyone if I get it wrong.
As a slightly clumsy segue, this leads to my other query, this one about Dr Smith. Why would a person answer what Dr Smith thinks about the biology of platypodes if they had never heard her pontificate on the subject? If I believed she knew everything about biology, and I believed platypodes were not mammals, I would answer the question about her views as “I don’t know, because she hasn’t told me, but I would guess her answer would be [x], because that’s what I think the answer is, and she wouldn’t be wrong”.
(Actually, I might well not add the last bit and leave it at hazarding a guess… but I wouldn’t make a certain statement about another person’s beliefs without actual – that is, non-inferential[sp?] – evidence.)
That’s what I would personally expect if your hypothesis were true. I’d expect people to answer with guesses, based on themselves, and identified as such – except where they had “sure” knowledge, based on however they believed God to “directly” communicate with them (religious texts, voices in their heads, etc).
Hey Anon,
Thanks for your thoughtful comment!
About whether the Ideal Observer is a sockpuppet:
We might just disagree about what a sockpuppet is. I would have said that X is my sockpuppet if I have (the right kind of) influence over X’s beliefs (or apparent beliefs). So, for example, a ventriloquist’s dummy is the ventriloquist’s sockpuppet because the ventriloquist can make the dummy (appear to) believe anything the ventriloquist likes. On that definition, it’s pretty clear that the Ideal Observer is not my sockpuppet. For example, the Ideal Observer would say that genocide is morally wrong, and there’s nothing I can do to make it say otherwise (although I certain can mistakenly believe it would say otherwise).
We might instead say (as you seem inclined to say) that X is my sockpuppet if my own beliefs provide my only evidence about X’s beliefs. I suspect that this is a less interesting conception of sockpuppetry. Regardless, I would argue that, even on this definition, the Ideal Observer is not my sockpuppet. It does seem to me that any evidence that I might use to form my own beliefs would be evidence that I would use to form my beliefs about the Ideal Observer’s beliefs. So, for example, genocide causes great suffering. This fact is evidence that genocide is wrong; it is also evidence that the Ideal Observer would believe that genocide is wrong. But this does not mean that my belief that genocide is wrong is evidence that the Ideal Observe would believe that genocide is wrong.
A final possibility is that X is my sockpuppet if my own beliefs line up with my beliefs about X’s beliefs. Then, yes, the Ideal Observer is my sockpuppet. But this doesn’t strike me as very interesting.
About Dr. Smith:
I don’t disagree with most of what you say here. But suppose that someone in Epley’s study said this: ““I don’t know what God thinks about affirmative action, because he hasn’t told me and I don’t know of any Bible passages that address the question, but I would guess his answer would be [x], because that’s what I think the answer is, and he wouldn’t be wrong.”
My guess is that Epley would have recorded this person’s response as another case in which the person’s beliefs line up with their beliefs about God’s beliefs. If that’s right, then Epley’s results are fully consistent with the possibility that people reason about God’s beliefs in a way that is precisely parallel with the way you say you would reason about Dr. Smith’s beliefs.