Here's a hypothetical dialogue between an agnostic and a theist.
A: Evolutionary theory (whether genetic or mimetic or a combination) can explain why you believe in theism. This explanation has nothing to do with the truth of theism. Hence, you should not be a theist.
B: The explanation has much to do with theism. For a good creator would want us to believe in him, and hence, if our beliefs arose through evolutionary means (which you grant!), he would have likely set up the evolutionary pressures in such wise that they should favor belief in theism.
A: You're begging the question by depending on the theism that my argument puts into question.
B: Are you a sceptic about our empirical, logical and mathematical knowledge?
A: I would be really stupid to be such, since scepticism about any of these areas would undercut my belief in evolutionary theory.
B: Good. But now consider this claim: One can give an evolutionary explanation (genetic or mimetic or a combination) of your empirical, logical and mathematical belief.
A: Yes, I can, and I see where you're heading. There is, however, a crucial asymmetry between the theistic case and the empirical, logical and mathematical ones. The explanation of why we hold these empirical, logical and mathematical beliefs depends on the truth of these beliefs. It is useful to believe tigers are dangerous because they are dangerous and the environment is such that true beliefs about dangerous things are useful.
B: But it is useful to believe that God exists because God has set up an environment in which having theistic beliefs is useful for the dissemination of genes or memes.
A: But in your "God has set up" conjunct, you're depending precisely on one of the beliefs put in question by my evolutionary argument.
B: And in the same way in your "the environment is such that" conjunct, you're depending precisely on one of the beliefs put in question by the evolutionary argument for scepticism about empirical, logical and mathematical beliefs.
A: But I can argue a priori that in most environments, having true empirical beliefs is useful.
B: I doubt you can. Maybe in the case of most environments governed by laws like ours you can argue for this, but then you will need a posteriori knowledge of the laws. But the claim that this is so in most logically possible environments? Arguing for that seems nigh impossible. And even if you succeed, what useful thing have you done? The a priori beliefs on which such an argument would be based have been put in question by the evolutionary argument just as the a posteriori ones have. You have failed to show an asymmetry between the theistic case and the empirical, logical and mathematical ones.
A: How about this asymmetry? I should not worry about sceptical arguments that put into question beliefs that all reasonable people should hold, and certain basic empirical, logical and mathematical beliefs are among these.
B: I agree with the principle, but it is not clear that it is of any use. For I think that all reasonable people should either accept theism outright or should accept all the premises of the cosmological argument. It is unreasonable not to do so.
A: You beg the question here.
B: This is exactly what the opponent who doubts basic empirical, logical or mathematical truths will say to you about your assumption that all reasonable people accept them.
A: Let's go back a bit. I see another disanalogy between the two cases. You are claiming that theistic beliefs are selected for because they are true, since the ultimate explanation of the selective pressures involves God's setting them up to produce theistic belief. But here we have an asymmetry between the theistic case and the other cases. For one cannot explain why belief in the dangerousness of tigers was selected for apart from saying that tigers are dangerous. But I can explain why belief in theism was selected for without making use of theism.
B: I can explain why belief in the dangerousness of tigers is selected for without saying that tigers are dangerous. You see, whenever there is a tiger in a given area, it is dangerous to stay in that area. So leaving that area is selected for, and a belief in the dangerousness of tigers leads one to leave the area.
A: Why is it dangerous?
B: Because there is a dangerous animal there.
A: You mean a tiger?
B: Yes, of course.
A: So, you see, you had to make reference to the tigers' dangerousness after all.
B: Yes, but only when pressed for further explanation. I gave you a perfectly good explanation of why it is good to leave the area, because it is dangerous to stay in that area. But this explanation raised a new question for you, namely why the area was dangerous. To this I responded that there was a dangerous animal. This raised a new question for you, namely why there was a dangerous animal. And to this I responded by saying that the animal was a dangerous tiger. Yes, we've arrived at the dangerous tigers, but only by asking further questions. The initial explanation was a perfectly correct explanation--if there is a tiger in an area, it's a dangerous area. But we can ask more and more questions here.
A: I'm losing sight of where this is leading.
B: Sorry. The point is that this is exactly parallel to the theism case. Why is belief in theism selected for? Because it correlates with certain kinds of behaviors which are beneficial, let's say. But now this raises a new question, namely why these behaviors are beneficial? Let's say one answers by citing facts about the environment. But then one can ask why these facts obtain. And eventually, I claim, one will get to a question where the explanation is going to be: "Because God set up the environment so as to select for theistic belief." So there is a parallel. In the case of both the theistic and non-theistic beliefs, a non-ultimate explanation of selectivity can be given without making reference to the truth of the claims. But in both cases, an ultimate explanation makes reference to the truth of the claims.
A: Hmm. How about this response? I can give an alternate ultimate explanation of the universe, an explanation that is non-theistic, and that implies selectivity for theistic beliefs. This story, maybe, involves the Big Bang happening out of a quantum vacuum.
B: I doubt you can plausibly flesh out your alternate ultimate explanation in a way that will stand up to criticism. In fact, the best atheistic responses to the cosmological argument involve denials that there is an ultimate explanation. But suppose you can. I will simply deny this explanation, on the grounds of my theism.
A: There is still an asymmetry. For while I can give a putative non-theistic ultimate explanation of why belief in theism is selective, you cannot give even a putative ultimate explanation of why belief in, say, the dangerousness of lions is selective without talking of the truth of that belief.
B: True, because I am not in fact a sceptic about empirical, logical and mathematical beliefs. But if I were one, I could easily come up with a dozen putative ultimate stories here, involving various demons or the like. These stories wouldn't be true by your lights, but neither is your story true on my lights.
A: Should I thus become a sceptic about empirical, logical and mathematical claims?
B: By no means. But you should abandon your claim that there is an asymmetry between the evolutionary argument for scepticism about theism and the evolutionary argument for scepticism about empirical, logical or mathematical truths. If one succeeds, you have no argument against the success of the other. If it is acceptable to make use of the beliefs put into question to argue that we would be likely to non-coincidentally come to true belief through the evolutionary process in order to refute the evolutionary sceptical argument, then this response will work for both. If one begs the question, so does the other.
A: But maybe this is the asymmetry: Accepting scepticism about empirical, logical and mathematical knowledge would undercut my belief in evolution, and hence would undercut the scepticism. This scepticism is, thus, self-defeating, unlike scepticism about theism.
B: I actually suspect that theism underwrites the scientific method, because the scientific method depends on simplicity being a good guide to truth, and the only plausible reason to think that simplicity is a good guide to truth is to think that a good God created the world, and either created us with doxastic faculties adequate to the world or maybe created a simple world because simplicity is innately good. So evolutionary scepticism about theism would also be self-defeating. But I see you're not impressed. Let me then take a different tack. One could, no doubt, choose a carefully delineated area of empirical knowledge which is irrelevant to your reasons for your beliefs in evolutionary theory--e.g., empirical knowledge about distant galaxies. If the only reason you have for rejecting the sceptical argument is its self-defeat, you will have to accept the sceptical hypothesis about this area. And surely you don't want to do that.
A: No, I see I shouldn't do that.
Hmmm...I'm not sure I entirely understand (despite having read the post twice over), but the theistic position appears to go from "evolutionary explanations of religious belief" to "evolutionary selection of religious belief" to "truth of religious belief." If so, there are problems bridging each gap. Most evolutionary explanations of religious belief are non-adaptationist (e.g., Dan Sperber, Justin Barrett, Pascal Boyer, Scott Atran, though see Joseph Bulbubia and Richard Sosis). If so, then religious belief is not itself adaptive or useful, but it the by-product of adaptive/useful cognitive mechanisms. But even if religious belief was adaptive in itself, this does not mean that the content of religious beliefs is true. To think so is to commit the genetic fallacy, is it not? Or have I missed the point of the post?
Is there a logical inference from 1) I can see a dangerous animal to 2) It will increase your chance of survival if you leave the area?
2) does not logically follow from 1)
So why would God create a world where logically invalid deductions are useful?
'One can give an evolutionary explanation (genetic or mimetic or a combination) of your empirical, logical and mathematical belief.'
I can give a much better explanation of my belief that 3 is the square toot of 9 than a genetic or mimetic one.
Mr. Jong:
I am responding here to the argument that once it is shown that religious beliefs have an evolutionary explanation, then there is no rational reason to hold on to them. (If the conclusion of that argument were that the beliefs are false, it would be an easy to respond to argument--I could just say it commits the genetic fallacy and be done with it.)
My response works just as well regardless of whether the preferred explanation is adaptationist or not. Whatever the evolutionary explanation is, the theist can claim that God set things up so that we would develop belief in God.
I am not arguing for the truth of religious beliefs, or even for the rationality of holding them. I am simply arguing that a certain argument against the rationality of holding them is a bad argument.
Mr. Carr:
If you think human beings are entirely the product of evolution, then I suspect that your explanation of your mathematical belief will depend on things that, on your view, themselves have an evolutionary explanation. Thus, maybe your belief is explained by your possession of a proof from, say, the Peano axioms. (This is unlikely.) But then we can ask about why you accept these axioms. Or maybe your belief is explained by the fact that you take certain laws of arithmetic--like the associative, distributive and commutative laws--on testimony (say, of your math teacher, or the mathematical community). But then what explains your willingness to accept the testimony of apparent experts?
(Note: These questions are about explanation, not justification.)
Hi again Alex,
A said: "Evolutionary theory (whether genetic or mimetic or a combination) can explain why I believe in theism. This explanation has nothing to do with the truth of theism. Hence, you should not be a theist."
A should have said "Evolutionary theory (whether genetic or mimetic or a combination) can explain why I would believe in theism whether it were true or false. Hence, you have an undercutter for your beief in theism even if it were true."
Suppose some object looks blue to you and would look blue to you even if it were another color. Suppose you know this. I think your belief that the object is blue would be unjustified in light of this. No?
Christian:
The reformulation sounds good. Thanks.
Your defense of the reformulated claim, however, is really hard to get working. The analogy you gives requires the truth of the following counterfactual:
(*) If God didn't exist, A (the theist in my dialogue) would still believe in God.
But (*) is true iff theism is false. Argument: If theism is true, then if God didn't exist, A wouldn't exist (since God is the first cause of the whole universe, including A,), and so (*) is false. If theism is false, then the antecedent of (*) is true, and so is the consequent, and so (*) is true. Hence (*) is true iff theism is false. QED
[Now maybe there is some clever way of getting (*) working, by introducing an appropriate context in which we hold fixed A's existence and his evolutionary history, but do not hold fixed the fact that A is a creature of God. But if we can do that, then we can in the same way introduce a context where we can get the following working in a perfectly ordinary case of seeing a red cube in normal light:
(**) If the red cube weren't there, you would still believe it's there.
How to get (**)? Just introduce a counterfactual context in which you keep fixed the observer and the light rays from the red cube. Then (**) is true--in that context. But plainly that's not the relevant way of reading that counterfactual.]
Suppose some object looks blue to you and would look blue to you even if it were another color. Suppose you know this. I think your belief that the object is blue would be unjustified in light of this. No?
No, it wouldn't. Why would I care about the counterfactual information? Suppose the bin is divided into blue items and red items. I'm told that all of the blue items bounce true and all of the red items do not. I choose a ball at random and learn that it appears blue and bounces true. I conclude that it is a genuinely blue ball.You inform me that had someone painted one of the red items blue, it too would have appeared blue and bounced true. I agree, but why would I care about that counterfactual information? I know that all of the blue items bounce true and none of the red one's do. So I know that if it bounces true, it is blue. You need more than the counterfactual information to get a defeater.
'I am responding here to the argument that once it is shown that religious beliefs have an evolutionary explanation, then there is no rational reason to hold on to them.'
Well, do they?
If we are told that 'God exists' by somebody called Karla who has a sterling character, and is an expert on the relevant topic, is that enough to allow us to believe in God?
Or if we found a Book of Mormon, can't we just claim that God set things up so that one day we would have the Book of Mormon?
Why do I take the word of experts that 3 is the square root of 9?
Because I have more than the word of a guy called Karla.
And why did God create a world where theists say our cognitive faculties are reliable when they make one logically invalid deduction after another?
Why didn't God create a world where we could logically deduce what the correct response to seeing a dangerous animal would be, rather than having to make a guess?
Mike:
But supposing that you believe they're blue only because they look blue. Would the counterfactual provide a defeater then? (I bet one can still find a counterexample...)
Mike,
I don't think your example is analogous to the example that I (at least am attempting) to describe.
I said you know that the object would appear blue even if it were not blue.
You gave an example where a randomly selected object appears blue but had another randomly selected object been painted, it would appear blue too. Thus, had the red balls been painted, any ball randomly selected would appear blue.
Is that it? If so, I think that one would have an undercutter in your case if they knew the red balls had been painted. My case is one in which, whatever one bases one's theistic beliefs upon, is such that it has been similarly painted. I took it that this was part of the evolutionary story under consideration. If not...let me add it. (btw...how do you italicize your text?)
Alex,
Let me see if I'm following. First, (*) seems true to me on the supposition about evolution. But why do you think this next claim is true?
"If theism is true, then if God didn't exist, A wouldn't exist."
Since,
(P > (~P > Q)) = ((P & ~P) > Q)
the claim quoted above is trivially true (here '>' is the subjunctive if-then). Is this why you accept it? But then you say this poses a problem for (*) and I don't see it. Anything follows from a contradiction, including the claim that A would not exist and that A would exist. Help?
Christian, on your subjunctive importation/exportation rule above: I/E isn't valid on subjunctives.
I said you know that the object would appear blue even if it were not blue.
In your case I am supposed ot have an undercutter in knowing that the object X that appears blue would appear that way were it not blue. Same in my example. The object X would also appear blue were it not blue, but merely a red item painted blue. So, in both cases, there is no difference in appearence between genuine blue items and non-genuine blue items. They look the same. But that information is nonetheless not an undercutter.
Christian:
If G causes A, and there is no overdetermination or preemption, then: Were there no G, there would be no A. But if theism is true, then God causes A (since God is the creator), and there is no overdetermination or preemption there. Hence, if theism is true, then were there no God, there would be no A.
"If theism is true, then if God didn't exist, A wouldn't exist." Since,
(P > (~P > Q)) = ((P & ~P) > Q)
The subjunctive equivalence above is false. But this one isn't,
([]G > (~G > Q)) -> ((G & ~G) > Q)
If theism is true, then God necessarily exists (says the Anselmian). Let []G be the assumption that theism is true. But then the closest worlds in which God does not exist are also worlds in which he does. But then Q follows trivially.
Jon: Fair enough. I didn't mean to suppose a Lewis/Stalnaker account. I don't think they are correct, but since in most context their account is assumed, I can see why you suggested my rule was invalid. I like suppositional views of subjunctives (Edgington, Quine, Mackie, Dummett, David Barnett). On suppositional views the rule is valid.
Mike: Thanks for the friendly ammendment with the subjunctive. I still disagree with your example though. I don't think we need more than counterfactual information to get a defeater. Consider this dialogue.
A: I think the object I'm looking at is blue because, and only because, it appears to me to be blue. But I trust you B, I know that you are as honest as they come, what do you think?
B: A, for a few reasons that I'm unable to tell you at the moment, due to national security, the object your looking at would appear to you to be blue even if it weren't blue.
A: Oh B, you are so clever. It appears I have an undercutter for my belief that the object I'm looking at is blue.
I guess what I'm saying Mike is build whatever information you think is relevant to the counterfactual information. I don't think we need to have it to have a defeater since A does not and he has a defeater. But suppose I'm wrong. Let me ask a question: Suppose the evolution story provides us "both" with the counterfactual information and the other information you think is important, do you think the evolution story would then provide a defeater for theists who believe in God on the basis of its seeming to them that there is a God (or by having some religious experience, whatever evolution allegedly predicts)? Second question: If you answered yes, then what additional information does the evolutionary theory need, in your view?
Alex: I understand your example. I just don't think your reading is right. In fact, I don't know how to read it at all, it's just weird. I mean we never say things like "If P were the case, then if not-P were the case, then Q would be the case." Insofar as I have any intuitive grasp of this claim, it seems to me to trivially imply Q. But if you want the theory, Mike has a modification above that's follows from a theory. What do you think of it?
Christian:
First of all, I am saying: "If theism holds, then it is true that were there no God, we wouldn't exist." Thus, we have first an indicative conditional and then a subjunctive. In your rendering, you have two subjunctives.
For clarity you may want to read what I say as follows. "Suppose for the sake of argument that theism holds. Then, we can say that were there no God, we wouldn't exist." This is non-trivial, just as the following is non-trivial: "Suppose for the sake of argument that Homer wrote the Odyssey. Then, we can say that were there no Homer, the Odyssey would not have existed."
For a quick example of non-triviality, consider the following: "Suppose for the sake of argument that George committed plagiarism yesterday. Then we can say that were George not to have committed plagiarism yesterday, the moon would have exploded today." This is false. But it has the same logical form.
Christian, this is not yet a defeater. Here's the problem. All that you've been informed about is the existence of a nearby world in which you exist (or your counterpart does) and he is peering at a similar item that appears blue but in fact isn't. This is compatible with your knowing that, in the actual world, there are no such objects or there are very few of them.
So learning that a blue item X I am looking at is such that I could not discern a difference between X and a red item Y, makes no epistemic difference to me unless I do not know the probability of my having a Y. If you tell me that there are such Y's and I know that such Y's are actually extremely rare, then my belief that I am observing a blue item is not undercut. Make sense?
First of all, I am saying: "If theism holds, then it is true that were there no God, we wouldn't exist." Thus, we have first an indicative conditional and then a subjunctive. In your rendering, you have two subjunctives.
Alex,
I think it comes out trivial, if the antecendent (viz., theism, T) entails that God exists necessarily (i.e., []G). Let -> be indicative and > subjunctive. Consider (1) first.
1. (T & []G) -> (~G > ~I)
(1) is trivial, wouldn't you say? I don't see how it could not be. But (since T entails []G) (1) is just equivalent to your claim in (2). So (2) is trivial.
2. T -> (~G > ~I)
Mike:
Makes perfect sense. And now I understand where you were going. Here's what I assume, was assuming...
You don't know how many red items there are, or the probability of drawing a red item. You simply have no information about the numbers and probabilities. It is in this condition that I think you have a defeater. Formally, I would suggest using the principle of indifference to partition the nearby possibilities, but I'm happy to go with the rough-n-ready intuition that one would have a defeater in such a case.
Alex: You're right that I mistakenly formalized you're claim as two subjunctives, my bad. But I still think my point point applies in the same way when read as a subjunctive inside the scope of an indicative. Each of your examples sounds strange to me. I think in every case the consequent is a trivial entailment. Let's take this example:
""Suppose for the sake of argument that Homer wrote the Odyssey. Then, we can say that were there no Homer, the Odyssey would not have existed."
I read this as follows:
Supposing Homer wrote the Odyssey and now, supposing Homer didn't exist, the Odyssey wouldn't have been written.
This is just supposing p and ~p. I think that's what one is doing when considering your example. Here's a general rule I like:
For any conditional (indicative or subjunctive), the antecedent is supposed and the consequent is stated/believed within the scope of the supposition. When a conditional appears to be stated from within the scope of a supposition, in fact, the antecedent of the conditional is added to the supposition and the consequent of the embedded conditional is stated within the scope of the now richer supposition.
Barnett defends this view in "Zif is If" in Mind (this year?). He defends it for subjunctives in a forthcoming paper. I can't do justice to the defenses here, but I agree with his view and that is why I'm reading your examples the way I do. Anyway, I wanted to respond without a theory guiding my response. But for what's worth, your examples just sound strange to my ear, except the last one, and I'm not sure why, but it does sound false (it "seems" that you're right and it is false).
Christian:
I think these examples are counterexamples against this kind of a reading of nested conditionals.
Maybe it will be better to do this in possible worlds. Take a possible world w very much like ours and where Homer authored the Odyssey (in fact, the actual world is one such world, but I do not need this fact). Now consider the truth at w of the following counterfactuals:
(1) Had Homer authored no text, the Odyssey would not exist.
(2) Had Homer authored no text, the Epic of Gilgamesh would not exist.
Now, it seems clear to me that (1) is true at w, but (2) is not true at w.
Here's another way to put it. If C causes E, then barring oddness (overdetermination and preemption), E counterfactually depends on C. If Homer wrote the Odyssey, then the Odyssey counterfactually depends on Homer's being an author. But it is false that if Homer wrote the Odyssey, then the Epic of Gilgamesh (still) does not counterfactually depend on Homer's being an author. Indeed, the Epic of Gilgamesh is counterfactually independent of all facts about Homer (but not necessarily conversely: it is not unlikely Homer was influenced by texts influenced by the Epic). The "counterfactually depend" claim can be unpacked as a subjunctive, and will yield my claims.
But I can side-step the issue. Your claim was, basically, that the counterfactual "Were there no God, B would still believe in God" provides B with an undercutting defeater for B's belief in God. But the counterfactual begs the question against B. For B believes that were there no God, B would not exist, and hence rejects your counterfactual.
(By the way, I've edited some minor slips in my post. This should improve it.)
Mike:
If God exists necessarily, it comes out both trivially and non-trivially. It comes out trivially because the antecedent is necessarily false. It comes out non-trivially because it is a true per impossibile counterfactual.
Compare: "If there were square circles, there would be circles." This is trivially true in that the antecedent is necessarily false. But it is a non-trivially true per impossibile counterfactual.
Alex,
I've never experienced a conversation like the one you describe above. In fact, I've only heard conversations running the other way.
Christian: Well, if there is no God how do you explain the fact that most people believe? [A disguised abductive argument.]
Agnostic: A plausible explanation is that religious sentiments are evolutionary bi-products. Here's a story of how one could explain widespread religious sentiment if God didn't exist . . .
Who's making the arguments you are imagining?
Alex:
The problem (I think) is that though (1) and (2) seem likewise true and false, they get evaluated apart from their context.
Here is their context:
"Take a possible world w very much like ours and where Homer authored the Odyssey (in fact, the actual world is one such world, but I do not need this fact)."
If w is the actual world, then I maintain exactly what I said above. For (1) would simply be the claim that:
(1*) Supposing that Homer authored the Odyssey and supposing that he authored no text, then the Odyssey would not exist.
And I think (1*) makes its consequent trivial.
If w is not the actual world, then we have a subjunctive inside of a subjunctive and I think the same point applies. The problem, I think, is with the assumption that subjunctives state categorical facts about possibilities, whereas I think they state hypothetical facts. The supposition does not refer to a possible world. Thus, subjunctives via a suppositional view makes them hypoethetical.
I didn't follow this part:
"But it is false that if Homer wrote the Odyssey, then the Epic of Gilgamesh (still) does not counterfactually depend on Homer's being an author."
Intuitively, the Epic does "not" counterfactually depend upon Homer's being an author. Was this just a slip with 'not' or am I missing something?
This here is cheap:
"For B believes that were there no God, B would not exist, and hence rejects your counterfactual."
For I know that I exist and it gets Pr(1). Thus, everything entails it. Thus, I could get no information that would count against God's existence for getting that information requires my existence which, if your strategy worked, would never give one a defeater for God's existence. But I can think of many non-trivial empirical defeaters (for some people) for God's existence.
Christian:
I did mean "true" in the Gilgamesh case. Sorry.
If B believes with probability 1 that if there were no God, B wouldn't exist, then indeed B would not (assuming rationality) admit any evidence against the existence of God. But if B believes this with probability less than 1, this is no longer so. (I think we should believe with probability 1 that God exists, but that's a different story.)
Surely you've got to have some way on your account of conditionals of expressing the following idea: "If Bob is indeed Mark's biological father, then Mark's existence counterfactually depends on Bob's." This seems a perfectly acceptable inference from the following universal claim:
(x)(y)(If y is x's biological father, then x's existence counterfactually depends on y).
Something like this has to be right, and is surely not trivial. How do you say that on your account?
Alex: Damn that sounds just fine to me! I'm getting confused now what I think about embedding the subjunctive in the indicative. Thankyou for nailing me on this. I want to maintain that:
If P were the case, then if not-P were the case, then Q would be the case trivially implies Q.
I want to maintain that:
If P is the case, then if not-P is the case, then Q is the case trivially implies Q.
These two claims still seems right to me. But your example seems like a good counterexample to what I was suggesting earlier:
If P is the case, then if not-P were the case, then Q would be the case tivially implies Q.
No, Q isn't trivially implied in your cases. Thanks! Given this I need to rethink the conditions under which some subjunctive claim counts as a defeater.
Christian:
Can you give me a reference for someone who defends a view of counterfactuals that implies these facts about nesting? I suspect that what I say is a clear counterexample to such a view--might be worth publishing.
Thanks for your patience!