Haldane on Dennett
March 11, 2006 — 8:30

Author: Trent Dougherty  Category: Atheism & Agnosticism  Comments: 57

There is a post here containing John Haldane’s brief response to Dennett’s book.
The main thrust is that whatever (surely speculative) story one can come up with about the rise and uses of religion, the bottom line remains, for the philosophically minded person, the truth of religious claims. Swinburne tried to engage him on that too of course, but it seems like Daniel doesn’t want to play.

Comments:
  • anon

    What do you make of Lindsay’s review of the book?
    http://majikthise.typepad.com/majikthise_/2006/03/breaking_the_sp_1.html

    March 11, 2006 — 13:35
  • Lindsay’s review is just OK. She’s entitled to thinking Dennett is right. I’ll state my critiques, while I try to salvage something useful from her review.

    The Truth Problem: she rightly points out that a lot of folks (even religious scholars) claim that religion is a good thing irrespective of its truth or falsity. However, she fails to consider the serious possibility that many religious believers might well have de re beliefs. This is the main thrust of Haldane and Swinburne’s criticisms of Dennett. Dennett is within his rights to conclude that many (or maybe all) religious beliefs are false. But he shouldn’t use this meta-claim as a premise.

    The Reducibility Problem: the problem here is whether the phenomena denoted by ‘religion’ is reducible to naturalistic explanation. Does religion strongly supervene upon human evolutionary history? Lindsay’s review starts getting better when she notices that Dennett takes on a bad assumption by religious scholars. She writes: “Dennett also takes on social scientists and scholars in the humanities who endorse secular religious studies but deny that science could ever explain anything about religion.” The basic problem in religious studies is that religious studies per se is supposed to be a serious and intellectually respectable field of study in the university. It is clearly not a hard science like physics. Is it a social science like psychology? Or is it a humanities field like history? Or should it make normative claims about religious beliefs in the same way philosophy does? Dennett punts on the last disjunct by failing to offer much of argument for his last claim. So if we let Dennett have what he wants, maybe religious studies is a social science after all. But if so, are there any robust kind-terms available for the social scientist of religion? I do not think so.

    One of the implications of Haldane’s review is that any attempt to define the term ‘religion’ in a way suitable for scientific study faces the charge of vacuity. There is a tremendous range of human phenomena that could fall under the term ‘religion’, but none of them have any common denominator. Many religious scholars in the last few decades have concluded that this problem is so serious that it is better to study particular religions rather than religion in general. This problem is analogous to discussions in philosophy of mind over whether psychology is indeed a unified science (see the Kim vs. Fodor debate).

    I think there are at least two different issues. First, we need to know whether religions in particular can explain religion in general. I do not have much confidence that a positive answer can be yielded, based on the inductive evidence. But since this is a modal claim, I am willing to allow that someone might well succeed someday. Second, we need to know whether some religions in particular have naturalistic explanations. Dennett might succeed in doing this for some religions, but the success of this claim does not at all imply the success of his overall project.

    March 11, 2006 — 15:14
  • Bilbo

    Isn’t it just the genetic fallacy to argue that because religious belief historically originated according to some naturalistic scenario, religious belief is not veridical?

    March 11, 2006 — 19:51
  • Lindsay makes the case in her review that Dennett does not commit the genetic fallacy. I’ve not read the book so it is hard to say, but her remarks seem pretty convincing. Clearly I’m speculating, but I’d be surprised if the central parts of Dennett’s discussion rested on such a fallacious inference for two reasons. First, he’s a good philosopher and that kind of gross mistake would be surprising. Second, for him to be guilty of the fallacy, you’d have to think he’d care about the truth of religious beliefs. I suspect he no longer regards questions about the truth of religious belief open.
    Remember, one of the complaints about Dennett is that he’s far too dismissive of religious folk and doesn’t think he needs to engage them by taking seriously the idea they might be right about something. He doesn’t offer his speculative explanation of religious belief to establish its falsity, he takes that to have been established long ago. There seems to be a great deal of discussion of this book and I’d be interested in knowing if anyone’s actually read the thing.

    March 12, 2006 — 15:30
  • Clayton,
    I take it you think she’s made the case in claiming that Dennett’s intent isn’t to undermine religious belief. He’s not out to undermine religious belief, but to break the spell of exempting religious belief from scientific scrutiny. I have two problems with this line of thought. One, despite Dennett’s hand waving that he’s not out to undermine belief with BS he’s hardly credible. Given his previous comments on religion and religious belief everyone would be well within their rights to be suspicious of Dennett’s aims. Two, if Dennett is really out to break some taboo, then I’d like to see some serious evidence that such a taboo exists. It’s hard to imagine he’d write a 450 clunker that simply tilts at windmills. There is a long history of scientific exploration of religion in psychology, and more recently by neuroscientists. We’ve seen studies on religious experience, the soul, life after death, efficacious prayer, and lots of studies within the social sciences on the origins, effects, costs, etc, of religion. I don’t know what the demarcation point is on science, but do meme stories, or memetics, count as doing science?
    It might be better for the defenders of Dennett to argue that the genetic fallacy isn’t always committed when one argues against the origin of a belief. Lots of beliefs may be veridical, but person might not have any epistemic virtues in holding them. The origin of evidence for a belief can be very relevant to the evaluation of the belief. This seems to be the very tack that Dennett takes in his recent CHE essay “Common-Sense Religion”.

    March 12, 2006 — 16:07
  • “…despite Dennett’s hand waving that he’s not out to undermine belief with BS he’s hardly credible.”
    Indeed, he reminds me of Charles Murray’s protestations that the point of the book he wrote with Richard Herrnstein wasn’t that black folk were natively intellectually inferior on average to whites, O no. Dennett does seem more open about putting the onus on religion to justify itself in the face of scientism and establishing an EvPsych story about religion as the story. Since, as you say, there is beaucoup ongoing Wissenschaftliche study of religion, from his proposal that there be scientific study of it one can only conclude he wants a prior concession to positivism and EvPsych, in the hopes that the Universal Acid Rain of what the late S.J. Gould called “Darwinian Fundamentalism” will blight the ground religion grows in.

    March 12, 2006 — 16:37
  • Alan, why not cross-post your comment at Majikthise?

    March 12, 2006 — 16:47
  • The Murray/Hernstein jab is patently unfair. I think the most careful critique of this book is Thomas Sowell’s, and it’s pretty clear to me that someone who has every reason to call those guys racists instead wants to be more honest and take their position to be what they state it to be. He clearly disagrees, but he thinks it’s an intellectual mistake and not a moral one. If it were immoral to get an empirical question wrong, then we’re in trouble.
    What may be different about Dennett is that he has at times made statements not just against religious beliefs but against those who believe them. I wish Mike Rea’s exchange with Dennett were still online, because Dennett really didn’t come out looking very charitable even to philosophers with religious beliefs.

    March 12, 2006 — 17:14
  • I second Matthew’s point about the taboo, since it’s rather difficult to see what the taboo could possibly be. It doesn’t seem to be among social scientists, for the reasons Matthew gives; and given how well works in cognitive science of religion sell, it doesn’t seem to be a general taboo among the populace. Conceivably Dennett has some specific targets in mind; but the first impression he and some of his reviewers give is that of a group arriving halfway through a party and saying, “Hey guys, let’s have a party.”

    March 12, 2006 — 17:30
  • Jeremy,
    Do you mean this exchange?

    March 12, 2006 — 17:31
  • Matthew,
    I’m not going to speculate on Dennett’s motives. I’m only interested in the question as to whether his reasoning is fallacious and in the way in which Dennett has come under attack. I think you’re right that it is quite unlikely that he has much respect for religious folk but of course citing this as a way to dismiss his ideas is just an ad hominem. The specific charge that he commits the genetic fallacy strikes me as implausible given the evidence produced in support of this charge. In fact, Lindsay isn’t the only one who presents a good case for thinking this charge is baseless. As you pointed out elsewhere (earlier post here?), it is far more likely that he is arguing there is an undermining defeater that shows the belief should be rejected because of a lack of evidence that makes probable the truth of the belief. [Does he succeed? Who knows?] I thought that on the standard definition of the genetic fallacy, it is essential that there is an inference from evidence about a belief’s origins to a claim about the falsity of a belief and to defeat by undermining, one can remain neutral with respect to the truth or falsity of a belief’s content.

    March 12, 2006 — 18:13
  • Huh. I guess he just moved it when he moved to Notre Dame. I must have been searching for it right during the time in between. I couldn’t even find a Google cache for it at the time.
    I have a suggestion as to what the taboo is, but it’s really just a wild guess. Perhaps he’s referring to religious relativists who care not a whit about the truth claims of the religions they’re so happy to tolerate without examining why anyone would believe such a thing. Perhaps he’s also referring to people who are fundamentalists of the sort that never critically analyze anything. This is surely a much smaller group than the broad-sweeping use he applies ‘fundamentalist’ to, but it’s a real group nonetheless.
    If that’s right, though, there’s a delicious irony. Most evangelicals are with him in this fight. One argument against interpreting him this way is that many of the people he’s opposing agree with him on this (on this interpretation of his opposition), including the Intelligent Design people. Maybe he’s too busy calling them fundamentalists to realize this (a possibility I wouldn’t dismiss too hastily). Otherwise, I think we may be stuck with a non-referring charge that he’s putting a lot of effort into refuting. Either way, it doesn’t look as if there’s much point in writing several hundred pages on the subject.

    March 12, 2006 — 18:38
  • It might be helpful to discuss some of the specific questions that Dennett thinks that science can help answer.
    *Why is religion a universal feature of human cultures?
    *How do we explain common features of religions that developed independently?
    *Why was there a general trend from folk religions to institutionalized religions? From polytheistic belief systems to monotheistic ones?
    *Why do some religions spread more rapidly than others?
    *How do religions evolve over time? What mechanisms are in place to ensure consistency? What countervailing forces encourage flexibility and dynamism?
    Suppose that human beings are genetically predisposed towards certain kinds of proto-religious ideation. A believer could argue that God designed human beings (through the proximate mechanism of evolution) to have the capacity and disposition to know Him. A scientific student of religion could trace the evolutionary history of these dispositions without considering their veridicality.

    March 12, 2006 — 23:41
  • Lindsay,
    Thanks for joining us. I have a few questions, since you’ve read Dennett’s book.
    First, I assume you have read the Haldane review. In the background of Haldane’s review is a long (since the 19th century) history of attempts to pull off projects like Dennett’s. It’s old hat to scholars in religious studies. Most of those projects are considered failures. They failed because they treated ‘religion’ like a disjunctive predicate. Disjunctive predicates ain’t a property. So they do not denote anything natural. Therefore, religion ain’t natural. But we knew that all along! What is Dennett doing that is new?
    Second, I wonder if the truth problem and the reducibility problem are each sufficently serious for me not to read Dennett’s book when I have other things crying out for my time. I think these worries imply reasonable requirements upon methodology that any entry into philosophy of religion has to address in order to be taken seriously. Since Dennett is an eliminativist about religion as it is safe to assume, I won’t expect him to address the truth problem at length. However, had he offered a significantly improved version of Mackie’s Miracle of Theism, it might have won APA book of the year!
    So does Dennett explicitly address the reducibility problem? Or does he assume the proof is in the pudding? You write that a scientific student of religion can address a number of scientific questions to explain why humans have religious dispositions. Suppose he does show that a science can answer these questions. So what? An eliminativist needs at least one more step. He needs to show that there is some categorical base to the dispositions which are entirely natural. Does Dennett address the rather obvious objection that the categorical basis for a religious disposition may be causal contact with a divine entity? If so, he has to mix it up with the literature in philosophy of religion about religious experience. Does he?
    I would like some answers to these questions so that I know it’s worth my time to read Dennett’s book.

    March 13, 2006 — 0:29
  • Oh, Lindsay, I did not mean to ignore your questions. I assume Dennett tries to address these questions in his book. Each one of your questions is worth a blog thread by itself. I think your first two questions about commonality and universality land closest to my worries that Dennett cannot pull off a reduction.

    March 13, 2006 — 0:36
  • I’d be curious to know if Dennett has read James George Frazer’s The Golden Bough: the Roots of Religion and Folklore. Frazer saw himself as a scientist using evolutionary ideas to examine why religion is universal feature of human cultures, why there is a move from the primitive to the more institutionalized, and how such myths get started in the first place. Frazer was a pretty smart dude who, to his credit, predicted the destruction of many of his ideas. Quite a healthy attitude for a guy who first published his book in 1890.
    I point to Frazer in part to show that suggesting evolutionary/scientific approaches isn’t anything new. I haven’t finished Dennett’s book, but I don’t see what separates his memetic story from that of Frazer, Freud, or any number of social scientists. It’s also unclear how the hard, or special, sciences could even go about answering these questions. I’ll ask again, do meme stories count as science?
    I also bring up Frazer because one of your questions, the idea that folk religions evolve into more complex religions, has been rejected by many people working in sociology of religion. The idea of such an evolution dates back to Frazer’s book. Many modern mythologists now argue that rather than being polytheistic, various animistic deities are simply embodiments of providence in general. If one looked to the body of research in religious studies one would find that many of your questions have been asked an answered. Of course like with many questions these are questions that are under constant evaluation, so much of what you ask is still a part of ongoing research.
    Let me add in religious studies 101, AKA sociology of religion, the first methodological point driven home is that you aren’t considering their veridicality of beliefs. That’s something left to the philosophers.

    March 13, 2006 — 1:17
  • idea that folk religions evolve into more complex religions, has been rejected by many people working in sociology of religion
    It’s going the way of the dodo among biblical scholars as well (and I’m not talking about just the biblical scholars who have anything at stake). Pretty much the whole foundation of Wellhausen’s account of the origins of the biblical materials has been toppled. His view required the material he was calling the priestly material to be very late, with the material he assigned to the Deuteronomist being much earlier, and then the very simplest and most pious accounts without any ritual were the oldest. Most younger scholars are arguing now that the material commonly assigned to the later priests for various reasons has to be earlier than Deuteronomy no matter how you slice it, which makes it at least pre-exilic, whereas Wellhausen had it post-exilic in order to fit with his sociological account. There have been other major developments besides just this, but this is the one relevant to the issue at hand. For reasons apart from the sociology, Wellhausen’s chronological account of the source materials simply doesn’t stand up, which is further support besides simple sociology that the development of religion (particularly Hebrew religion) didn’t take the path from folk religion to a more institutionalized religion, at least not during the time the materials recorded in the biblical texts were coming together. I would say that most biblical scholars under 50 would agree with this.

    March 13, 2006 — 7:36
  • Brian Leiter has posted a cheer for Dennett on this point, urging that the origins of a belief are sufficent reason for it to be unjustified. It seem now that the genetic fallacy has become a virtue, at least in some in cases! We have already discussed the implausibility of Dennett’s own genealogical story in its own terms. Perhaps now we should examine whether it meets reasonable standards for justification.
    On a reliabilist notion of justification, what counts is whether the belief-forming process is truth-conducive. Perhaps the origin of a belief may contaminate the belief-forming process, or perhaps it does not. Alston suggested a number of years ago that a direct theory of reference may account best for how folks refer to God, and do a better job of explaining the tremendous diversity in the descriptions of what God is light. The causal theory of reference developed in part because counterexamples to the descriptive theory reference showed that one could successfully refer to objects without needing a correct description (Donnellan 1966). So Kripke and Putnam concluded that descriptions were neither necessary nor sufficent for successful reference. Analogously reference to God, if one applies the causal theory, need not require a correct description. One could perfectly well refer to God while describing God as a law professor from Austin Texas.
    I am on shakier grounds with an internalist notion of justification, but as I understand internalists, a belief is justified if and only if its epistemic status supervenes upon other beliefs lending support to this. I think internalists owe a huge explanation as to what the support relation is really about. But I think this problem could be set aside, and we can focus upon evidentialism as one sort of internalism. On evidentialist grounds, one could consider the wierd origins of a belief an underminer. But it would be a far stronger claim to consider it a defeater. It’s not clear to me how the origin of a belief constitutes a negative argument against the belief without begging the question. Perhaps Dennett, Leiter and his compatriots want to run a queerness argument against theism. However, I would urge a huge caution at this point, as those arguments are confessions of faith rather than real arguments.
    I would appreciate it if the epistemologists especially of the internalist variety would speak up and point out problems I might be having.

    March 14, 2006 — 20:23
  • Is it wrong to conclude that Leiter has now referred to anyone who holds to reliabilism and a causal theory of reference as “(not very good) philosophers”?

    March 15, 2006 — 6:46
  • Uh, Jeremy, I think BL really stuck his foot in his mouth with that one. I suspect Nietzsche’s genealogical method has run amok. Alston holds to the causal theory of reference and accepts a certain amount of reliabilism. I suspect Alvin Goldman would have the same sorts of commitments, because it’s really hard for me to see how reliabilism and descriptivism would work together.

    March 15, 2006 — 13:40
  • Gordon Barnes

    I want to make a couple of points. First, John Haldane overstates his case when he says that there is no definition of “religion,” because there is nothing that all religions have in common. On the contrary, here is something that all religions have in common: belief in invisible, intelligent agents who cause some of what happens in the world. All religions involve belief in entities of this kind, and there is no belief system that involves a belief in such an entity that would not thereby qualify as “religion.” So I think that Haldane is just wrong here.
    Second, there is a very plausible epistemic principle that entails the relevance of the origins of religious belief for the rational justification of religious belief. It is what I call The Principle of Explanatory Defeat: If E is my only evidence for a belief B, then if I discover that the best overall explanation of evidence E does not imply that B is true, then E is defeated as evidence for B. For example, suppose I am looking at an object that appears red to me, and I form the belief that it is red. My only evidence for my belief is that the object appears red to me. Now suppose I am told that the object is being illuminated by red light. Then the best explanation of my evidence, which is the appearance of redness, fails to imply the truth of my belief that the object is red, and for that very reason, my evidence for my belief is defeated. Now, I believe that, on a charitable interpretation, the scientific debunker of religion is best interpreted as saying that the only evidence that people have for religious beliefs are their religious experiences, but these experiences are best explained by scientific theories of religion that do not imply the truth of these beliefs. Therefore people’s religious experiences are defeated as evidence for religious belief. Obviously this assumes that natural theology fails, and that’s another matter, but I think that this line of argument demonstrates the relevance of scientific theories of religion to the rational justification of religious belief.

    May 10, 2006 — 13:42
  • “The Principle of Explanatory Defeat: If E is my only evidence for a belief B, then if I discover that the best overall explanation of evidence E does not imply that B is true, then E is defeated as evidence for B.”
    Really? Suppose my sole evidence E that Jones was not hurt in an accident is that my friend informed me. My friend is honest, but her honesty played no role in her decision to inform me. The only reason she informed me about Jones was that she wanted to relieve my concern. The best overall explanation of why I have evidence E is that my friend wanted to relieve my concern. That obviously does not imply that Jones was not hurt. But that also does not defeat the evidence E that Jones was hurt.

    May 10, 2006 — 14:24
  • Gordon Barnes

    This example doesn’t refute the principle. The appearance is due to underdescription of the example. Suppose that your friend tells you that “Jones was not hurt,” and that testimony from your friend is your evidence. Then suppose that you discover that your friend said this to you because she wanted to make you feel better, and thus she would say this whether or not it was true, just to make you feel better. Then is your evidence for your belief about Jones defeated? You bet it is. You have discovered that you would have this evidence (your friend’s testimony) even if Jones was hurt, and that defeats the testimony as evidence for your belief. What you did here was to confuse the issue by saying that your evidence is that your friend told you CORRECTLY that Jones was not hurt. This confuses matters, because in describing your evidence in this way, you have suggested that you somehow have more evidence than just your friend’s testimony, but then the principle does not apply. Finally, if your friend told you this because she knew that it was true, then even if making you feel better is part of the explanation of her telling you, it is not the whole explanation, UNLESS she would have told you even if it were not true, but then the evidence would be defeated, as argued above. So no matter how you slice it, this example does not contradict the principle.

    May 10, 2006 — 15:08
  • Matthew

    Gordon-
    Most people working in areas connected to religious studies are quick to admit that ‘religion’ lacks a suitable definition. (Funny that both religion and science should suffer from something of a demarcation problem.) Your common thread is lacked by some sects of Buddhism and by many Unitarians, then there is the Einsteinian religion of Dawkins and the religious nonrealism of Don Cupitt. This is just the short list. With a little reflection we could come up with more examples. Your common thread doesn’t extend to everything we might call religion. I suppose that you can stipulate your own definition but I don’t see why we’re under any pressure to accept it, especially since it seems to deviate from the norm. So, I think you’re just wrong on this front.

    May 10, 2006 — 15:34
  • “What you did here was to confuse the issue by saying that your evidence is that your friend told you CORRECTLY that Jones was not hurt”
    I said nothing of the kind. I said nothing even close to this. We’re basically talking about Pollock’s undercutting defeater principle (or something similar) right? I said that my friend was honest but that she did not tell me *because* she is honest. So I have a full explanation of why she told me that appeals to the fact that she wants to relieve my concern. But that explanation (contrary to what you want to conclude) does not therefore undermine the reliability of her information.
    Now your claim is that if the best overall explanation of why I have the evidence (in this case it is her desire to relieve me of my concern) does not imply that Jones was not hurt (and it plainly does not imply this) then my evidence is defeated. But this is pretty clearly false. Though it is true that the best overall explanation of my evidence DOES NOT imply that Jones was not hurt, it is nonetheless true that my evidence is not defeated: this is because an honest person who relates information E from a motivation consistent with both honesty & dishonesty does not therefore give us any reason to doubt the honesty of her report.
    So we have a case in which all of the following are true:
    1. The best explanation of my evidence E does not imply that Jones is not hurt.
    2. I would not have evidence that Jones is not hurt were he not hurt (the counterfactual is supported by her disposition to honesty).
    3. My evidence that Jones is not hurt is not defeated.
    4. Principle EP is false.

    May 10, 2006 — 15:52
  • Gordon Barnes

    Most buddhists do believe in unseen agents that act in the world. They are called “bodhisattvas,” and they are enlightened people who have chosen to remain in this world and help those who are here, rather than go to nirvana. These bodhisattvas are not all people who are still alive and physically present in this world. For example, one bodhisattva is “Kuan Yin, the Chinese manifestation of Avalokiteshvara, the Bodhisattva of Compassion. Although originally depicted as male, he gradually became represented as female. She appears to all who need her help, especially those threatened by water, demons, sword or fire. Childless women often turn to her for help.” So Buddhism does, in fact, fit my definition. As for Dawkins and Cupitt’s “religion,” I think it is plausible to say that these are not religions, for a variety of reasons. Here I think that people in Religious Studies would agree with me. Finally, although there might not be a generally accepted definition of “religion,” it does not follow that there is not a natural kind that is referred to with that term, even if we do not yet understand the kind yet.

    May 10, 2006 — 15:54
  • Gordon Barnes

    I assume that true causal statements support counterfactuals. This is controversial and the issues are tricky, but we can discuss them if you like. Now, if the cause of your friend saying this to you was ONLY that she wanted to relieve your distress, and NOt also that she wanted to inform you of the truth, then she would have said this to you even if it were false. If she would have said this to you even if it were false, and you know that, then what she said to you is not evidence for its truth. So what she said to you is not evidence for its truth. Now, there is a simple, valid argument. So which premise is false, and why?

    May 10, 2006 — 15:59
  • “Now, if the cause of your friend saying this to you was ONLY that she wanted to relieve your distress, and NOt also that she wanted to inform you of the truth, then she would have said this to you even if it were false.”
    The case involves a person disposed to be honest. But the fact that a person is disposed to honesty does not entail that, no matter what she does, we must explain her actions by appeal (even in part) to an aim to be honest. In the case I described, it was not a part of her aim or motivation to say honest things; her aim rather was to relieve me of some concern. She did so honestly not by aiming at it, but by being in possession of the relevant virtue. But you say, if Smith really was motivated solely by the desire to relieve me of my concern, then (1),
    1. If it were the case that Jones was hurt, then Smith would have told me that he was not hurt.
    But (1) is false, I think. If (1) is true, then so is (1′).
    1′. If it were the case that Jones was hurt, then Smith would have acted contrary to her character and told me he was not hurt.
    (1′) you’d agree is false. I say this is so despite the fact that Smith was not motivated by honesty. It formed no part of her motive. But it did form a part of her character.

    May 10, 2006 — 16:25
  • Matthew

    Gordon-
    I’ll grant your claim that most Buddhist do believe in unseen agents, but my claim was that some do not. All I need is one representative exception for your claim to “something that all religions have in common” to fail. You don’t deny that both the Buddhist and Hindu traditions have strong agnostic traditions within them. Are atheist Unitarians not religious? They get classed as a religion in most taxonomies. Don Cupitt is a priest within the Anglican church who thinks we can dispense with the supernatural and retain religion. I don’t think many in religious studies would say he wasn’t engaged in religious practices and part of a religion. Your claim about a common thread is tempting, but it has been rejected time and again because it is simply to restrictive. I think your claim is simply to strong to hold up.

    May 10, 2006 — 16:43
  • Gordon Barnes

    If part of my evidence is that my friend, who told me this, is honest by nature, then when her testimony is explained by her motive to relieve me, my total evidence is not thereby explained, because part of my evidence is that she is honest by nature, and that has not been explained by her motive to relieve me. So we do not yet have a counterexample to the principle. By the way, how would you explain the defeat in the example that I originally gave (which is, I think, due to Pollock)–the case of the objects bathed in red light? I think that my principle explains that defeat very nicely.

    May 10, 2006 — 17:07
  • Gordon,
    Pollock’s red lit objects example actually begs the question in a very interesting way. But it is a little lengthy for a post here. I will try to shorten it, or maybe send you a brief attachment with the argument (if you’d like). Let me know.
    But you say this,
    “If part of my evidence is that my friend, who told me this, is honest by nature, then when her testimony is explained by her motive to relieve me, my total evidence is not thereby explained. . .”
    Right, but you cannot be requiring an explanation of my total evidence. My total evidence includes such things as that I am not a BIV, that this is not a prank, that solipcism is false (there really are other minds and people), that I indeed have sense organs capable of reliably receiving the information from Smith and so on and on and on. You cannot be requiring that the explanation of Smith’s bit of evidence to me (which consists entirely in conveying to me that Jones is not hurt) ALSO explain all of this? If you are actually requiring that in your principle, you might make that explicit, I think.

    May 10, 2006 — 19:24
  • Gordon Barnes

    Mike,
    Let’s see if we can find exactly where we disagree. Let me start with this proposition, which I accept:
    (1). If I believe that P on the basis of evidence E, then if I discover that I would have possessed E, exactly as it is, even if P were false, then E no longer gives me evidence for P.
    Now, (1) seems intuitively obvious to me. If I discover that something that I took to be evidence for P is something that would obtain regardless of whether P is true or false, then I should no longer take this putative evidence as evidence for P at all. Now, are you saying that you do not share this intuition at all? So if I seem to see pink elephants in the room, but then my friend tells me that he just put acid in my lemonade that would cause this experience whether or not there were any elephants there, you think that I am still justified in believing in the elephants on the basis of my experience? I’m just very puzzled at what you’re saying here.

    May 10, 2006 — 20:26
  • Gordon Barnes

    Matthew,
    The agnostics among Hindus and Buddhists are heterodox in those traditions, and so I do not think that they qualify as genuine members of those religious traditions. Compare: do you think that an atheist is genuinely a Christian? Maybe you’ll say so, but I would be surprised. So although there are people who SAY “I’m a Hindu, but I don’t believe in any supreme mind,” I do not think that they qualify as genuine Hindus. Likewise for Buddhists who deny the existence of bohdisattvas. As for Unitarians, they have historically all been theists, and so they also fit my definition. If there are Unitarians now who are atheists, then I’m not sure that they really belong to any religion either. I mean, is what they do really that different from a social club that meets on Sundays? Finally, though, all that I need to argue here is that there is a natural kind that is defined by the belief in and worship of invisible, intelligent agents who act in the world. Even if our folk concept of “religion” fails to match this natural kind perfectly, it might still be possible to give a scientific explanation of that natural kind, which would, in turn, be a scientific explanation of a very large class of beliefs that we call “religious.”

    May 10, 2006 — 20:35
  • Gordon Barnes

    The first sentence in that last post should refer to Hindus only, since obviously Buddhists are atheists, who believe only in the bohdisattvas.

    May 10, 2006 — 20:37
  • Gordon,
    Yes, I agree that (1) is intuitive. But I’m pretty sure it is mistaken. Let me offer another counterexample that (as it happens) is analogous to one that Gibbard sent to me. It pretty clearly shows that (1) is mistaken. Here’s (1),
    (1) If I believe that P on the basis of evidence E, then if I discover that I would have possessed E, exactly as it is, even if P were false, then E no longer gives me evidence for P.
    Now the counterexample. Here’s what we are to suppose is actually true:
    All saints always do what is right and have no flaws. All nonsaints always do at least one thing wrong and have observable flaws. Now suppose saints are much in the minority. Suppose we randomly select a person whose name, let’s say, is Peter. The prior probablity (or the probability based on background information k alone) that Peter is a saint is low (since by hypothesis there simply aren’t many saints). But suppose we learn E that Peter has no observable flaws. Pr(Peter is a saint/k)

    May 11, 2006 — 7:57
  • That should say
    Pr(Peter is a saint/k) is much lower than Pr(Peter is a saint/k & E)

    May 11, 2006 — 7:59
  • Iorwerth Thomas

    ‘Likewise for Buddhists who deny the existence of bohdisattvas.’
    Bodisvattas are part of the Mahayana tradition, which post-dates Theravada Buddhism, which itself has no such entities. I’m not sure that calling the former heterodox when they predate the latter is entirely justified.

    May 11, 2006 — 9:45
  • Iorwerth Thmas

    That should be:
    ‘I’m not sure that calling the latter heterodox when they predate the former is entirely justified.

    May 11, 2006 — 9:48
  • Gordon Barnes

    Mike,
    OK, very interesting. I have two thoughts. First, I actually have the intuition that my evidence that Peter is a saint IS defeated in the case as decribed. If I know that even if saints had flaws, they would be good at hiding them, then I think that I am no longer in a position to judge, merely on the basis of observable flaws, whether or not someone is a saint. However, with that said, I realize that I am in dangerous territory here, which started as soon as I “went counterfactual.” Robert Shope’s work, years ago, on the conditional fallacy, showed that counterfactual analyses of epistemic concepts are extremely problematic. Ok, so in light of that, I want to offer another version of my principle. More precisely, it is a different principle, but it’s in the same neighborhood. Suppose that I replace the use of explanatory and counterfactual concepts with a causal concept, and without making any assumptions about the conceptual connections between causal concepts and explanaroty or counterfactual concepts. So, the new principle is The Principal of Causal Defeat. Here it is:
    (C) If my only evidence for my belief that P is E, then if I discover (1) that there is a condition C that is causally sufficient for E, and (2) that C does not imply, or even raise the probability of E, then E is defeated for me as evidence for P.
    Now, let me go back to your original counterexample, involving testimony, which might seem to be a counterexample here as well. But now what I want to say is that the causally sufficient condition for Smith’s telling me that Jones is ok involves both the fact that Smith wants to relieve my stress, AND the fact that Smith was given this information from a reliable sourse and Smith is honest. Now, when we include this in the causally sufficient condition, it DOES imply that Jones is ok. Moreover, this condition, (C), will account for the pink-elephants-due-to-acid-lemonade example. Finally, I have growing sympathy for your skepticism about my principles, but there must be some good account of the defeat in that acid-lemonade case, right? So how would you diagnose that case?

    May 11, 2006 — 14:00
  • Gordon Barnes

    Iorwerth,
    On Theravada Buddhism, ok, fair enough. Now, at this point I am inclined to stress the similarities between Karma and an invisible, intelligent agent that acts in the world. I realize that Karma is not conceived as an intelligent agent. However, it seems to me that Karma governs the universe as if it WERE an intelligent agent. What I mean is that Karma governs the universe for the sake of certain ENDS. That is to say, it’s activity is teleological. To my mind, that makes Karma very much like an invisible, intelligent agent that acts in thw world. So I think that my condition for “religion” is still defensible here, with just a little tweaking. More specifically, I’ll replace the notion of an aganet with the broader notion of an invisible force that acts teleologically, for the sake of ends that specifically involve the welfare of human beings.”
    By the way, Iorwerth is an extremely cool name. I suspect it is Welsh. Are you Welsh? Dwin siarad cymraeg, dim ond typpin bach.

    May 11, 2006 — 14:10
  • Iorwerth Thomas

    A fi hefyd!
    ‘Now, at this point I am inclined to stress the similarities between Karma and an invisible, intelligent agent that acts in the world. I realize that Karma is not conceived as an intelligent agent. However, it seems to me that Karma governs the universe as if it WERE an intelligent agent. What I mean is that Karma governs the universe for the sake of certain ENDS. That is to say, it’s activity is teleological. To my mind, that makes Karma very much like an invisible, intelligent agent that acts in thw world.’
    I’m not sure; (not being much of an expert on Buddhism – beyond the Theravada-Mayahana distinction – what follows should be taken with a pinch of salt) Karma seems to me to be more teleogical in the sense that Lagrangian or Hamiltonian mechanics in physics is telelogical than in the way an agent is telelogical. (Assuming classical physics, if I wrote a Hamiltonian for the universe, it would behave telelogically as I integrate over time, but I shouldn’t conclude from that that it’s an agent!)
    ‘…for the sake of ends that specifically involve the welfare of human beings.’
    I think that probably captures the things that we’re interested in, though I doubt that it entirely exhausts the notion of ‘religion’. But that may just be me…
    As a sidenote, as a non-philosopher, I’m finding the discussion of (1) pretty fascinating. Myself, I’m a little dubious about it; I think that it’s the kind of principle that leads to underdetermination problems in physics. In order for it to apply, do you need
    a) the mere logical possibility of an alternative explanation
    or
    b) for an actual alternative to be described?
    Either seems a little strong.

    May 12, 2006 — 4:14
  • “Moreover, this condition, (C), will account for the pink-elephants-due-to-acid-lemonade example. Finally, I have growing sympathy for your skepticism about my principles, but there must be some good account of the defeat in that acid-lemonade case, right? So how would you diagnose that case?”
    Gordon, I haven’t all the details of the pink-elephants case. Is the case analogous to Plantinga’s alpha centuri case of coincidentally detecting large gray beings in an improper enviornment? How does it go, exactly?

    May 12, 2006 — 16:40
  • Gordon, I think I’ve located the example. But it doesn’t seems to illustrate what Pollock says it does. He says (if I have the right case) that we *know* all of the following:
    1. People generally tell the truth.
    2. Robert says the elephant beside him looks pink.
    3. Robert becomes unreliable in the presence of pink elephants.
    4. Robert’s statement gives us prima facie evidence that the elephant does look pink, which gives us a reason for beliving that it is pink.
    5. But (4) combined with Robert’s unreliablity in the presence of pink elephants gives us a defeater for thinking that the elephant looks pink.
    6. So we should not conclude that the elephant beside Robert looks pink.
    So I think Pollock concludes that our evidence for the claim that the elephant *is* pink is defeated. But I’m pretty sure this isn’t true. Here’s a little proof. (a) Robert is either reliable or not in his report about the presence of a pink elephant. (b) Suppose he is reliable: in that case we should believe that there is a pink elephant beside him. (c) Suppose he is not reliable: in that case we should believe that there is a pink elephant beside him. (d) Therefore we should believe there is a pink elephant beside him.
    We know that (b) is true for obvious reasons. But why is (c) true? (c) is true because we know that Robert is only unreliable in the presence of pink elephants. So if he unreliably reports the presence of a pink elephant, there must be one present.

    May 13, 2006 — 12:51
  • Gordon Barnes

    Mike,
    Sorry to disappear. Been busy grading. Your argument against Pollock seems exactly right. However, I should have clarified that I was thinking of my own little example, mentioned briefly in an earlier post. It is this. Suppose that I have a visual appearance as of pink elephants. This gives me prima facie evidence that I am in the presence of pink elephants. However, then my friend tells me that as a practical joke he spiked my lemonade with some acid that is known to cause people to hallucinate pink elephants. Then it seems to me that my prima facie evidence is defeated. Moreover, I think that the reason is this: I now know that there is a causally sufficient condition for my visual experience (the acid), and this causally sufficient condition does not make it probable that I am in the presence of pink elephants. So I think that my latest, causal version of my principle explains this case of defeat. What do you think?

    May 13, 2006 — 21:25
  • Gordon Barnes

    Noswaith Dda, Iorwerth,
    Your points about teleology are well taken. I will have to think about that some more. Moreover, I agree that, even if correct, my principle would not constitute an exhaustive characterization of religion, but it might still constitute an indispensable component. However, even there the challenges that have been made give me pause. I will have to think about it some more.
    I’m intrigued by your question concerning the debate over (1), but I’m not completely sure that I’m understanding it completely. Could you elaborate on it a bit more>

    May 13, 2006 — 21:32
  • Gordon Barnes

    Matthew,
    Unfortunately, I no longer have an electronic copy of “Modal Inquiry.” However, I have developed one of the central arguments from it into a paper that is forthcoming in Philosophical Studies, and I would be happy to send you that, for what it’s worth. Much of the rest of it has been duplicated since then in work by David Chalmers and Frank Jackson. I’m glad to hear that you’re also interested in modal epistemology. It’s fascinating stuff.

    May 13, 2006 — 21:37
  • Hi Gordon, here’s your example,
    Suppose that I have a visual appearance as of pink elephants. This gives me prima facie evidence that I am in the presence of pink elephants. However, then my friend tells me that as a practical joke he spiked my lemonade with some acid that is known to cause people to hallucinate pink elephants. Then it seems to me that my prima facie evidence is defeated.”
    I’m not sure. If, for instance, (i) your visual experience is caused by the presence pink elephants and (ii) you learn that you would have those experiences whether or not the elephants were present (the acid would have caused them had they not been actually present), then I’d say that your evidence is not undercut.
    But you might be saying instead (a) what is actually causing your experience is the acid and not the presence of the pink elephants. In that case, it is pretty clear that your evidence is undercut (really, but hypothesis.
    You might also be saying (b) you do not know whether your experience is caused by the presence of the pink elephants or the acid, given your information. Whether your evidence is undercut here I think will depend on your intuitions about internalism/externalism.
    Am I following your argument?

    May 15, 2006 — 13:49
  • Gordon Barnes

    Hi Mike,
    Yes, that’s the argument, and I agree that the verdict will depend on exactly which of the scenarios that you distinguish is involved. I am thinking of the third scenario. Suppose that I have no independent evidence for the existence of the pink elephants. All I have is my visual experience. And now I know that the acid in my drink usually causes just such an experience. However, I have no independent evidence that, in this particular case, the acid is the cause of my experience. In exactly that situation, I’m inclined to think that (using an internalist notion of justification) my justification for my belief in the pink elephants has been defeated. Now, if we use an externalist notion of justification, then I agree that it really depends on the actual causal process that produced my belief in this case. I will have to think about it some more.

    May 15, 2006 — 17:34
  • Iorwerth Thomas

    Gordon,
    I think that I was wondering about what conditions would be sufficient for (1) to apply; given that it seems to me plausible that, in a given scientific field (say physics), evidence for E could be explained away another by another theory [1]. Since, in the space of all possible theories explaining E, one of this type is certain to exist, should this possibility undercut the evidence for E, or should the alternative actually have to be spelled out for this to occur?
    Admittedly, these awful ramblings are largely due to the amount of Duheim I’ve been reading recently, so it’s likely that my understanding of the issues is utterly dated.
    [1] Ignoring the issue of the relative parsimony of the theories for a moment; let’s say they’re equally so.

    May 16, 2006 — 9:39
  • Gordon Barnes

    Prhynawn Da, Iorwerth,
    Got it. I think that (1) would apply only when one of the alternative theories is an actually-existing theory that would BETTER explain the data than the theory under consideration. Now, as you are aware, the “better” in “better explains” raises very large questions. I am just assuming that there is some clear way of explicating that notion.

    May 16, 2006 — 13:41
  • Gordon, I’m not sure that you’re still interested in this line, but I did want to point up one way in which undercutting defeaters have gotten some play in discussions of evil. Steve Wykstra’s CORNEA principle (modified in ’96 in Dan Howard-Snyder’s well-know collection on evidence and evil) is supposed to be an undercutting defeater.
    Wykstra’s argument–to be very brief–is that (what he calls) God-purposed goods are such that, if there were any, you would not be able to observe them. From this he wants to conclude that the observation that there are no God-justifying goods is not evidence that there are no goods.
    This is the general line of that very unfortunate defense of theism dubbed ‘skeptical theism’. The undercutting defeater is supposed to work this way.
    1. After careful consideration, I observe no goods that justify (Rowe’s) evils e1 and e2.
    2. Therefore there are no good justifying e1 and e2.
    The evidence in (1) (and for (2)) is undercut, say the ST’s, because,
    3. It is possible that there are God-purposed goods justifying evils e1 and e2.
    4. If there were God-purposed goods justifying e1 and e2, then we would not observe them.
    But (3) and (4) do not undercut the evidence in (1) unless we have reason to believe that it is more than just possible that there are such goods. Suppose we do not already have reason to believe that among the goods that exist there is a subset of God-purposed goods. In that case, despite (4), we have (5) true.
    5. The observation of no goods justifying e1 and e2 is good evidence that there are no goods justifying e1 and e2.
    The fact that there are possible goods that we are in no position to observe or that are way beyond our ken does not alone constitute an undercutting defeater for our observation that no goods justify e1 and e2. To get an undercutting defeater we have to have reason to believe in addition that such God-purposed goods are actual. And that of course is exactly the point in question.

    May 16, 2006 — 14:14
  • Christian

    Dear Mike,
    I’m going to jump in on this because I really like the point you are making. I have some questions, though.
    1. After careful consideration, I observe no goods that justify (Rowe’s) evils e1 and e2.
    Thus, [It is therefore likely that]
    2. There are no good justifying e1 and e2.
    ST’s challenge the inference by proposing the following general principle:
    General Principle: The fact that (1) is true makes it likely that (2) is true only if: if there were goods that justify e1 and e2, one would likely observe them.
    Now, this is different from your (4):
    4. If there were God-purposed goods justifying e1 and e2, then we would not observe them.”
    (4) seems to me stronger than the general principle. I take ST’s to argue the general principle is enough to undermine the inference from (1) to (2) without appeal to (4). I would also assume that they would argue that the general principle undermines (5). So, I’m wondering whether you think that ST’s are committed to your (3) and (4) together rather than, say, (3) and the genrsl principle?

    May 16, 2006 — 15:13
  • Hi Christian,
    I don’t understand how the general principle (GP), as you state it, is supposed to help the ST’s. Shouldn’t the consequent of your GP instead be “one would *not* likely observe them”?
    But GP is false unless we have reason to believe that there are God-purposed goods. It is the GPD’s that are unobservable, not any old good that might justify those evils.

    May 16, 2006 — 15:22
  • Christian

    It’s been quite some time since I looked at this type of argument so forgive me.
    I recall an analogy: I’m sitting by the window looking towards the garden. I think, “I don’t see any slugs in the garden” and I form the belief, on the basis of my obeservation, that “there likely are no slugs in the garden.” This is supposed to analogous to the move from (1) to (2). The idea then is that the first case is bad and the reason that it’s bad is that I should only make the inference if my observation makes it likely that there are no slugs in the garden. This is just (GP). But my observation doesn’t make it likely that there are slugs in the garden.
    Anyway, you asked, “Shouldn’t the consequent of your GP instead be “one would *not* likely observe them.”
    I’m wondering why you think this? Why not accept (GP) as stated and leave it as an open question whether (3) is true or whether your modifaction is needed? Isn’t this what a STist will say?
    Incidentally, a ST need not claim that the relevant goods are unobservable right? For example (I can’t remember his name, Steve somebody) argues that the goods may be observable, but that we are missing them because we are corrupted.

    May 16, 2006 — 15:48
  • Christian

    Mike,
    I meant “Why not accept (GP) as stated and leave it as an open question whether (4) is true.” What do you think?

    May 18, 2006 — 11:47
  • Sorry Christian,
    I’ve got all of this work to do and I’m letting discussions lapse. I reread your use of (GP) and it says this,
    General Principle: The fact that (1) is true makes it likely that (2) is true only if: if there were goods that justify e1 and e2, one would likely observe them.
    I’m not sure I agree. The inference from (1) to (2) need not depend on (GP). Suppose there are no goods at all justifying e1 and e2. Perhaps we are living in a nihilistic world. So the observation of no goods justifying e1 and e2 is evidence that there are no such goods. It might nonetheless be true that the closest non-nihilistic world in which e1 and e2 are justified is a world in which (i) our epistemic powers are so limited that we do not observe them, or (ii) the goods are so weird and remote that we do not observe them or (iii) there are only sentient beings suffering and no rational beings capable of knowing this, or … or….
    In any case, (GP) is consistent with (4) unless there is reason to believe that there are God-purposed goods.
    4. If there were God-purposed goods justifying e1 and e2, then we would not observe them.
    So (GP) can be true along with (4). I don’t think there is any doubt (given the description of God-purposed goods) that they would be unobservable. So I don’t doubt (4). What I am claming is that (4) is consistent with (GP), unless we have reason to believe that there are God-purposed goods. Make sense?

    May 18, 2006 — 13:56
  • Christian

    Mike, it’s that time of year I know. Grading, grading and then more. Damn finals.
    So, I see what you are saying. I realize now that “God-purposed goods” is a technical term as you are using it (But your use makes 4 trivial if I’m not mistaken). I was using it non-technically, as a good that would justify god in permitting some evil. Let me stipulate that I am talking about goods in my sense. And I also think that (4) is consistent with (GP), or at least, I cannot think of a good reason to think they are inconsistent. I didn’t mean to imply otherwise.
    Now, here is my worry: STist can accept (GP) and deny that (4) is relevant to the validity of the inference from (1) to (2). (4) will not undermine the move from (1) to (2) for the reason you suggest. However, if (GP) is true, then the move from (1) to (2) is undermined. So, ST should accept (GP) because they think the move from (1) to (2) is bad. Furthermore, they can produce examples to support (GP).
    You propose a counterexample to (GP). A world w where there is evil and no justifying good (This will be an impossible world to many!). In this world if we observe the evil and no justifying good you suggest it is likely that there are no justifying goods on the basis of the observation (not the stipulation). You say that perhaps the closest world to w is such that people cannot observe the good that justifies the evil, so the counterfactual:
    “if there were goods that justify e1 and e2 in w, one would likely observe them” is false. This renders (GP) false because if it is true, then it is necessarily true.
    I like this example. I need to think about it.

    May 18, 2006 — 17:55