I promised to solicit responses from both Anthony Grayling and Richard Dawkins about their recent debate on the possibility of evidence for God. As you recall, their conversation had left me a bit unclear about their positions, and I asked these gentlemen for clarification. Richard has not yet responded, though he posted my analysis on his website, where it’s stimulated a fair amount of discussion. But Anthony did respond, and I want to record our email exchange here:
Before I do, let me summarize what I see as the three main positions advanced by atheist readers on the nature of evidence for god:
- The whole question is incoherent because one cannot construct a valid hypothesis to test. This is the position of Steve Zara and P. Z. Myers, with Myers asserting that “There is no valid god hypothesis, so there can be no god evidence, so let’s stop pretending the believers have a shot at persuading us.” I believe this is the consensus of readers who have posted both here and on Pharyngula. This is, of course, telling the faithful, “Forget about giving us evidence for god: there’s nothing you can do to make us believe in him.”
- There could be evidence for god, but what has been offered so far is pathetic—not even remotely convincing. Thus there is no basis for believing in a celestial deity. Nevertheless, it is possible (though the possibility is almost zero) that some evidence might arise that we would find convincing. This is my position and, I believe, the position that Richard took in The God Delusion, where he put himself as a 6.9 on a 7-point scale ranging from 1 (“I know that God exists”) to 7 (“I know there is no God”).
- There could have been evidence for God, but none has ever surfaced. Therefore we can reject the god hypothesis and need not consider further evidence. This is, I think, the position that Grayling takes, and is laid out in his email response to me (all emails published with permission, and put in chronological order, separated by lines):
Jerry – oh dear, obviously what I said sounded too much like throat-clearing when actually it was, or was intended to be, genuine spitting. The point I made (and I here reprise a passage from my ‘To Set Prometheus Free’ on Russell and ‘we cannot prove that there is not a god’) is that when you understand what proof is in the non-deductive, contingent sphere of reasoning, you see that you can prove there are no supernatural entities. And both Richard and I were pretty clear that putative proofs FOR supernatural entities would, if they genuinely were proofs, prove that supernatural entities are natural, not super. – Note that your position, if you seriously mean that you think it might be conceivable or possible that there could be evidence for a deity, is agnostic, not atheist; and the following remarks are directly relevant to you therefore, because agnosticism is incoherent. The relevant passage follows:
Russell felt bound by logic to admit that he would be at a loss to find arguments to disprove the existence of (for example) the Olympian deities. His position in this respect merits challenge. As a logician he should have distinguished between proof in a formal deductive system (demonstrative proof) and proof in the empirical setting (scientific proof). The former consists in deriving a conclusion from premises by rules, and are literally explications in the sense that all the information constituting the conclusion already exists in the premises, so a derivation is in fact a rearrangement. There is no logical novelty in the conclusion, though often enough there is psychological novelty, in the sense that the conclusion can seem unobvious or even surprising if the information constituting it was highly dispersed among the premises.
Demonstrative proof is watertight and conclusive. It is a mechanical matter; computers do it best. Change the rules or axioms of a formal system, and you change the results. Such proof is only to be found in mathematics and logic.
Proof in all other spheres of reasoning, and paradigmatically in science, consists in adducing evidence of the kind and in the quantity that makes it irrational, absurd, irresponsible or even a mark of insanity to reject the conclusion thus being supported. The definitive illustration of what this means, not least for the use that theists would like to make of the myth that ‘you cannot prove a negative’, is Carl Sagan’s dragon-in-the-garage story. On this basis someone who on the basis of evidence and reasoning concludes that it is irrational, absurd, irresponsible or even lunatic to believe that there is such a thing as deity, might further ask whether it is nevertheless none of these things to believe that there might be such a thing as deity.
Consider an analogy. Suppose someone thinks: ‘My belief that rain will wet me if I do not use an umbrella is (only?) inductively justified; therefore I am entitled to believe that it is possible that rain might not wet me next time I do not use an umbrella when it rains.’ Is the belief that ‘rain might not wet me next time’ less irrational or absurd than the belief that rain does not wet at all? Obviously not. For this reason Russell’s use of ‘agnostic’ as functionally equivalent to ‘atheist’ but with the reservation of a quibble about proof is seen to turn on an assimilation of proof concerning matters of fact to proof of the demonstrative kind – and it is a quibble that does not, pace our man with the umbrella, hold water.
Pointing this out matters because misapprehensions about the nature of proof continue to support the apparent plausibility of agnosticism. But agnosticism, as the position that entertains the possibility that there might be or could be one or more supernatural agencies of some sort, is an irrational position, for precisely the same reason as holding that there might be or could be fairies or goblins or the Olympian deities or the Norse gods.
There is another consideration. The claim that there are supernatural entities/states of affairs is not exactly the same as that there are gods or a god. The supernatural embraces ghosts, fairies, goblins, and the like. The question, as the foregoing passage shows, is about the rationality of believing such things. But particularly as regards a god: on the standard definition of an infinite, omniscient, omnipotent, benevolent etc being – on inspection such a concept collapses into contradiction and absurdity; as omnipotent, god can eat himself for breakfast…as omniscient it knows the world it created will cause immense suffering through tsunamis and earthquakes, and therefore has willed that suffering, which contradicts the benevolence claim…etc etc…to say nothing of local suspension of the laws of nature for arbitrary reasons e.g. in answer to personal prayer, which makes a nonsense of the idea that the world or the deity is rationally comprehensible: and if either or both are non-rational then there is nothing to talk about anyway.
In response to this, I wrote Anthony:
Thanks; I understand your position better. I think you didn’t express this in the conversation with Dawkins as clearly as you did here. Do I have your permission to post all this? I basically agree with you, but will have a bit more response when I digest it and post it. And I do agree that the “you can never prove a negative” stuff is stupid.
So—and let me get this straight—you do think that the possibility of a “supernatural” god (I do quarrel with how “supernatural” is used, since any god who interacts with the world is doing something natural by definition) is sufficiently low that THERE IS NO LONGER ANY EVIDENCE OR EVENTS that could ever convince you that a god exists?
Jerry – post it by all means. And you are right: there is no evidence that would convince me that there are goblins or fairies or Norse gods or a Christian one. To my earlier remarks I should add that when one reflects (a) on the meaning of ‘rational’ from ratio = proportion to the evidence, and (b) the requirement of overall fit between a given claim, e.g. ‘fairies exist’, with the understood, widely experienced and examined nature of the world, and particularly in the fairy case of the fit between fairy-talk and zoology, and (c) when one understands where fairy talk comes from (childrens’ stories, folklore, superstitious explanations of phenomena like fainting spells or losing one’s shoelaces) one sees what an evaluation of the claim involves. Now think of claims about deities, and apply the same tests and considerations. They are exactly parallel. This is the point of talking about the rationality (ratio-nality) of beliefs and their credibility. – Anthony (post this too if it adds)
And I responded to Anthony this morning:
I’ll post our exchange of emails, including this one, this morning. I did want to respond briefly to your comments. I see three atheist positions on God-evidence: P.Z. and Zara’s, which is that the whole hypothesis is incoherent and not worth considering (i.e. there can never be no good evidence for God); the idea that there could be evidence, but there hasn’t been any and so we can act as though God doesn’t exist with near, but not complete, certainty (the position of Richard and me); and the idea that there has been ample opportunity to get evidence for God, but since none has ever surfaced we can abandon the idea, and completely refuse to consider further evidence. This last one appears to be your position.
Our disagreement boils down, then, to the difference between positions 2 and 3. But I’m not sure why you say that at one time (perhaps 50 AD?) you would have considered evidence for God, but for you the case is now closed completely. This position seems manifestly unscientific for several reasons. A god, for example, might not have chosen to show himself until now. Granted, that seems dumb, but who knows?
More important, it seems unscientific to say that “the case has been fully and irrevocably decided at point X—no more evidence could ever count.” That is not the way scientists treat scientific hypotheses. Take Darwinism, for instance. I think all of us—certainly Richard and myself—accept the truth of evolution with as much tenacity as we accept the idea that there is no god. There has been a ton of evidence in favor of evolution, and no convincing evidence for creationism. And there is a ton of evidence in favor of “no god,” and no convincing evidence to the contrary.
And yet you won’t find an evolutionist who says, “Okay, there’s now enough evidence for evolution that we can regard it as an absolute truth. We needn’t consider any further evidence that purports to overturn it.” For there’s always the possibility, however remote, that such evidence could appear. Haldane, of course, cited the fossil of a rabbit in the Precambrian. I have a list of other things that could overturn evolution, such as the pervasive finding of adaptations in some species that only benefit members of other species.
Indeed, I regard the evidence for evolution as so strong that I would consider anybody who rejects it to be, as you say, “irrational, absurd, and irresponsible,” though I don’t know that I’d go so far as to say “insane.” And the same is true of the evidence against God. As Victor Stenger has emphasized, you can indeed prove a negative if the evidence could have been there but isn’t. That is the situation with the god hypothesis, and with creationism.
Would you then call Richard and me “agnostic” on not only the God question, as you do, but also on the evolution question? That seems unfair. We are evolutionists in the same way we are atheists: we tentatively (but firmly) accept evolution and the absence of a God, for there is no absolute truth in science. Do note that in Sagan’s essay, which you cite, he says that “the only sensible approach is to tentatively reject the dragon hypothesis, to be open to future physical data, and to wonder what the cause might be that so many apparently sane and sober people share the same strange delusion.” The key bits here are “tentative” and “be open to future physical data”. Sagan, then, is on our side, not yours.
Richard and I, then, are atheists in precisely the same way we are Darwinists. If you choose to call us “agnostics” on both issues, that is your choice, but I find that characterization misleading. Would you tell a creationist that Dawkins and Coyne are “agnostic on evolution”? When I say I am an atheist, I mean that I am 99% sure that there is no god, and I don ‘t see any evidence on the god side.
You say that that this position is “agnostic”, because it leaves room for the possibility that there is a god. I’m not sure why you’re so worried about that, unless you think that somehow it gives fodder to the faithful—perhaps by implying that if we can’t disprove god, one must somehow exist.
If the question of god is not a logical one but a scientific/empirical one, as we both think it is, then we should behave according to the dictates of science, which is to always regard “truth” as provisional, even if it seems almost impossible that that truth could be overturned. That holds for both evolution and atheism. It grants no credibility to creationism to hold evolution as a firm but tentative truth, and it grants no credibility to religion to consider God’s existence in the same light.
We’ll now leave this issue behind—at least for a while—but I do think that although few have changed their minds, the volume of comments shows that it’s been a stimulating topic.
UPDATE: Anthony has written back that I misunderstood him—that he aligns with P.Z. and Zara all along. His email:
No, I don’t think that every effort has been made to look for evidence and none has turned up: I don’t know how you got that out of my remarks! You and Richard think it’s an empirical matter whether there are deities (or fairies? goblins? consider why you think the latter are zoological non-starters) and I think it’s a matter of coherence of the concept (so I’m with PZ and Zara) because of a variety of considerations that apply: the contradictions in the concept, the mistaken notions of proof in play, the requirement that enquiry about it be rational, our understanding of what work the concept did historically (as a substitute for explanation in the ignorant past of mankind, e.g.), and the psychological motivations for belief in it. The point is that ‘god’ is not like ‘ether’ – it is not amenable to empirical investigation, and does not occupy a slot in some systematic framework of thinking about the world that might be improved on in the light of better theory or observation. It does no work because it purportedly does all work; like a contradiction it entails anything whatever; it is consistent with all evidence and none. These considerations constitute the proof that it is an empty concept. – If you treat the word ‘god’ as a name for a putative entity that might or might not exist and such that something might count as evidence for or against its existence, as you do, then you are committed to agnosticism about everything that can be given an apparent name. But ‘god’ is not like ‘yeti’ (which might – so to say: yet? – be found romping about the Himalayas), it is like ‘square circle’. Trying to explain to someone who thinks that ‘god’ is like ‘yeti’ (namely, you) let alone to someone who thinks ‘god’ is like ‘Barack Obama’ (names an actual being, as Christians and Muslims do) that it is actually not like ‘yeti’ but like ‘square circle’ and that nothing can count as evidence for square circles, is harder work for ‘god’ than ‘square circle’ only because religious folk have been squaring the circle for so long!
I reject Anthony’s assertion that God is not amenable to empirical investigation, since one can empirically investigate claims about how God interacts with the world. The efficacy of prayer is one of these. I believe Grayling is referring here to a deistic god, since theistic gods need not be “consistent with all evidence.” The existence of earthquakes, for example, is not consistent with a benevolent theistic god. I still maintain that if one claims that a god interacts with the world in certain ways, then those claims can be investigated empirically. To me the existence of a deity is not a matter that can be ruled out by philosophy or logic from the get-go; it’s a matter for empirical observation and testing.