If you have more time, please read the much better extended version here.
The extended version has this key exchange about whether one can prove or disprove God's existence through science and how to best approach this subject with people who believe. I think Krauss understands that beliefs are not simply related to evidence and not everyone takes an academic approach to religion. Thus, while Dawkins' statistical arguments are valid in an academic sense, they may not have much impact on believers other than to further antagonize them against science and this usually results in a breakdown of communication. I think Sagan made an effort to understand where the believers are coming from, and thus despite his open atheism, he had respect across the board. Krauss, here, is making a similar effort.
Krauss: You have cogently argued in The God Delusion that religion is bad science. I would argue, however, that this is particularly inappropriate, and in fact falls into the same trap fallen into by those who push Intelligent Design in science classrooms, as well as those who fund Templeton Foundation grants that attempt to foster scientific evidence for God. I have framed this issue in language that hearkens back to Carl Sagan, who said that absence of evidence is not the same as evidence of absence. Would a world without God necessarily look any different than the world we live in? Most scientists would say no, and thus claim we do not need the God Hypothesis to explain anything about nature. On the other hand one might also ask: Would a world with a God necessarily look any different than the world we live in? People of faith would argue no, and in so doing feel vindicated in their faith. The problem is that both groups are correct, and nothing either can say is likely to influence the other.
Dawkins: I have several times said that a universe with a God would be a very different kind of universe from one without. You have translated this into operational terms, and consequently arrived at the legitimate question of whether the two kinds of universe would look different. Not be different (my question) but look different (your question, where 'look different' can presumably mean any difference, detectable in any way by any of our sense organs or scientific instruments). I agree that yours is an important question, and I agree with you that it might be surprisingly hard to detect, by observation or experiment, whether we live in a god-free universe or a god-endowed one. Nevertheless, I still maintain that there is a cogent sense in which a scientist can discuss the question. There still is a sense in which we can have an interesting and illuminating scientific discussion about whether X is the case, even if we can't demonstrate it one way or the other by observation or experiment. How can I argue this and still claim to be doing science?
In The God Delusion, I made the distinction between two kinds of agnosticism. Permanent Agnosticism in Principle (PAP) is exemplified by that philosophical chestnut, "Do you see red the way I see red, or might your red be my green or some completely different hue (‘sky-blue-pink’) that I cannot imagine?" Temporary agnosticism in practice (TAP) refers to things that we cannot (or cannot yet) know in practice but nevertheless have a true scientific reality in a way that the 'sky-blue-pink' conundrum does not. Bertrand Russell's hypothetical orbiting teapot might be an example. Some people think the question of God’s existence is equivalent to ‘sky-blue-pink’ (PAP), and they wrongly deduce that his existence and non-existence are equiprobable alternatives. I think we should be TAP agnostic about God, and I certainly don’t think the odds are 50/50.
Statements such as 'There are (or are not) intelligent aliens elsewhere in the universe’ are clearly TAP statements insofar as we are talking about the observable universe this side of our event horizon. At any time, a flying saucer or a radio transmission could clinch the matter in one direction (it can never be clinched in the other). What, though, of statements about the existence of intelligent aliens in those parts of the universe that are beyond our event horizon, where the galaxies are receding from us so fast that information from them can never in principle reach us because of the finite speed of light? In this case, at least according to the physicists I have read, the aliens would forever be undetectable by any means whatever. On the face of it, therefore, we would have to be PAP agnostic about them, not just TAP agnostic.
Yet I would resent it as a scientist, not just as a person, if you tried to rule out any scientific discussion of aliens beyond our event horizon, on the grounds that it is beyond the reach of empirical test (PAP). Suppose we take the Drake equation for calculating the odds of alien intelligences existing, and apply it to the whole universe rather than just our galaxy. Clearly it will yield very different results depending on whether we hold to a finite or infinite model of the universe. Those two models of the universe are discriminable by empirical evidence, and that empirical evidence would therefore have some bearing on the probability of alien life existing somewhere in the universe. Hence the probability of alien life is a question of TAP rather than PAP agnosticism, even though direct empirical experience of the aliens might be impossible. It is not obvious to me that gods are beyond such probability estimates, any more than aliens are. And a probability estimate is the limit of my aspiration.
Krauss: First, I have to say that I have nothing against trying to think about phenomena that might never be directly measurable. I do this all the time in my work in cosmology, where I consider the possibilities of other causally disconnected universes. Of course I do this to see if I can resolve outstanding puzzles in the physics of our universe. If this approach turns out not to work, then I find the issue less interesting. I also agree with you that probabilities are important, but I think your example of the Drake equation is quite relevant here, but perhaps not in the way you intended. First of all, the Drake Equation is really applied locally, within our galaxy. If the probabilities turn out to be small that there is more than one intelligent life form in our galaxy, I think most astrophysicists will not be particularly interested in worrying about the civilizations that might exist in other galaxies but which will be forever removed from us. But more important is that fact that the probabilities associated with the Drake equation are almost all so poorly known that the equation really hasn’t driven much useful research. Varying each of the conditional probabilities in the equations by an order of magnitude or so, one can derive results that either argue strongly in favor of extraterrestrial intelligence, or strongly against it. The proof is likely to come from empirical searches. As bad as this is, I would argue it is far worse when attempting to quantify probabilities for the existence of divine intelligence or purpose in the universe.
Indeed, I have argued that questions of purpose in the universe are generally not a part of science, and the best example I know is that of Georges Lemaître, the Belgian priest who was also a physicist, and the first person to realize that Einstein’s General Relativity implied there was a Big Bang origin to our universe (a claim initially much derided by Einstein). Following this realization Pope Pius XII issued a statement that said science had proved Genesis. Lemaître responded appropriately. He wrote to the Pope and urged him to stop saying that. The theory in question was a scientific theory whose predictions could be tested. What religious implications one took from the theory depended upon one’s metaphysical leanings. One could take it to validate Genesis, by implying that the Universe had a beginning—a revolutionary scientific claim at the time. But one could equally well take it to imply that there is no need for a God, that the laws of physics are all that are required to understand the universe right back to the beginning. The point is that the science is accurate in describing how the universe works, independent of the metaphysical implications one derives from it. The same is of course true for evolution, which happened and is happening, whether or not one chooses to believe in God.
Dawkins: Of course Lemaître was very wise (although I must add that I am left wondering why he remained a priest at all). As for the option that his physics might or might not be taken to support Genesis, why is it even an interesting question? There never was reason to expect that the writings of an unknown scribe, probably less than a millennium ago in Babylon, might have any special insight into the origin of the universe. If Genesis happened to get something right, why would that be anything but a trivial accident?
Krauss: Well, the key point that you are neglecting here is that there is a reason to believe it might capture some truth about the universe, but only if you believe in God. Presumably Lemaître did. Getting back to the issue I raised earlier, I do not mean to say, of course, that science could never provide positive evidence of design or purpose. If, for example, tonight the stars suddenly lined up in the sky and spelled out ‘I am Here,’ most astronomers would be willing to consider a supernatural cause. However, the absence of such evidence of design—in spite of what the con artists and misguided pseudoscientists who presently claim that living systems currently provide such evidence—does not logically rule out the possibility that our universe, and life within it, has some purpose.
Dawkins: I am perennially baffled as to why anybody thinks it is an important point that you can't logically rule out some possibility. There is an infinite number of possibilities that we can't logically rule out but which we nevertheless don't take seriously because we have no positive reason to do so. This was the original point of Russell's teapot.
Krauss: My point is that if you cannot rule out some possibilities, probably best not to dwell on them, other than saying that they may be unlikely. You have argued, and I agree, that the complete absence of direct empirical evidence for God does suggest that the existence of divine intelligence is unlikely. However I think that is as far as one can go.
Dawkins: How much farther could one want to go? Unlikely is unlikely is unlikely. It is not the same as impossible, but science is replete with estimates of likelihoods that fall short of the demonstrably impossible. Global warming is highly likely to be happening and caused by human activities, but the alternative cannot be ruled out. It is very probable, but not certain, that the dinosaurs were killed by a large object colliding with Earth. It is almost but not totally certain that humans are closer cousins to chimpanzees than to gorillas. Just about everything we know in biology is supported by statistical evidence and is not totally certain. If you agree with me that the existence of divine intelligence is statistically unlikely, that is all I would claim. But I would claim that this low estimate of likelihood that we both agree upon is a scientific estimate, not something that is in principle immune to scientific discussion.
Krauss: Yes, but I don’t think the likeliness of God can be quantified in the same ways as dinosaurs, or global warming, and therefore it doesn’t make sense to spend a lot of time trying. Why work so hard when inevitably it is too slippery a subject, and has been for some time? In this regard, arguing that detailed probabilistic arguments that have been used to suggest that life is a rare phenomenon imply at the same time mathematical support for the non-existence of God is something that I don’t buy, since I cannot see how one can use physical arguments to restrict the possible existence of something that, by definition, supercedes the laws of physics.
Dawkins: Theologians resort to this definitional argument as their only defense against the statistical argument that we both accept. But why should we allow them this remarkably convenient get-out? Why should theologians be allowed to call the shots and immunize God against scientific scrutiny by a sort of definitional prophylactic injection? Suppose I were to say that the bolide that killed the dinosaurs 65 million years ago was hurled by Zeus. The data (iridium layer in the rocks, crater in Yucatan, etc.) is equally compatible with both the Zeus theory and the meteorite theory. Theologians of the Olympian school are at liberty to interpret the scientific data in terms of Zeus (and theologians of the Valhallan school are at liberty to interpret the same data as manifestations of Thor's hammer). Those theological concerns are by definition beyond the reach of science. You don't really believe that, Lawrence. So why allow Judaeo-Christian theologians to get away with evading the statistics issue in declaring, by definitional fiat, that their God is beyond the laws of physics?
Krauss: A valid point, but I think what most sensible theologians really argue is that the ‘intentions’ of God are beyond the laws of physics. Namely, if one could determine in detail the origin of the bolide that killed the dinosaurs, and demonstrate that it was in fact kicked out of its orbit around the Sun by the gravitational perturbation of the planet Jupiter, would this then demonstrate that there was no God, Zeus or otherwise? No, because God could have intended that life would evolve in an environment of random and rare catastrophes, that would help drive evolution forward.
Dawkins: That is strictly true, but I really do mean strictly. Why do you keep bending over backwards to be nice to superfluous and highly unparsimonious add-ons to science, when you would kick them out of the window if they were not protected by the label, “Religion. Handle with kid gloves for fear of giving offense”?
Krauss: For me it a matter of ignoring the add-ons rather than being nice to them, because I don’t see any debate as being fruitful. In a more general sense, arguing that religion is bad science merely invites those who want to introduce religion into science classes to continue harder to try and do so. But I believe it is essential to intellectually separate science and religion. It may be true that faith is not based upon reason, but this fact would only make it bad science if the claims of faith were in general falsifiable. As long as the tenets of faith go beyond reason, i.e. go beyond issues that can be settled by evidence or lack of evidence, faith lies in the realm of human activity that has little to do with reason. Going back to my earlier point, if this realm was restricted to religion alone one might have a good argument for trying to squelch religion. But, like it or not, it is a central facet of much of what it means to be a human. All of us share some characteristics with Lewis Carroll’s Queen, who believed six impossible things before breakfast each day. For most people religion is one way of making sense of an irrational world, a world that is not fair, in which human justice is an afterthought.
Read the full version here