How should we study religion?

Prospect Magazine

How should we study religion?


The philosophers Daniel Dennett and Richard Swinburne debate the correct approach to the study of religion

Dear Richard
5th January 2006

It’s high time science took a good hard look at religion. Why? Because it has become evident in recent years that if we are to make progress on the world’smajor problems, we will have to learn more about religions and the influence they wield over people’s lives and actions. Failure to appreciate the dynamics of religious allegiance, and the psychological impact of religious differences, may lead us to invest heavily in counterproductive policies. The phoenix-like rebound of religion in the former Soviet Union suggests to many that just as prohibition and the war on drugs have proved to be disastrous, if well-meant, attempts to deal with the excesses of these popular indulgences, so any ill-informed effort to rein in the fanatical strains of religion will probably backfire badly if we don’t study the surrounding phenomena carefully and objectively.

From a biological perspective, religion is a remarkably costly human activity that has evolved over the millennia. What “pays for” this profligate expense? Why does it exist and how does it foster such powerful allegiances? To many people, even asking such a question will seem a sacrilege. But to undertake a serious scientific study of religious practices and attitudes, we must set aside the traditional exemption from scrutiny that religions have enjoyed.

Some people are sure that the world would be a better place without religion. I am not persuaded, because I cannot yet characterise anything that could replace it in the hearts of most human beings. (Perhaps we should try to eliminate music while we’re at it. It inflames the passions and seduces many young people into wasted lives.) What people care about deeply deserves to be taken seriously. Exempting religion from scrutiny is actually a patronising way of declaring it to be all just fashion and ceremony.

Either we take religion as seriously as we take global warming and el Niño, and study it intensively, or we treat it as mere superstition and backwardness. As with the other marvels of nature, I find that paying scrupulous attention to its elegant designs increases my appreciation of it, but others may think that too much knowledge of the backstage machinery threatens to diminish their awe, to break a spell that should not be broken. This is not just a difference in taste, or a purely academic disagreement. Our futures may well depend on how we decide to proceed.


Dear Dan
10th January 2006

You think that religion needs rigorous scientific investigation. I agree, as would most religious people—in Britain, if not the US. I have been doing this for some decades, as have thousands of philosophers, historians and sociologists over the past 3,000 years. Welcome to the club. But you suggest in your new book Breaking the Spell that we should investigate religion “with the presumption that it is an entirely natural phenomenon.” Such a presumption needs to be justified. If there is a God, then all regular processes—codified by physics, biology, psychology or whatever—occur because of the sustaining activity of God and so are “supernatural” (even if God never intervenes). So the first thing is to investigate “scientifically” whether or not there is a God.

A scientific theory is rendered probable by its data in so far as 1) if the theory is true, the data will probably occur; 2) if it is false, the data will probably not occur; 3) it is simple; and 4) it fits in with our background knowledge of what happens in other areas of enquiry. But the fourth criterion tends to drop out when we are dealing with a large scientific theory (such as a general theory of all physical phenomena) for which there are few other fields of enquiry. Above all, criterion four drops out when we have a theory such as theism which purports to explain all data: that is, all observable phenomena. The most important of these observable phenomena which theism can explain is that there is a large physical universe, that everything behaves in it in a totally regular way, that the boundary conditions of the universe and the laws of nature are such as to lead to the evolution of human bodies, and that human bodies are connected to conscious lives.

There is not the slightest reason to suppose that these phenomena will occur unless a theory somewhat like theism is true—why should every atom in the universe behave in exactly the same way? (It is of course a “law of nature” that they do; but laws of nature are just the way things behave. They don’t explain them.) On the other hand, if there is a God of the traditional kind—omnipotent, omniscient, perfectly free and perfectly good—we have every reason to expect that he will bring about the existence of good things; and one especially good thing is the existence of embodied creatures such as ourselves who have a choice between good and evil and can influence the world and each other in various ways. The supposition that there is such a God is a very simple one. For it is the supposition that there exists one “person” (not many persons), who is the simplest kind of person there could be. A person is a being with some power to make a difference to things, some knowledge of what the world is like, and some degree of freedom as to which differences to make. God is postulated as a being in whom there are no limits to these qualities. (Scientists have always preferred theories postulating infinite degrees of qualities to theories postulating large finite degrees, when these are equally compatible with the data.) God’s perfect freedom means that there are no irrational influences deterring him from doing what he sees reason to do, that is what he believes good to do; being omniscient, he will know what is good and so he will be perfectly good. So by scientific criteria the data make it probable that there is a God. Given that, we should investigate religion on the presumption (now established by reason!) that there is a God.

In Breaking the Spell, you are very dismissive of arguments about the existence of God, and devote a mere seven pages to such arguments. The only one to which you give any serious attention is the ontological argument, in which the existence of God is logically derived from his perfection. But this argument is a unique attempt to prove the existence of God; it is a philosopher’s quirk and has no connection with the thinking of ordinary believers. You refer the reader to another book of yours—Darwin’s Dangerous Idea—which you claim gives more attention to such arguments. You do draw attention there to the fact that the constants of the laws of our universe and the variables of its boundary conditions had to be precisely what they are to within one part in a million million if human life were to evolve. But you go on to suggest that maybe there is an infinite number of universes, each with different laws and different boundary conditions, so that it would have been inevitable that in a few such universes human-type life would emerge. But a supposition that by mere chance there exists an infinite number of such universes, each varying in a precise way from the next, would be totally unscientific—an enormous violation of my criterion three. And if you are to suggest that there is some machine which generates such universes, then the question arises as to why the universe-generating machine is itself of just this kind—why didn’t it produce an infinite number of universes all of the same kind? It would have had to be a very special complex universe-generating machine to produce the right spread of universes, and there’s no reason why there should be a machine of exactly that kind rather than one of exactly another kind. But, on the other hand, if there is a person in charge of the universe who acts for the sake of the good, then we have plenty of reason to suppose that there would be a universe inhabitable by humans, rather than a chaotic one.

Having bypassed the discussion of this all-important question in Breaking the Spell, your main plea is for investigation of the origins of religion with the aid of biology, psychology and sociology. Given that there is a God, we would expect him to sustain a world where things behave in simple regular ways—for only in this way can finite embodied beings learn how things behave and so use the regularities of nature to make a difference to the world (agriculture, for instance). But we would also expect God to respond differently to the different ways in which we choose to behave, and so sometimes to intervene in the processes which he sustains. So we should investigate the phenomenon of religion on the “presumption” that it is a God-sustained movement in which God occasionally intervenes miraculously. That, I suggest, is the proper scientific approach.

With best wishes

Dear Richard
13th February 2006

I am proposing that religion should be studied as a natural phenomenon in the sense scientists understand, and therefore the first order of business is not whether or not there is a God, as you insist. As you say, theologians and philosophers have been debating this issue for millennia, but they haven’t produced any consensus to show for all that labour, and I wanted to open up new territory. Still, I agree that a book about religion as a natural phenomenon that never got around to considering the evidence for or against the existence of God would surely leave many religionists unsatisfied, which is the only reason I treated the topic at all.

You note that I devote a “mere seven pages” to such arguments, but anybody who is mainly interested in that topic can find references in those pages to my earlier discussions, and those of many others. Let’s, briefly, consider what you have to say about those earlier arguments of mine. In Darwin’s Dangerous Idea I considered various hypotheses by physicists and cosmologists about multiple universes and I claimed that these hypotheses, still untested, nevertheless nicely undercut the rhetorical question: “Look at the fine-tuning of the constants of nature! What could possibly explain that other than the existence of a foresighted, intelligent God?”

You say these hypotheses violate your criterion three of simplicity, but I utterly disagree. You imagine the hypothesis as positing “some universe-generating machine” and ask why it didn’t “produce an infinite number of universes all of the same kind.” This question is actually not as embarrassing as you suppose. To speak of a machine suggests design, but no such “machine” is required; in the absence of design, in an infinite time, all possible values would occur. Leonard Susskind, the founder of string theory, recently put it very simply and succinctly: “The same process that forged our universe in a big bang will happen over and over. The mathematics are rickety, but that’s what inflation implies: a huge universe with patches that are very different from one another. The bottom line is that we no longer have any good reason to believe that our tiny patch of the universe is representative of the whole thing.”

I sketch an account of how first folk religions and then their descendants, the organized religions, evolved out of our instinctual habit of adopting the intentional stance about anything puzzling or startling—our query “Who’s there? What do you want?” a habit that populates our heads with phantom agents, which then compete for rehearsal in our brains, with only the most unforgettable founding lineages. Each elaboration, each new adaptation, has to pay for itself in the evolutionary coin of competitive reproduction—of either genes or memes or both.

You counter my sketch of the evolution of religion with your own: “If there is a God of the traditional kind—omnipotent, omniscient, perfectly free and perfectly good—we have every reason to expect that he will bring about the existence of good things; and one especially good thing is the existence of embodied creatures such as ourselves who have a choice between good and evil and can influence the world and each other in various ways.” How convenient! Who could ever imagine anything more wonderful than us? As William James so memorably put it, more than a century ago: “Religion, in short, is a monumental chapter in the history of human egotism.” You imagine a “good creator” God who made sure things were regular enough for us finite beings to understand, so that we could eventually discover agriculture, and even hit upon the idea of the God who was behind it all, and who “occasionally intervenes miraculously.”

Presumably this same foresighted creator anticipated the amusement the unbelievers would feel when contemplating the recent declarations by Pat Robertson to the effect that Ariel Sharon’s ill health was God intervening to punish him for ceding Gaza. I’m sure you’ll tell me that our expectations about what a good creator would want, and do, don’t extend to such particulars as these, but why are your expectations any better grounded than mine? You haven’t told us what the rules of this game are.

If your hypothesis about the rise of religion is a rival to mine, there ought to be some empirical details on which we differ, something you can explain and I cannot, something you predict that I don’t, or something that refutes my sketch. I don’t claim to know it is true; on the contrary, I insist that it is just a sketch, presented so that we can have something on the table that we might try to fix, or replace by a better scientific alternative. Do you have any proposals for a better theory?


Dear Dan
16th January 2006

Some sort of multiverse theory might well be true. My point was that if there is a multiverse, it’s a multiverse of a kind which will produce at least one universe which will produce humans. But it’s logically possible that there might instead have been other quite different kinds of multiverse, or just one universe, not productive of humans. So why are the most general laws of the multiverse as they are? Why do all particles behave in exactly the same way as each other, so as together ultimately to produce human life? This enormous coincidence in particle behaviour requires explaining. I’ve got a good theory which explains it; you haven’t. And if you are really telling me that the production of humans is not, objectively, a good thing, I find myself wondering if you really mean something so implausible.

The relevance of having good reason for believing that there is a God to your project of explaining the origin of religious practice is as follows. It is an important scientific principle that in assessing the evidence for the causation of some particular phenomenon we should take into account our “background knowledge” by way of general theory about how the world works (criterion four of my previous letter). For example, suppose you are an astronomer investigating some pattern of dots of light in the sky, and you see that the simplest explanation is that these dots are the debris of a star explosion. Then it is right to believe that that they are so caused—unless you have a well-established theory of physics which says that stars can’t explode. In the latter case you should adopt a less simple explanation for the dots of light, unless the explosion theory is so simple and so well able to explain many data that you give up your whole theory of physics.

Similarly, if there is good evidence to suppose that there is a God, then there is good evidence to suppose that the processes by which religious practices developed are ones sustained by him, and so—since he is good and so will want occasionally to interact with humans—ones which he may occasionally set aside. Then this background theory will lead us to attribute some event about which we have some historical evidence to the intervening action of God, when given a rival background theory it would be wrong to do this. Now I am inclined to think that since we have so little historical evidence about the early stages of religious practice, any theories about this are pretty speculative. But I don’t wish to deny that some of the features to which you draw attention in your book (shamans, and looking for animistic explanations) operated in developing religious practices. But theism also leads me to suppose that God will have intervened from time to time to help humans to be aware of the divine and of what a good God wanted them to do.

My point is, however, crucially relevant when we come to the later period. Consider the origins of Christianity. It is pretty well agreed that the belief of early Christians in the bodily resurrection of Jesus was crucially influential in their commitment to that religion and to its widespread propagation throughout the world. What caused that belief? My view is that the resurrection of Jesus caused the belief in the resurrection of Jesus. You, I suspect, have a different view. There’s quite a lot of evidence from witness testimony that Jesus did indeed rise. But if I believed that laws of nature are the ultimate determinants of what happens, I would not believe that Jesus rose, because I know that the laws of nature do not allow humans to come to life again after 36 hours. But as I have good reason to believe that laws of nature only operate as and when God allows them to, and I believe that on this occasion he had good reason to set them aside in order to put his signature on the teaching of Jesus, I have good reason to suppose that Jesus rose.

Our respective background theories rationally determine how we assess the historical evidence. You ask why are my expectations about what a good creator would do better than yours? Well, I give reasons which try to take off from moral views which you might share—say the goodness of parents interacting with their children, and the goodness of people having free choices and responsibility for others. If you agree with me so far, then I try to show what follows for what an omnipotent and perfectly good God would do. If you don’t concede my initial moral views, then I appeal to examples where you might see that my moral views are plausible. Our moral understanding thus progresses by what John Rawls called the method of “reflective equilibrium.”

That brings me to what you have to say in Breaking the Spell about morality and religion.You suggest that we might investigate scientifically whether religion makes people morally better. The trouble is that the content of morality, and so what it is for someone to be morally good, is affected in part by whether or not there is a God and what he has done. While theists and atheists might agree to quite an extent about what are the necessary truths of morality—that it is good to feed the starving, say, or that parents ought to educate their children—some such necessary truths have different consequences for the theist and for the atheist. Among the necessary truths of morality are, I suggest, that it is good to reverence good people, and obligatory to show gratitude to and please our benefactors. So if we are created by a perfectly good God and sustained in existence by him every moment of our lives, we have an obligation to reverence him greatly and show him gratitude—that is, to worship him and to please him in other ways.

Pleasing God will involve obeying any commands which he has given to us. He will issue commands by giving his signature to some teaching, as I believe that he did to the miracle of the resurrection. And God may also in the same way reveal to us necessary moral truths which we are not clever enough to work out for ourselves—about abortion and euthanasia, for example. So again, as with the historical story of religion, so also with whether religion makes us better people, our conclusions will be much affected by whether we have good reason to believe that there is a God and other claims of religion. There’s no avoiding those issues.

With best wishes

Dear Richard
21st January 2006

You and I have different ideas about what constitutes a good explanation, and what needs an explanation at all—and this fact itself needs to be acknowledged before we can engage in a constructive discussion. You find it improbable that there would be a multiverse of all physically possible universes, including ours. Is it less improbable than that there would be an omnipotent, benevolent universe-creator? I don’t think so, and here Bayesian probability theory gives no leverage, so far as I can see. Both are mind-boggling prospects—but that doesn’t give yours the edge. As Philip Morrison observed, the prospect that we are alone in the universe is mind-boggling, but so is its denial, and one of them has to be true. From my perspective, your imaginative attempt at an inference to the best explanation is telling for the one thing it lacks: a single striking prediction. That’s why it can’t be taken seriously as a contender against a purely secular and materialist theory of cosmic and biological and cultural evolution. Compare: “Wouldn’t it be cool if our whole visible universe was actually a sort of zoo exhibit for a race of extraterrestrials who get a kick out of watching us, and there’s a super-engineer out there who is in charge, and who signals us every now and then by turning the local laws of nature on and off!” You will have to admit that this is logically possible, but before we have any reason to take it seriously, anyone proposing it would have to describe some telltale signal predicted by the theory. I’ll make one up, to show that this is not impossible: my interpretation of what I take the earlier signals to have been leads me to the hypothesis that the super-engineer has planted hints to the effect that we will find his signature (“Kilroy was here”) if we drill a hole exactly a mile deep from the summit of Everest. We do, and there it is. Wow. It’s not just that I make the prediction, but that I use my theory to make it. Science, including evolutionary biology, offers so many such predictions that they don’t strike us as remarkable any more (grab any bird from the air anywhere on earth, and any fish caught in any ocean, and I will tell you some specific similarities and differences you will find in their DNA). Your Christian hypothesis has nothing like that to offer, only retrospective interpretations. It is relatively easy to make such interpretations cohere, and the history of revision of Christian doctrine offers many examples of such improvements—a “just so story” indeed.

I do agree that “the production of human beings” is a wonderful thing, but I suspect that there are other possible beings (in some philosophical sense of “possible” that you and I can accept) that would be still more excellent. So I see no reason to go along with your hypothesis that we’re just what to expect from a perfect and omnipotent creator.

Finally, your claim that we need to factor in God’s existence or non-existence before we can empirically assess whether religion makes people morally better than they otherwise would be raises a problem, but for you, not for me. I’ll grant for the sake of argument that if God exists and commands reverence, then irreverence is morally bad and worship is morally good, but in any reasonable political attempt to achieve global consensus, this claim would have to be leveraged later off the points we can all pretty much agree on at the outset: murder, rape, torture, theft, perjury… are (other things being equal) morally bad. Maybe eating pork, or dancing, or saying God’s name is morally forbidden, but persuading us all of any such claim is a task facing those who so believe, not something that they can announce as a fixed point at the outset—since it amounts to unilateral refusal to engage in reasoning from common ground. If God commands you to worship him, then it’s your problem to convince the rest of us that we should hear this command as well—and you would dishonour your God (I would think) by refusing to explain why this is so. This leaves you with a stark choice: preach to the converted in your own flock and abdicate responsibility for convincing (not terrorising) the rest of the world to see it your way, or accept that you have to begin from the neutral position that presupposes no God and prove to us that you’re right.

I’m all for the latter option, of course. And in the meantime, I note that on the matters of morality on which there is something approaching a global consensus, the evidence to date does not show that adherence to a religion—any religion—has a positive effect on the probability that a person behaves morally.


Dear Dan
30th January 2006

I don’t think that it is in any way important that science should make predictions. What is important is that science should make probable the occurrence of certain phenomena which we observe, when their occurrence would not be probable otherwise. What is unimportant is whether we observe these phenomena before or after we formulate our scientific theory. Newton’s theory of mechanics made no new predictions which could be tested for the next 50 years. But it was accepted as an obviously true theory of mechanics because it gave a simple explanation of a vast range of phenomena previously known—movements of planets, movements of moons, behaviour of pendula, behaviour of tides and so on.

Certainly, one can always devise theories which “cohere” (in the sense of “are logically consistent”) which will lead one to expect known phenomena; but it is very difficult to devise a simple theory which leads you to expect the known phenomena. If you can, that is very strong evidence that the theory is true. Another example to illustrate this point is a detective story, in which the detective learns all the evidence in the opening chapters, but only solves the crime by finding, in the last chapter, a simple explanation of all that evidence; and when he has got one, he doesn’t need a new prediction to render his explanation probable. So the fact that we know the phenomena which I cite as evidence for theism before we consider the theory of theism has no bearing on their evidential force. All that is relevant is how well theism satisfies the criteria which I listed in my first letter. The criterion of simplicity dictates that we should postulate one simple entity responsible for innumerable other entities when these others behave in exactly the same way as each other. It would be absurd to dig up a whole hoard of coins in a Roman villa with exactly the same markings on them, say “What an interesting coincidence!” and leave it at that. We should look for a single cause: that they were all produced in the same mould. If there is one universe or many universes governed by the same general laws, this consists in an enormous number of fundamental particles behaving in exactly the same way as each other. It is deeply unscientific not to look for a simple cause.

If it were the case—and I don’t think you have adequate evidence to show this—that serious commitment to a religion has no “positive effect on the probability that a person behaves morally” this would indeed be surprising—since almost all religions put quite a lot of effort into dissuading people from your “murder, rape, torture, theft, perjury.” But if this were so, I can think of two possible explanations. The first is that the teaching methods which religions use are counterproductive; and if we learnt that this was so, then of course the methods could be changed to good effect. Alternatively, it might be the case that those more inclined to “murder, rape…” are those who are so aware of their inclinations that they feel a need for a God to forgive them and help them, and so are more open to the claims of religion. (Jesus seems to have thought this—see Luke 7:40-43.) But that is compatible with religion making each person more moral than they would be otherwise. And it does not imply that those whose wrongdoing is less obvious are less in need of God’s help and forgiveness. I do of course agree with you that, if I am to persuade you that there is a God, I have to begin from a neutral position that presupposes no God and prove to you that I am right. I’m trying hard!

But the same applies in reverse—you have to begin from a neutral position that does not presuppose that there is no God and prove to me that you are right. And, to repeat, my worry about your book is that it advocates investigating the origins and consequences of the practice of religion with an atheistic presupposition. But with the need for rational argument, we are in total agreement.

Best wishes

  1. September 22, 2009

    cosley sathekge

    yes i do agree with the fact that religion should also be studied as nature but not creating theories that scientists create to refute the nature of God.

  2. July 23, 2010

    Philosopher of Religion


    For the record, Swinburne is not a theologian. He is a philosopher of religion (one that happens to be a theist).

  3. August 16, 2011


    as with the historical story of religion, so also with whether religion makes us better people, our conclusions will be much affected by whether we have good reason to believe that there is a God and other claims of religion. There’s no avoiding those issues.

  4. April 4, 2014

    Pat Hayes

    It saddens me that Dan Dennett is wasting his valuable time having such an utterly pointless conversation with such a fool. Richard claims to have all kinds of ‘explanations’, but never actually gives us any. Take just one: the reason that all atoms (of a given element, I presume) are alike is because God fixed this so that moral human beings could exist. This is a complete non-sequitur. Why couldn’t moral creatures exist in a universe made of dissimilar atoms? We have no idea what such a universe would be like, but I see no clear reason why consciousness could not exist in it. And why, for that matter, would a good God wish to create other conscious beings, in any case? Having other consciousnesses exist does not seem to be particularly morally good or beneficial in itself, given that God is already (by hypothesis) perfectly perfect. And so on: all through these absurd responses there are no actual clear arguments at all, only empty verbiage crafted to fit a narrowly Christian doctrine. A waste of time to read, and a waste of time to have written.

Leave a comment

  1. The Divine Conspiracy Blog » Blog Archive » Studying Religion10-02-09


Daniel Dennett

Daniel Dennett is a philosopher, cognitive scientist and author of the forthcoming “Intuitions Pumps and Other Tools for Thinking” 

Share this

Most Read

Prospect Buzz

  • Prospect's masterful crossword setter Didymus gets a shout-out in the Guardian
  • The Telegraph reports on Nigel Farage's article on Lords reform
  • Prospect writer Mark Kitto is profiled in the New York Times

Prospect Reads

  • Do China’s youth care about politics? asks Alec Ash
  • Joanna Biggs on Facebook and feminism
  • Boris Berezosky was a brilliant man, says Keith Gessen—but he nearly destroyed Russia