Why Science Cannot Address the Existence of God

Steven Dutch, Natural and Applied Sciences, University of Wisconsin - Green Bay
First-time Visitors: Please visit Site Map and Disclaimer. Use "Back" to return here.

There are a lot of arguments claiming that science cannot prove or disprove the existence of God. Many of these can be summed up in a theme elaborated by Stephen Jay Gould that science and religion have mutually exclusive spheres of authority, or as he called them, "nonoverlapping magisteria." The problem with this approach is that science and religion can't agree on the boundaries and therefore, even if the magisteria don't overlap in reality, they do in practice. More fundamentally, the division assumes a distinction between a world of fact (the domain of science) and a world of subjective belief (religion), but no religion regards its dogmas as subjective belief. Every religion regards its dogmas as objective, demonstrable facts.

One theological argument against definitive proof is that a definitive proof would violate free will by forcing people to believe. This approach has a number of flaws. For one thing, it makes pretty specific assumptions about the existence and nature of God. Using this approach, you can only claim the existence of God is unprovable if you already believe in God. Then too, free will is free. The history of pseudoscience shows unmistakably that even the most convincing evidence still leaves people free to deny it. Even if God came down from the sky and walked across the English Channel on global television, many people would rationalize the event away (it wasn't my god, therefore it must be fake; it was a computer-generated hoax, etc.). Thus, there is good reason to doubt that even the most overwhelming demonstration would violate free will. Regardless how convincing the event was at the time, it would soon be dismissed as legend, hallucination, or hoax. Finally, it has not always been agreed among religious believers that free will even exists; many Protestants, particularly Calvinists, have doubted the existence of free will.

Then there are psychological issues. Militant atheists accuse believers of immature wish fulfillment. Militant believers accuse atheists of petulant defiance of authority. People who subscribe to either view are not likely to be persuaded by anything the other side says. And to the [considerable] extent that the accusations are true, people whose attitudes are armor plated won't be open to any argument that challenges their deeply held beliefs. So apart from the honestly held and debated evidence concerning the existence of God, we have to navigate a thicket of intellectual dishonesty. Some of the ideas advanced by people with ideological axes to grind have merit, others are merely specious. Then there's the whole issue of competence. Some arguments on the existence of God are the result of deep and informed reflection, others are embarrassingly juvenile and ill-informed.

But apart from the matter of domains of applicability, theological assumptions, and psychological factors, there are other reasons to doubt that science will ever be able to say anything about the existence of God. These reasons are grounded in the nature of scientific proof as applied to complex reality. The more complex a question becomes, the more likely it is that there will be numerous interpretations that fit the data equally well, the harder it is to isolate relevant evidence, and the more complex and indirect the proof must be. The more complex the evidence and the logic, the greater the likelihood that fallacies of reasoning and faulty data might contaminate the results.

One final note: I'm not talking here about moral or esthetic arguments for or against the existence of God. This is entirely about the use of empirical data and scientific methodology.

The Nature of Complex Reality

Let's consider a question that can be addressed using scientific methods. Why did the United States have high crime rates in the 1970's and 1980's, then experience a precipitous drop in crime rates in the 1990's?  Among the answers proposed to this question have been:

  1. Tougher sentencing, more police and longer prison sentences have had a deterrent effect, as well as removing criminals from circulation.
  2. The legalization of abortion in 1973 meant that a lot of babies who would otherwise have been raised in dysfunctional environments were not born, meaning that twenty years later there are fewer angry young adult males to commit crimes.
  3. The outlawing of leaded gasoline at about the same time eliminated an environmental trigger of violent and aberrant behavior.
  4. The end of the Baby Boom meant fewer babies born beginning in the 1970's, so there were fewer young adult males (who dominate the crime statistics) beginning in the 1990's.

There is no problem about gathering data on this question by scientific means. The data are abundant and in forms suitable for standard statistical analysis. It's deciding which data are relevant and what the data mean that gets sticky.


In science, it's fairly easy to establish facts if you can control the circumstances. The classic way to do this is the controlled experiment. If you think two atomic particles interact in a certain way, two chemicals undergo a specific reaction, or a certain drug cures a given illness, you set up an experiment to see what happens. Especially if you can predict in advance what you hope to see, and it actually happens, then you can be highly confident of your ideas. This method works because the experiments can be repeated at will, and the system can be reduced to simple terms so that the phenomenon being tested can be studied in isolation with extraneous factors eliminated. If we suspect any error, fakery, or overlooked factors, we can run another experiment to check. Also, repeatability safeguards against drawing the wrong conclusions merely because some experiment turns out a certain way just by chance.

If you can't control the circumstances but you can observe a lot of instances of a certain phenomenon, that's also a good way to establish facts. No two hurricanes, eclipses, comets, volcanic eruptions, or mountain ranges are exactly alike but they are repeatable in the sense that they all have features in common, and if you observe enough instances, you can start to draw conclusions about how these phenomena work. We can conclude that all hurricanes have eyes, all eclipses of the sun occur at new moon, all volcanic eruptions are powered by gas pressure, and all mountain belts have thrust faults. Furthermore, you can make predictions about what you expect to see the next time you observe that particular phenomenon. For example, when a supernova erupted in a nearby neighbor galaxy in 1987, astronomers were delighted to see pretty much everything their theories predicted happen on schedule.

At least in principle, all the hypotheses for the drop in crime rates could be tested in controlled ways. We just back off on judicial rigor, prohibit abortion, have another 70 million babies in a 20 year period, and reintroduce leaded gasoline. We'd have to test each of those variables in isolation, and since three of them link a change to events twenty years later, it would take two decades at least to run one test. To be really sure of our conclusions, we'd want to see the same results in several runs. It would take centuries. And the costs would be unacceptable. Would we really want to trigger an increase in crime deliberately? Alternating twenty years of legal abortion with twenty years of prohibition would alternately be intolerable to anti- and pro-abortionists. And we added lead to gasoline because it seemed to offer better performance without adverse effects. Now that we know leaded gas is harmful, it would be morally out of the question to run an experiment to see if leaded gasoline leads to a rise in crime. And who in their right minds wants another Baby Boom? Even if we were insane enough to try to repeat any of these experiments, so many other variables would have changed in the meantime that we'd be hard put to be sure we had really successfully duplicated the results. So in practice, there is no way to do a conclusive experiment to settle the matter.

It's when you can neither control the circumstances nor observe a large number of similar phenomena that things get dicey. In the case of the drop in crime rates, we have a unique event and the only way we can establish why it happened is to examine all the possible causes and evaluate the evidence for each. How do you establish cause and effect for a one-time event? Well, if you've got a good theoretical understanding, it's pretty easy to establish cause and effect. We know hurricanes create storm surges, so even though Hurricane Katrina was unique in flooding the only major American city that lies extensively below sea level, we don't need to look for exotic causes. Everything we know about hurricanes and storm surges tells us that Katrina was entirely sufficient to do the observed damage. The 2004 Indian Ocean tsunami was by far the worst tsunami disaster ever, but it originated the way all other tsunamis did and behaved just like all other tsunamis. It was just unusually big and hit unusually populated areas.

But what if we have a one-time event but we don't have a good theoretical understanding? In many cases, we can establish that an event happened even if we have no explanation why. For example, we know that the English left a colony on Roanoke Island in 1587 and it wasn't there a few years later. We still don't know why it vanished, but that it vanished is beyond doubt. Sometimes the event can be broken down into components that can be successfully analyzed. We only have one moon, and for a long time none of the available theories of lunar formation stood up to scrutiny. All of them had fatal flaws. In the last twenty years, studies of planetary formation, lunar chemistry, and impact mechanics have persuaded a large number of planetary scientists that an oblique impact between a Mars'sized protoplanet and the proto-earth can account successfully for the moon's formation. In the case of the crime question, it's clear that crime rates fell even if there's no clear consensus why they fell.

It seems very difficult to imagine any general theory in the physical sciences that can't eventually be tested conclusively. It may take huge amounts of resources and a very long time, but since we have physical theories for many phenomena and know how to go about developing more, it seems likely that any theory we can propose can eventually be settled. The fact that we can keep revisiting a question until we get a workable explanation makes all the difference. Historical sciences like geology pose the problem that they deal with events that were not witnessed, and so we have to attack problems by examining physical evidence. If there is no physical evidence, it may be impossible to answer some questions. For the most part, we will probably never know where ancient rivers ran or why some given species became extinct or what exactly the earth's plates looked like two billion years ago. We will probably never identify any of the stars that formed in the same cluster as our solar system, or know what the constellations looked like in the days of the dinosaurs.

Yet all of those problems are, in principle, solvable given the right information. The question why U.S. crime rates fell in the 1990's is of a different order of complexity. There's no question of the availability of data, but there are no theories to guide us in interpreting the data. Even the most basic questions about the functioning of society are contentious. There's disagreement about whether tough penalties deter crime or not, whether social spending reduces poverty or creates dependence, whether a strong or weak military is more likely to promote peace. When we get to history and the social sciences we often deal with unique events where there is neither consensus nor any prospect of testing hypotheses.

For example, did the cooling event called the Little Ice Age contribute to the spread of bubonic plague in the mid-14th century? There's a plausible causal chain: cooling would result in poor harvests, reduced resistance among the population, and possibly ecological changes that would bring rats, which carry the plague, into more direct contact with humans. The last major plague pandemic happened in the 1660's, and the cessation of plague has been attributed to the host species of rat being displaced by another species. There's no way of testing these hypotheses conclusively. They seem plausible or at least possible, and they have the virtue of being rather harmless. Our view of the Middle Ages won't topple if they're proven wrong. (And to forestall comments, there's a school of thought that a virus akin to Ebola, not bacteria, was the agent that caused the plagues of the Middle Ages.) On the other hand, the question why crime rates dropped in the 1990's is not harmless; it has serious implications for social policy.

One way to try to get a handle on problems like this is to use statistics. If you break an event down into separate subevents, or look at the event as it played out in different places, that's sort of like making multiple observations. So, to return to our crime question, we could compare the drop in crime rates in different states or counties with differing criminal penalties, abortion rates, birth rates or use of leaded gasoline. The researchers who suggested the leaded gas link were impressed with the fact that they found correlations in countries with differing dates of abolition of leaded gasoline and different judicial systems. Statistics can strengthen our confidence in a theory, but social scientists are fond of pointing out that "correlation is not causation." (We should also point out that "correlation is not causation" is not disproof of a causative link, as many social commentators seem to think.) We can equally validly point out that discrediting statistics is not necessarily disproof of a theory. It merely means we can't be as confident of that particular line of argument. Also, it's commonly overlooked in science that while statistical methods are completely objective, the criteria used for deciding whether the results mean anything are not. The decision to use one or two standard deviations, or 99% confidence, as the criterion for deciding whether something is statistically significant is quite subjective, and dependent on the problem. A 70% correlation coefficient may be stupendous in the social sciences but hopelessly sloppy in physics.

Historians are fond of saying "history has no predictive value," a statement that is about 80 per cent true. If history had no predictive value, there'd be no point in studying it. After all, apart from a few delusional compulsive gamblers, nobody studies old winning lottery numbers because they have absolutely no predictive value even if they may have won millions for the lucky ticket holders. We do study history because we can glean some predictions out of it. The U.S. got involved in the former Yugoslavia very reluctantly because history shows clearly that occupying Yugoslavia may be easy, but subduing it may be another matter altogether.

Multiple Factors

To return to the crime question, it is distinctly possible that all the proposed causes for the drop in crime are correct. It may well be that deterrence and removing criminals from circulation, legalization of abortion, outlawing of leaded gasoline and the end of the Baby Boom all acted jointly to lower crime rates. An additional complication when multiple factors are involved is synergy - two agents acting together have a stronger effect than the sum of their individual effects. So it's entirely possible that statistical tests of any of the proposed causes in isolation might be inconclusive, but all four acting together could have a powerful effect.

Motives For Advancing Hypotheses

Some of the hypotheses for the drop in crime rates were not advanced with the purest of motives. Although the motives of someone for holding an idea never - repeat, never - constitute evidence, they can raise red flags. Motives can alert you to the potential for hidden fallacies, selective use of data, maybe even outright fakery. You are never justified in rejecting an idea solely because of the motives of the person presenting it, but you are entitled to give intense scrutiny to the evidence and logic.

The baby boom hypothesis is probably the most innocent. Since young adult males dominate the crime statistics, a natural approach is to ask what happened when they were children. In a quick scan of history twenty years before the crime rates began to fall, the end of the Baby Boom leaps out. So it's a simple and sensible hypothesis. One way to evaluate it would be to see if crime rates track population growth with a time lag of about 20 years. But major population dips and rises are rare, and there are many other factors that could affect the results. On the other hand, this explanation might supply a ready rationalization to someone who is ideologically predisposed to reject any of the other hypotheses.

The legalization of abortion would also leap out at anyone looking for significant societal events from the early 1970's. But this hypothesis has a powerful appeal to anyone looking for an argument in favor of abortion.

The hypothesis that tougher sentencing has had a deterrent effect makes sense, but also has a powerful ideological appeal to social conservatives.

The leaded gas hypothesis is also fairly innocent. The researchers noted that leaded gas was phased out in the 1970's, were intrigued enough to look at crime statistics from other countries, and found a widespread correlation. But this hypothesis might appeal to people who want an alternative to some other proposed cause. Opponents of tough criminal laws might latch onto this explanation to argue that it's a viable alternative to the deterrent hypothesis, and opponents of abortion might seize on it to discredit the abortion hypothesis. And it's a wonderful argument for environmentalists.

So, apart from genuine conviction, people can embrace hypotheses either because they support what they want to believe, or offer an alternative to something they don't want to believe. Thus we have the complication that people can believe the right things for the wrong reasons.

Motives For Rejecting Hypotheses

Most of the debate over the drop in crime rates seems to have had less to do with advancing an explanation than finding reasons not to accept some proposed explanation. People who otherwise aren't concerned with why crime rates fell become very concerned when the drop is credited to some cause they reject. The motives for rejecting hypotheses, apart from genuine conviction, mirror those for acceptance: people either don't want to believe the hypothesis or want to believe some alternative.

In science, the motivation issue tends to be settled by the bystanders who don't have a deep emotional investment, but who do have a vested interest in their own disciplines. I don't care one way or the other how astronomers sort out the Big Bang, dark energy, and the like. I would start to care if they came up with something that radically revised the age of the earth. They in turn don't care how far back we can trace plate tectonics, but they'd get very interested indeed if we recalculated the age of the earth and doubled it. That would play havoc with their models of stellar evolution. So regardless how vehemently an outsider like H.C. Arp argues that galaxy red shifts don't always correlate with distance, I will trust the astronomers to sort it out, as long as their conclusions are useful, or at least harmless, to my own field.

An important group of bystanders are scientists in the pipeline as undergraduates, graduates, and junior scientists. They may be intrigued by and dabble in some of the off-brand theories, but they don't have the emotional attachment that the originators of those theories, or defenders of the status quo, have. So when conflicting evidence comes along, they take it with a lot more equanimity. They tend to gravitate to the position best supported by the evidence, so the alternatives either linger on as marginal theories or eventually die out.


It may never be possible to resolve exactly why crime rates in the United States dropped in the early 1990's. Even if there is really only one correct explanation, there may be no way to establish its correctness using any known methods. There may be many equally valid ways to interpret the evidence. "Valid" here doesn't mean as seen by some hypothetical omniscient observer. "Valid" means that using the best available evidence and the best available reasoning and analysis, there may still be mutually incompatible explanations that fit the data equally well. It's like watching replays of a controversial football play. Some angles suggest that a receiver had control of the football when he hit the ground, others don't, and two equally skilled referees might differ on the call they make. And even though, on the whole, football referees make remarkably good decisions, even the best available evidence and training doesn't guarantee there won't be occasional errors.

The Nature of Gods (or Other Beings)

Note that the word "god" (capitalized or not) has not been mentioned since the introductory section. Everything since then has been about the problems of proof when evidence is ambiguous, theoretical frameworks are lacking, and there may be several interpretations that explain the data equally well.

The only question about a god that is meaningful or interesting is whether or not there is a god who interacts with the universe. Pantheism, the idea that the sum total of everything that exists is a god, is trivial. Deism, the idea that a god created the universe but does not interact with it, is of no imaginable interest or relevance.

In practical terms, deciding the existence of a god amounts to testing for the existence of some rational and extremely powerful supra-human being or beings. Whether it's a single infinitely powerful deity, a number of finite but still powerful supernatural beings, or a powerful natural alien civilization, the practical problems of evidence are the same. Whether or not a god exists may not be testable by scientific methods, but there are very similar questions that science certainly can address. Suppose, instead of asking whether there exists an infinite, omnipotent deity acting by supernatural means, we ask if there is some very powerful intelligent Entity interacting with our planet. The Entity need not be infinite or omnipotent, merely far more powerful than we are. The Entity need not act supernaturally, but merely by means of natural laws we have not yet discovered, or technology we have not yet developed. To free ourselves from any distractions imposed by the supernatural, let's consider the hypothesis of a purely natural, but extremely powerful and knowledgeable Entity.

What's the Entity's agenda? It could be malevolent or sadistic, seeking to harm us or cause prolonged suffering rather than destroying us outright. That might be a plausible explanation for war, famine, and disease. Or it could be dispassionate, watching to see how long we can avoid destroying ourselves. But let's consider only the possibilities that the Entity is benign and actively trying to help us. Furthermore, it is knowledgeable enough about human psychology and the workings of our planet that clumsiness or ignorance are not an issue. If it had tried to prevent World War II, for example, it would not have inadvertently triggered some worse alternative history. It would not, say, have gotten Hitler into art school only to have Stalin conquer Europe. And remember, this Entity is not a deity; it is merely a very intelligent and very powerful, but 100 per cent natural, being.

Why wouldn't the Entity reveal itself? We can easily think of a host of reasons why an intelligent super-being might conceal its existence or make it non-obvious. It might, of course, have ulterior or selfish motives, and the epistemological issues get really interesting if we allow that there might be multiple entities with conflicting agendas. But let's restrict ourselves to the case of an Entity without ulterior motives (or at least ulterior motives harmful to us). There will be some reasons that make sense to us. Then we have the possibility that the Entity has reasons that make no sense to us or that we cannot comprehend. But let's stick with the answers that we can comprehend, since those are the only ones we can discuss profitably.

  1. If the Entity openly reveals itself, then what? Will it continue to respond openly to every request for information or guidance? Will it respond to requests to endorse one nation's agenda over another's, or one political party or social class over another? Will it respond to requests to use its advanced powers to eliminate disease or aging, stop environmental degradation or eliminate poverty? Will it respond to requests for information about itself, or explain why it does or does not act in certain ways?
  2. If the entity does keep responding after initial revelation, would it soon end up micromanaging the world or destroying human initiative? Would so many humans become helpless dependents or lapse into nihilism that even those who retained their initiative found it impossible to get anything done? It may be logically impossible for the Entity to reveal itself openly without scrambling the world beyond repair. Open revelation might perturb human society in ways that conflict with the Entity's aims.
  3. If the Entity revealed itself openly but thereafter remained silent, might we not after a few centuries dismiss the event as a fabrication or a myth? And if that's the case, how can we be sure events we now ascribe to myth weren't in fact real?
  4. The Entity may be communicating all the time, but since we are constantly immersed in the signals, we simply perceive them as natural background noise. Fish don't know that they're wet.
  5. The Entity may be deliberately selecting for those individuals who can detect its communications and who choose to interpret its faint signals as communications.
  6. The Entity may communicate only with specific individuals on specific occasions, for its own specific purposes.
  7. Whether it's secular philosophies like Marxism, Naziism, radical environmentalism or militant atheism, or religious movements like the Inquisition, radical Islam, or Christian fundamentalism, people who believe they have an objective basis for dictating the behavior of others are the most spiteful, vindictive, and downright evil people on the planet. Just imagine what would happen if certain people had real objective evidence for their beliefs. Maybe the Entity wants to protect us from such people.

If the Entity doesn't communicate unambiguously, can we still somehow test for its existence somehow? It seems pretty clear that a controlled experiment is out of the question:

  1. The Entity is not likely to submit to human control. It will not respond to demands that it perform replicable actions merely to meet our standards for verification. The Entity knows it exists. Why should it care whether it meets our criteria? "Wow, I don't meet your criteria for proof? I'll slink back to the mother ship, curl up into the fetal position, and suck my thumb for a while."
  2. To use terms that seem crass from a human perspective, the Entity may find it demeaning to be expected to perform on demand in an experiment, or to submit to human criteria for proof, or justify its decisions.
  3. The Entity will certainly not reveal itself in ways that allow that knowledge to be used against it, say to overpower it or block its actions. Even the idea that there is a benevolent Entity might inspire humans to be complacent about pressing problems, secure in the belief that the Entity will rescue them. People might use the existence of the Entity to rationalize their own selfish actions by claiming to do them in the interests of the Entity, or rationalize their desire to dominate others.
  4. What about people willing to cooperate with the Entity? The Entity may communicate freely with them, only to have their claims dismissed by others as delusional or mistaken. The Entity may choose not to communicate with them, or communicate only sparingly, because open communication might influence their behavior adversely. They might, for example, come to consider themselves elite or privileged. Or the Entity may not be able to communicate openly with them because doing so would furnish information to people with ulterior motives.
  5. Since there are a host of sound reasons why the Entity might not reveal itself openly or participate in an experiment, that leaves us with observing nature and trying to tell whether or not there is some pattern in nature that we can detect. From the Entity's perspective, it may be manifesting itself in ways so obvious that only a very rudimentary intelligence could miss it.
  6. Humans are extremely ingenious about ferreting out even faint patterns. If the Entity is concerned about ways that knowledge of its existence might be misused, it might manifest itself subtly enough to defeat even sophisticated statistical tests. Or it might use the inherent ambiguity of complex reality to camouflage its actions. Maybe its version of an intelligence test is to see whether some people choose to see evidence of its existence or not.
  7. Is it superfluous to add that irrational actions people commit because of their belief or non-belief in the Entity have nothing at all to do with whether or not the Entity actually exists?

And those are the problems facing us in trying to determine the existence of a completely natural Entity. Maybe someday it will slip up, nod off at the wheel, and some interplanetary probe will sneak up on its base undetected, and then we'll know. Or maybe it will decide the time has come to communicate overtly. Or maybe not.

The Nature of People

Let's imagine two scientists, both highly competent, working on the intelligent entity question. On the one hand, data set after data set has been number crunched and nothing conclusive has emerged. Some runs come out in favor of an entity, others against, and the grand average is statistically indistinguishable from zero. On the other hand, there are nagging subtle things that just hint at patterns. There are events that could have turned out disastrously except that a whole concatenation of unlikely things got in the way, or coincidences so intricate that appealing to the laws of chance is unsatisfying.

Looking at the statistics, there's no conclusive evidence, and those nagging subtle things could be nothing more than our pernicious tendency to find spurious patterns. On the other hand, important scientific discoveries frequently start out on the very borderline of detectability. So there are good reasons to reject the existence of the entity but also good reasons to think there might be one as well. As some researchers have said about ambiguous research topics: "There's too much there to be nothing, and not enough there to be something."

But one scientist desperately wants to believe there's something out there because it's comforting to think there might be a guardian preventing us from destroying ourselves. And the other one just as desperately wants not to believe because he's still rebelling against his strict parents after all these years. So we can have perfectly sound rival interpretations of the evidence, but unsound reasons for picking one interpretation over the other. Just as with the crime question, people can be right for the wrong reasons.

To see how rationalization and wishful thinking can distort any attempt to investigate the God question scientifically, consider Francis Galton's 1872 paper Statistical Inquiries Into The Efficacy Of Prayer, in which he argued that, since royalty tended to die at a younger age than other affluent classes of society, despite all the prayers offered for their health, that prayer was ineffective. As Galton noted:

The prayer has therefore no efficacy, unless the very questionable hypothesis be raised, that the conditions of royal life may naturally be yet more fatal, and that their influence is partly, though incompletely, neutralised by the effects of public prayers.

In other words, if we attempt to explain the difference by some other factor (say inbreeding or hemophilia, both of which plagued the royal houses of Europe), that's an ad hoc explanation. One wonders what Galton would have said if the results had turned out the other way. Well, actually, we don't need to wonder, because Galton also wrote:

We are justified in considering the clergy to be a far more prayerful class than [lawyers or doctors]. ...We do not, however, find that the clergy are in any way more long lived in consequence. It is true that the clergy, as a whole show a life-value of 69.49, as against 68.11 for the lawyers, and 67.31 for the medical men; but the easy country life and family repose of so many of the clergy are obvious sanatory conditions in their favour.

So when royalty turn out to have shorter longevities than everyone else, it's not permissible invoke some additional ad hoc factor to explain it, but when clergy turn out to have longer lifespans, it is. We can be equally certain that a believer would point triumphantly to the longer lifespans of clergy as evidence for the efficacy of prayer, and invent reasons for disregarding the shorter lifespan of royalty.

We can derive one useful conclusion: You have no business claiming objective evidence for your beliefs unless you would be willing to reverse your position in the face of equally good contrary data.

Science And The Existence of God

The Nature of Complex Reality

When we are dealing with one of a kind events that are very complex, for example the drop in U.S. crime rates in the 1990's, there may be so many interacting factors to consider that a definitive answer becomes impossible. It is entirely possible that radically different, possibly even mutually exclusive hypotheses, may be equally supported by data. It is also entirely possible that no hypothesis successfully explains all the data, or that every hypothesis is contradicted by some data. But in a world full of interacting forces, occasional contradictions don't disprove an idea. The fact that lightning hits a valley and misses a hilltop doesn't disprove the idea that lightning tends to hit high points. It merely proves it doesn't hit high points all the time.

The question "what does it all mean, and is Anybody in charge?" is the ultimate one of a kind, complex event. We have no consensus about what would constitute a proof, no possibility of repeating events exactly and no theoretical basis to guide us in designing a strategy for investigation. 

The Nature of God (or Powerful Aliens)

If there is anything even remotely like the god of Western monotheism, it will not allow itself to be known in any way that allows the knowledge to be used to thwart its actions or subvert its designs.

The amount of actual information about God in the Bible or the Koran is miniscule. We are told there is a Deity, it makes moral demands on humans, there are consequences for obeying or disobeying those demands, and that's about it. Also there are accounts of God acting in history from which we might be able to draw conclusions about its intentions. All the rest of the stuff in those yards and yards of books in the typical religious bookstore is mostly fluff, with a few per cent or less being useful insight.

And even the scanty information we have is subject to gross abuse. The mere belief in the existence of God leads many people to fatalism or to a belief that God will protect them from even the most irresponsible behavior. If God forgives misdeeds, many people take that as license to behave immorally and then perform some superficial act of atonement. If God makes moral demands, we have those who interpret that as a license to impose those demands on others. And everywhere we see "the will of God" being used as a cover for "my will."

So we can be virtually certain that God will not permit itself to be experimented on or statistically analyzed. If it acts, it will be in a way that defies statistical discovery. The signal will be buried in the noise, and if it's discoverable at all, it will be by methods so elaborate that their very complexity raises doubt about their validity.

The Nature of People

One possible resolution of the God question that explains all the facts is that God exists; the arguments deducing his existence are correct; and the arguments against his existence are all fallacious. Since many of the people who deny the existence of God are highly intelligent and strive for intellectual honesty, the reason for their acceptance of fallacious arguments must be some combination of faulty logic, delusion, and wishful thinking.

Another possible resolution of the God question that explains all the facts is that God does not exist; the arguments deducing his existence are incorrect; and the arguments for his existence are all fallacious. Since many of the people who believe in the existence of God are highly intelligent and strive for intellectual honesty, the reason for their acceptance of fallacious arguments must be some combination of faulty logic, delusion, and wishful thinking.

In other words, two diametrically opposite and mutually exclusive conclusions explain the empirical data equally well. So, whichever outcome is true, we have to conclude that a large fraction of the most intelligent and rational people on the planet nevertheless fell prey to faulty logic, delusion, and wishful thinking. And these are the most intelligent and rational people. The intellectual landscape below that level is bleak and scary indeed.

One approach to sorting out this mess is to focus on people who have particular claims to rationality, though how you'd judge a head to head contest between David Hume and Thomas Aquinas is hardly clear. Actually I suspect Hume and Aquinas would find themselves closer to each other than either would be to the rank and file in the God debate.

Another approach might be to look at people who converted from one position to the other, or at least endured major crises of belief, on the grounds that these people had given their beliefs special scrutiny. Even so, we'd want to exclude people who converted as the result of some traumatic experience, say, lost their faith when a child died, or became believers after a narrow escape from tragedy. Their reasons may be utterly sincere, but they're not reliable evidence.

In fact, we should be suspicious of extreme conversions in general. As Eric Hoffer pointed out sixty years ago in The True Believer, the easiest conversion is a complete 180 from one extreme to the other. A militant, Madeline Murray O'Hair atheist who becomes a rabid fundamentalist hasn't converted the core of her being; her life still revolves around anger and intolerance. The guy who says "I used to get high on drugs, now I get high on Jesus," still has to get high to cope with reality. Nuanced conversions, say from fundamentalism to moderate agnosticism, or militant atheism to moderate Protestantism, probably have much more to teach us. One of the few cases I know of where a hard-line religious believer became a moderate is Frank Schaeffer, son of the famous conservative theologian Francis Schaeffer. Schaeffer's autobiography, Crazy For God, describes his disillusionment with and ultimate abandonment of radical fundamentalism. Even here, I wonder if Schaeffer didn't shake off a shell he'd inherited from his upbringing and revert to a more natural (for him) moderate stance.

The Wastelands


At the far end of town
where the Grickle-grass grows
and the wind smells slow-and-sour when it blows
and no birds ever sing excepting old crows...
                                                        Dr. Suess, The Lorax

It's not getting any smarter out there
                                                        Frank Zappa

If even people who are seriously committed to intellectual honesty and rationality can fall prey to fallacies and wishful thinking, the intellectual landscape below that level is a blighted and desolate wasteland indeed. We have people who parrot platitudes learned in childhood and become defensive or militant at anything that shakes their confidence. We have people who, having convinced themselves of their position, believe it's perfectly permissible to lie, fabricate evidence, conceal unpleasant facts, or use dishonest arguments if the ultimate goal is to convince others of the truth. We find people who never read, think about, or associate with anything or anyone who might shake their beliefs. These are the people who won't go see the Narnia movies because of their Christian origins, or go see The Golden Compass because its author is an atheist, even though both movies can be enjoyed as pure fantasy. And we find people who are simply careless, shoddy, or superficial in their thinking and who feel no obligation to improve, indeed, who feel they shouldn't make any effort to improve their thinking because it might cause them to start having doubts. I've actually had people get angry with me when I suggest we have a moral obligation to examine our beliefs for validity. And finally we have people who believe being on the right side is a license to disregard ethics and civilized standards of conduct.

If we all went around naked, it would quickly become obvious that most people are unattractive. Now that we have the Internet and people parade their naked thoughts, it is just as glaringly obvious that most people are intellectually and spiritually unattractive.

Then we have people so messed up mentally they can't even figure out what an idea means. Like the people in some logging areas who pressured schools to remove The Lorax from their libraries. One idiot even went so far as to write a pro-logging rebuttal, even though logging companies long ago realized they have to restock forests. Or people who criticized the movie Kingdom of Heaven because it portrayed Christian Crusaders as corrupt and brutal, even though history is clear that they were corrupt and brutal.

Return to Pseudoscience Index
Return to Professor Dutch's Home Page

Created 23 January, 2008;  Last Update 28 April, 2011

Not an official UW Green Bay site