“You can’t prove a negative”

October 14, 2013 • 5:43 am

Before I begin, I emphasize again that I am not a philosopher, having taken just a few philosophy courses in college and done a fair bit of reading thereafter. What I present below are the lucubrations of a scientist grappling with theology.

UPDATE: I should have made clear that I’m talking about a theistic God here. If you posit a deistic God who doesn’t do anything, or some nebulous apophatic “ground of being” God, then of course you can’t disprove it. But there’s no reason to take it seriously, either. Those who posit an ethereal deity for which there’s no evidence are subject to Hitchen’s Dictum: “That which can be asserted without evidence can be dismissed without evidence.”


When discussing matters religious, the conversation often ends with a believer asserting: “You can’t disprove a negative.” What she means, of course, is this: “No matter what arguments you adduce against God, you’re unable to convince me—or anyone—that He doesn’t exist. That’s because you can’t prove that anything doesn’t exist.”

This argument is made by both believers and nonbelievers.

In an interview at Five Books, for instance, atheist Susan Jacoby said this:

“Of course an atheist can’t prove there isn’t a God, because you cannot prove a negative. The atheist basically says that based on everything I see around me, I don’t think so. Every rational thing I see and have learned about the world around me says there isn’t a God, but as far as proving there isn’t a God, no one can do that. Both the atheist and the agnostic say that.”

Biologist Ken Miller, an observant Catholic, said something similar on the BBC:

“The issue of God is an issue on which reasonable people may differ, but I certainly think that it’s an over-statement of our scientific knowledge and understanding to argue that science in general, or evolutionary biology in particular, proves in any way that there is no God.”

I don’t agree with either of these.

The “you can’t prove a negative” argument is wrong. You can prove a negative, which means disproving a positive (i.e., God exists)—if you construe the word “disprove” as meaning “showing that the existence of a phenomenon is so unlikely that one would have to be blinkered or perverse to still believe it.” And that is the case for God.

Scientists, of course, don’t use the word “prove”.  We have greater or lesser degrees of confidence in phenomena.  And when a phenomenon is supported by so much evidence that you’d have to be perverse to deny it (as Steve Gould put it), then we regard it as a fact, or “proven” in everyday jargon. I am immensely confident that the earth rotates on its axis, that a water molecule has on oxygen and two hydrogen atoms, and that we evolved from other creatures very different from modern humans. I regard those claims as “proven” in any meaningful sense, but to preserve the provisional nature of scientific truth, I avoid the word “proof” in both technical and popular presentations.

Now mathematicians can indeed “prove” things, for their domain is not the real world but the consequences of a series of axioms.  Mathematicans can prove that, in Euclidean geometry, the square of the length of the hypotenuse of a right triangle is equal to the sum of the squares of the other two sides. That’s why mathematicians, but not natural scientists, consider things proven. “QED,” as they say: “which had to be demonstrated.” As Sean Carroll has said, there’s no conceivable world in which you can disprove mathematical propositions like the Pythagorean Theorem; no empirical observation that can refute it. (Let’s not talk about other geometries, okay? That’s beside the point.)

But we can “disprove” the existence of something, for all intents and purposes, by showing that the evidence that should be there if that something existed is missing. Victor Stenger has made this point repeatedly, and a famous earlier example is Carl Sagan’s argument against “The Dragon in My Garage” from The Demon-Haunted World.  Read it (free at the link).  It’s about someone who claims there’s a fire-breathing dragon in his garage, but the dragon is invisible, and its advocate keeps countering the skeptic’s observations of a lack of evidence with claims like ‘it’s invisible,” and “it floats, so you can’t detect footprints,” and “its fire is heatless, so you can’t feel it.”

Sagan’s point was, of course, that it doesn’t make sense to believe in things for which there’s no evidence, and that the “you can’t prove nonexistence” claim is fatuous when the evidence should be there. As he noted at the end of this parable:

Once again, the only sensible approach is tentatively to 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.

(Sagan was a lot more vociferous against religion than most people think!)

And so it is with other things.  Can you disprove that I don’t have a heart? Of course you can: just do a CAT scan! Can you disprove that I am not married? For all practical purposes, yes: just try to find the records, ask people, or observe me. You won’t find any evidence. Can you disprove the notion that fairies live in my garden?  Well, not absolutely, but if you never see one, and they have no effects, then you can provisionally conclude that they don’t exist.

God is like those fairies.  Not only is he a supernatural being who’s supposed to exist, but, unlike fairies, a theistic God is supposed to have designated effects on the world. In particular, he’s supposed to be omnibenevolent, omnipotent, and omniscient.  Some further believe that there is an afterlife in which one goes to either Heaven of Hell, that prayers are answered, that God had a divine son who was resurrected, and so on.

If these are true, there should be evidence for them.  But there is none. In particular, this is what we find:

There is no evidence of divinity or miracles in the present world, and no palpable evidence of God-inspired miracles (prayers don’t heal amputees).

God, despite being omnipotent and desirous of our knowing him, has never appeared despite his manifest ability to do so. He could, for example, write “I am Yahweh; obey me” in the stars.  This is the “hidden God”, the Deus absconditus. As philosopher Herman Philipse has noted, God should want each individual to know of his existence to create a reciprocal relationship.

Tests of intercessory prayer show no effect.

There is no good justification, assuming a benevolent and all-powerful God, for “natural evil,” the suffering of animals and innocent children due to diseases and natural disasters.  Theologians’ attempts to explain why, for example, children get leukemia, why ten million civilians met their deaths at the hands of the Nazis, and why thousands are killed by tsunamis, are laughable, and not remotely convincing to anyone who hasn’t already bought into religious delusion.

Earlier “evidence” for divinity has been dispelled (creation, Adam and Eve, Great Flood, etc.)

A benevolent God would not kill off humanity in 5 billion years. Nor would a benevolent and powerful God use evolution or natural selection to create modern life and humans. That just doesn’t make sense, though theologians concoct amusing arguments not only why evolution makes sense, but why it should be God’s preferred way to bring species into being.

There is no explanation for why a benevolent God would allow more than 99% of the species he wanted to exist to subsequently go extinct without issue.

Most of the universe inhospitable to life, and nothing lives there. Why this largesse of uninhabitable space if God created Earth for humans? Even if life exists elsewhere, it can’t be common, and the trillions of uninhabited stars serve no purpose.

In the case of God, then, the absence of evidence is indeed evidence for His absence.  We can provisionally but confidently say that there’s no evidence for a God. and therefore reject the notion that He exists. (This could be revised, of course, and in earlier posts I’ve given some possible evidence that would convince me of divine beings.)

Needless to say, all the above observations make sense—indeed, are expected—if God doesn’t exist.

200 thoughts on ““You can’t prove a negative”

  1. The “you can’t prove a negative” line is a red herring because the onus is obviously on whoever posits the existence of God to justify it. It’s Russell’s flying teapot point, often quoted by Dawkins(and maybe slightly misremembered by me.
    Not sure it’s even true to say, as a matter of logic, that you can’t prove a negative. I think Gödel did. At least, he proved that certain things are unprovable, and in my book ‘unprovable’ is a negative.

    1. What Gödel showed was that there are things [in a closed system, e g mathematics] which we know are true but which cannot be proved.

      1. To be precise, they cannot be proved from within the system, which means that no system is complete (i.e., all axioms are proved within the system).

        1. Er, that’s not what complete means at all. A formal system is complete if every statement in the language of the system can be either proved or disproved in the system.

  2. Don’t we first need to have a coherent definition of the thing we’re disproving? This is the main issue I see. Which god are we talking about? What are the characteristics of said deity? I don’t think there has ever been much agreement on theses basics.

    1. The term “God” does not refer to a specific thing. It refers to a wide variety of concepts/entities, and the answer to the question could (and should) be different for different referents.

      For example, let’s say for the sake of argument that the question is “What would convince me of the existence of the God of Abraham?”

      My answer is ‘nothing’. There is no evidence that could possibly convince me of the existence of such a being, because there is no evidence that can possibly constitute justification for the claim that such a being exists. The very nature of the being in question is such that its existence would violate the set of assumptions I need to be able to make in order for any sort of empirical evidence to support any claim.

      Consider this: The claim is that a being exists which, among other things, has the ability to alter at will any evidence that I can possibly observe. It also has the ability to alter my own perceptions of the world in any way it wishes, and even to alter my own memories of past observations.

      How then, could I possibly consider any evidence I observe to be reliable evidence concerning the nature of that being?

      The moment I accept that any being or beings with such capabilities exists, I forfeit the ability to say anything further about it/them. Maybe it’s Yahweh. Maybe I’m a computer simulation and somebody in the real world is manipulating my experiences. Maybe aliens with technology I can’t even begin to imagine are manipulating me. Maybe I have a brain tumor that is causing severe audio-visual hallucinations. Regardless, I clearly cannot draw any conclusions about the nature of whatever phenomena are responsible for the “evidence” in question.

      Of course, this particular response does not apply to all conceptions of God, but it applies to a lot of them. For other conceptions of God, my response would be different.

      1. Cara,

        I do not agree with your reasoning. Your argument seems to take the form of:

        If some entity has the power to affect X without our knowing it, we can not come to a rational conclusion about X.

        But this doesn’t necessarily follow at all in normal reasoning. Strictly speaking, our beliefs about the world are never absolutely “proven” true, but are provisionally held (even if strongly) inferences from the evidence we have. And, generally speaking, we navigate these issues via tools like parsimony in our explanations, and principles like: “things are as they seem UNLESS we have other reasons to doubt things are as the seem.”

        My wife has the power to be having an affair that I’m unaware of at the moment. But that leaves open the question: is she ACTUALLY doing so? And the only warrant I’d have for believing things are other than they look (that our marriage is sound), would be additional evidence that my wife is actually having an affair.

        Twelve happy magical unicorns could have affected the results of the Higgs Boson experiments, fooling us about the results, in a manner undetectable to us. This is a proposition you can not strictly disprove; it’s logically possible. But there’s no reason to adopt that additional hypothesis that the experiment was affected by magical powers; nothing suggests it and until there’s additional reason to think magic beings are affecting our experiments, we say “we infer from the way things seem to be, and don’t posit additional hypotheses.”

        So even if you posit a God who COULD affect your experience in the way you suggested, it doesn’t follow that he IS affecting your experience. We would still be left asking “IS it the case God is affecting our experiments?” If “Fred” said “but maybe God fiddled with the Higgs Boson results undetectably to fool us” we’d rightly ask “Ok, that’s a hypothesis: how do you support that your hypothesis is more likely over the one where we’ve simply inferred from how things seem to us?” If Fred said “well, see, there wouldn’t be any detectable difference” then Fred’s hypothesis gets tossed aside for exactly the same reason all other such hypotheses get tossed: it posits gratuitous assumptions that are unnecessary for explaining the phenomenon.

        Even if you’ve already accepted the existence of an entity like God, it’s still gratuitous in a hypothesis to propose God IS DOING X without supporting evidence for that additional proposition. You’d need some additional support, for instance being able to show that God would have the motivation to be fooling with the experience, or observation/experiment in question.

        Now, if we are talking about the Christian God, you can find passages both in support of God not being a trickster and some that suggest He can be. But that moves on to a different conversation; what I’m disagreeing with is your claim that it follows *simply from the fact* of a God’s power to fool us that this undermines our ability to hold reasonable conclusions about our experiences and empirical inferences.

        And when you start talking of “maybe” being in a computer simulation, or maybe aliens are manipulating you, again there is no reason to take leave of normal empirical inference and parsimony. Is there any additional EVIDENCE for, or greater explanatory power provided by, the existence of such people fooling aliens, or computer-realms? No. Then you’ve no need to worry that such gratuitous concepts undermined your ability to know the world. No more than those propositions already render your inferences, or science’s inferences, invalid.



        1. Hi Vaal! Aye, God works in mysterious ways! (AThe ultimate jade’s trick.)I agree. Descartes talked abot the Devil deceiving us, but it’s really all the same thing, and doesn’t vary according to what non-existent thing you ae discussing. It doesn’t matter what colour the Teapot is.

        2. Hi Vaal,

          I think you’re only half right.

          When you’re speaking of reasonable beliefs to follow scientifically speaking, then assuming that God tinkered with the universe to give you the God particle is useless to you. Scientifically speaking you can’t use this hypothesis. You can’t reject it or support it. You can’t even test it. So you’re left only with the hypotheses that you can test.

          But when the conception of God is such that he isn’t so much tinkering at the edges but rather created the universe, the people in it and their logical processes – including as currently being discussed in another thread the very idea of logic – then all bets are off.

          As for parsimony, that in itself is a logical construct.

          In the end we’re discussing the nature of reality, and in this discussion science can only take you so far. Logic can only take you so far. And faith can take you somewhere completely different.



    2. Agreement is a problem. I think an equally large problem is the believer’s slippery mid-game changes in their definition. When you see them in the churches or posting on Facebook they believe in a literal Abrahamic god who flooded the earth, divinely inspired an infallible book, is a chaos of unimaginable superlatives (omniscient, omnipotent, the very definition of love), helps sick relatives recover and harried moms find parking spaces, and is just generally leaving his fingerprints on everything. When you address the counter-evidence for these claims about god, god suddenly becomes remote, ineffable, and deistic.

      To be fair to them, I think it is often less a matter of conscious mendacity than it is part of their own deep befuddlement. Since they feel that their god-character is real, whatever kind of god they can get you to agree is possible is taken by them to be *their* god and they feel in their gut that they have demonstrated what needs to be demonstrated. They can’t imagine their god as fictional long enough to understand that the syllogism “some kind of god is possible” therefore “jesus” is absurd.

      Two apropos comics:


  3. As Sean Carroll has said, there’s no conceivable world in which you can disprove mathematical propositions like the Pythagorean Theorem; no empirical observation that can refute it.

    This is beside the main point of the post, but one of my hobbies is claiming that maths and logic are subsets of science, and that they are empirically derived.

    The only reasons that we accept the proof of Pythagoras’s theorem is that both the axioms and the logic used are empirical products, they are found to work in our universe.

    Also, all mathematical proofs are done by humans and all humans are fallible. It is in principle conceivable that all humans have a blind spot when it comes to logic such that their proof are invalid. (Similar to the fallibility shown by the colour illusions recently on WEIT.)

    How would one go about verifying mathematicians as infallible, and thus their proofs as “proven” beyond those of scientific “proofs”?

    Thus the reason we actually accept Pythagoras’s theorem is the copious empirical verification.

    [Sorry for the tangent, I agree on the main point of the post.]

    1. I don’t think one can really claim empiricism or science on math unto itself.

      Pythagoras’s theorem is just the result of definitions that must be reached if there are to be no contradictions, it can be done from the armchair in the library. It is a useful idea that can be applied to real world things. Just because I can imagine a line with no thickness and infinite length it does not mean that line exists (except perhaps as an idea). Such a line is just a useful idea that can be applied to a hypothesis and then tested against experiment.

      Math is part and parcel to all great science, it is a method of thinking that is both expansive and meticulously precise, but it cannot be science unto itself.

      1. it can be done from the armchair in the library.

        Only by using axioms and rules of logic that — ultimately — derive from empirical experience of our world.

    2. Indeed, this is an important point, and the subject of much argument among mathematicians and philosophers since the realization that non-Euclidean geometries can be constructued. For this discussion, I’d like to see people avoid the tagline “unlike mathematics, which can be proven”.

    3. There are plenty of mathematical results that can’t be tested empirically and are still viewed as valid.

      Banach-Tarski paradox, the proof that e is a trancendental number, Cantor’s diagonalization argument for the cardinality of the reals being greater than that of the integers, the structure of all possible finite fields, and many, many more.

      Don’t mistake a truth about why we have mathematics in general as a field of study, (because it produces many results that work out and are useful in the real world) with something that must apply to every mathematical result.

      1. Science doesn’t require every result to be empirically testable.

        For example, were I to run a program backwards and predict that “there was a solar eclipse visible from Antarctica in 18706 BCE”, I may not be able to empirically verify that claim but could still regard the statement as valid if I had verified the laws of physics, the planetary data, and the computer program by testing them on things that I could compare with data.

        I consider that your mathematical statements above are of that nature, being implications of axioms and logic that can be empirically verfied.

        1. I consider that your mathematical statements above are of that nature, being implications of axioms and logic that can be empirically verfied.

          Not quite. I can’t think of any reasonable sense in which the Axiom of Choice can be empirically verified.

          1. I can. Create lots of sets of disjoint, non-empty sets and see if you can form a set from each them which satisfies the Axiom of Choice. Then do statistics on the results (which will be 100% success) and when you reach some desired level of significance that the results could not be random, declare that the Axiom of Choice has been experimentally verified. (This is how the Higgs Boson was verfied – although the form of the experiment was different and more expensive.)

            I also have commented on a couple websites (perhaps including this one, previously) that I see math as having empirical foundations and therefore it should be considered as a form of science rather than something separate. (I will spare you my examples this time.)

          2. I don’t quite follow. Are you talking about finite sets? For those you don’t need statistics, AC can be proven for finite sets.

            Here is an example where your “experiment” would fail: Consider R/Q, the set of cosets of the reals modulo the rationals. You can’t select an element from each coset without appealing to AC, for the resulting set would not be Lebesgue measurable, and the statement “all subsets of the reals are Lebesgue measurable” is consistent with the remaining axioms of set theory.

        2. The most general scientific statements aren’t testable either, at least directly. They have to be enriched with subsidary assumptions first.

          (Incidentally, this is one reason why there’s no dividing line between science and science-oriented philosophy.)

      2. But mathematical idealizations (such as infinities) corresponds to physical toy models. They can be used to derive valid principles, but they don’t need to be testable to do so.

    4. This is not so far out there. We have always known that axioms are made up. The hope for a long time was that there might be only a few axioms so that this made-up aspect of math could be limited. Of course Godel demonstrated that any formal system must be either inconsistent (contain statements A and not-A) or incomplete (contain true statements that are unreachable in the formal system). Still, there was hope for a time that these unprovable statements would be limited in scope. It has since been shown that there are an infinite number of diophantine equations whose solution is unreachable in any formal system. This is more of a real math problem than the esoteric construction of Godel, which might have been thought to be a freak example, and suggests strongly that an infinite number of axioms will be needed to keep going forward with math. This interpretation places mathematics firmly in the realm of the empirical, since it is from experience that we get our axioms.

      Nonetheless, while math as a field is forever incomplete and grounded in empiricism, any particular set of axioms and production rules will allow for proofs which are unassailable within the system, for which the word “proof” means what people expect it to mean.

      1. Also, thanks for the diophantine example! I was aware of the Godel construction implications, but I didn’t know of this (as you note) stronger example.

        Do you have refs handy? My google-fu got jinxed. (I found a paper on “On the Inherent Incompleteness of Scientific Theories” [and therefore God]. Neither readable nor searchable. :-/)

        1. Searching for “gregory chaitin diophantine equation” should yield relevant results as he has been a big champion of the interpretation that says math is more experimental than we might have liked. I think Matiyasevich’s theorem is also apropos, and this is of course tied to Hilbert’s Tenth Problem (Answer: No).

          I believe I first read about this in Information, Randomness, and Incompleteness by Gregory Chaitin. Searching in Google Books I see an informal sketch of this proof on p. 77. That informal sketch cites this:


          as a starting point.

      2. It’s possible to translate statements about a Turing machine halting into statements about diophantine equations having solutions (and vice versa). This is an interesting (and surprising) result. Since there are many (provably) undecidable problems about the halting problem, you get undecidable problems about diophantine equations.

        However, I don’t see that recasting the problems in terms of diophantine equations shows up more of a “real math problem” than talking about the Turing machines more directly.

        It’s true that every consistent formal system that’s powerful enough to do arithmetic and represent diophantine equations has an infinite number of diophantine equations that have no solution but cannot be proven within the system to have no solution. However, that doesn’t seem like a reason to add more axioms; if you add more axioms, you’ll be in a new formal system that still has an infinite number of problematic diophantine equations.

        1. This result is, like, fifty years old now. I agree that from our perspective now this seems like an unremarkable distinction. The range of problems that have been shown to be undecidable is now large enough, and the much simpler to prove and understand Turing machine version of Godel’s theorem has been absorbed much more widely, and can be seen to have practical implications more easily, so that I think even without the explicit diophantine result few would be surprised now to learn that problems we’d actually want to solve are undecidable. From a more historical perspective, though, I think there were mathematicians who suspected that these undecidable propositions were just one-off things, some freak set of constructed problems that showed an unreachable corner case in mathematics, but that all of the routine math we’d like to do to actually solve real world problems might still turn out to be decidable. Of course, computer scientists can see the utility of wanting to have a Halting Program, but the utility of wanting to prove a Godel sentence is not so clear. Solving diophantine equations is an example of something that there is every chance we’d want to do, but find out that we may not be able to.

          It is not a direct reason to add more axioms, we do not know which diophantine equations are unsolvable in our current system, only that some infinite subset are, so we can’t just go adding axioms to try to fill in the missing diophantine equations and, as you note, no amount of axioms is ever going to get out of this problem anyway. But psychologically it highlights that for some problems we encounter, real problems we want to solve, we may have no choice but to turn to empiricism.

          This, too, is pretty uncontroversial now. The axiom P != NP is an empirical result. Many would balk at calling it an axiom because there is still active effort to prove it with other axioms. Proofs based on this empirical observation are often phrased as “If P != NP then..”, to sort of hedge that we don’t have any proof that this axiom is true. Nonetheless, in practice it is used exactly like a new axiom and there are whole books of results whose premise is that P != NP, and whole fields of practical endeavor, such as cryptography, that are predicated on this quasi-empirical axiom.

          So I think all of this adds up psychologically to emphasize that undecidability infects everything, even something seemingly mundane as diophantine equations, and is a sort of signpost that this empirical aspect of math isn’t going to go away.

  4. Mostly believers just want to rely on an argument from ignorance, and reality has backed them up so far that the only ignorance they can still use is logical uncertainty. Dr. Conye has it right, essentially there is no absolute certainty for the real world like there is in math, just reasonable and useful certainty, but that is not a license to start making things up and claim equal legitimacy for all claims.

    1. Although it is a good bit of fun to take that same license and run with it, as the IPU and FSM followers attest.

  5. There is a problem with taking this concept (ie. Absence of evidence = evidence of absence) into the wider range of philosophical and scientific debate (look at quantum physics, extra dimensions etc.) because of so many surprises that turn up in modern fields of inquiry. This is why even hardened atheists are reluctant to take this argument up.

    1. Absence of evidence is evidence of absence if and only if evidence of presence has been looked for properly (putting the burden on that adverb, which depends on whatever theory most adequately and simply explains those things for which there actually is strong evidence). Where’s the problem?

      1. Almost.

        Absence of evidence is always evidence of absence.

        Absence of evidence becomes proof of absence when the search is so through that there’s no chance of the phenomenon in question remaining in some as-yet-unsearched domain.

        There most certainly are elephants in the universe, for example. But there most emphatically are no herds of angry elephants stampeding through my office as I type these words. The absence of evidence of such an herd is all the proof I need to know with absolute certainty that the phenomenon in question (an angry herd of elephants stampeding through my office) does not exist.

        Similarly, there’s a reasonable chance that there are very powerful entities somewhere in the universe, and we know that there are entities that have the best interests of humanity at heart — ourselves. But, as surely as there are no elephants in the room with me, there are no very powerful entities with the best interests of humanity at heart.

        The theological gods also fall on simpler grounds. Just as we don’t need to look for evidence to rule out the possible existence of married bachelors or the largest prime number, we also don’t need to look for evidence to rule out the possible existence of a prime mover or the greatest power.



        1. “we know that there are entities that have the best interests of humanity at heart — ourselves”

          This doesn’t strike me as a given. At least not from listening to my fellow humans talk and watching the news to see what my fellow humans are doing. 😉

          1. Considering how the on-again, off-again negotiations over the default were again in the “off-again” phase this morning, you may well have a point….


  6. These are the conclusions that a rational person would make. Unfortunately, the majority of Americans are not raised to be rational.

  7. Of course in the vast majority of religinut cases, when asked for evidence they’ll simply say “here’s the bible, that’s the evidence”

    I’ve found that most apologists fail to understand the difference between evidence and a claim. They consider all their claims to be evidence and all evidence that counters them to be claims.

    Evolution, Cosmology, Geology, Physics etc etc don’t in themselves provide evidence to disprove *a* god – they simply demonstrate that the particular “god” in question at that time is not valid.

    It is possible to conceive that the known universe (all of spacetime) was created by another civilization in some version of the LHC (scale and time being relative this wouldn’t be *impossible*) and that creator civilization could be considered to have “played god” but they are NOT the christian/Muslim/Hindu “god”.

    1. “It is possible to conceive that the known universe (all of spacetime) was created by another civilization”.

      Not with inflation, it isn’t. At least the contents (structure formation) is decisively derived from quantum fluctuations.

      Also, “the known universe” (from inflation) is just a volume ~ 1000 times larger than the observable universe.

    2. “I’ve found that most apologists fail to understand the difference between evidence and a claim. They consider all their claims to be evidence and all evidence that counters them to be claims.”

      That will be useful, thanks.

  8. “You can’t prove a negative” is not even wrong. In any context that the world “prove” makes sense if you can prove a negative, for if you have proven P you have also proven ~(~P).

    Even what the phrase means is that you can’t prove that something doesn’t exist, again the assertion is obviously wrong. In mathematics we prove “negatives” every day. For example: there doesn’t exist a field with 6 elements, here doesn’t exist a closed polygon with 7 angles and 6 edges, there is no simple polyhedron with seven edges, etc. These statements prove things about the real world as well, the last one for example tell’s us that is impossible to carve a potato in a particular way.

    1. I’ve always been partial to this reply. Of course you can prove a negative! ~(P & ~P) is a theorem!

      It’s open to the “but I meant a negative existential claim” reply. To this I always answer: “Are you saying that science can’t prove there does not exist air at the top column of a mercury barometer?”

      Also, there does not exist an x such that x is F and x is not F.

  9. “…theologians concoct amusing arguments not only why evolution makes sense, but why it should be God’s preferred way to bring species into being.”

    I wonder if theologians could concoct arguments for why water runs down hill or any number of phenomena as being Gods preference and not merely guided by the laws of nature? Since the theory of gravitation seems to satisfy most of us concerning gravity, it would be pointless, but still amusing. As far as evolution serving as an explanation for diversity, why does it satisfy some but not others? Yes, I am amused.

      1. Thanks for the link – I’ve heard the term for years, but didn’t know the source. The article is better written than most serious screeds (the difference of a real education, I guess).

  10. Notice that “You can’t prove a negative” is itself a negative. Yet people who assert it seem to think it’s provable.

    There are some incontrovertible cases of proving a negative. Logically impossible objects, such as the Penrose triangle, can be known with certainty that they don’t exist. Same with logically contradictory things like square triangles. So there are two cases of *proving* a negative.

    As for logically possible objects: “But we can “disprove” the existence of something, for all intents and purposes.” The qualifier “for all intents and purposes” is key. In the domain of logically possible entities, all we can possibly do is assign a probability to the existence or non-existence of things. Nonetheless, it’s no trivial matter when the probability of, say, Allah existing is virtually zero.

    So: atheists could indeed *prove* that God doesn’t exist *if* “God” is defined in a way that makes him logically impossible. For example, a three-in-one God seems to be impossible (on logical grounds), and therefore such a God doesn’t exist.

    On the other hand, if God is understood in a coherent, logically impossible way, then all we can do is make observations and, based on the evidence acquired, assign a probability to his existence. In my best judgment, the evidence for such a being is incredibly small and uncompelling.

    1. “if God is understood in a coherent, logically impossible way” –> “if God is understood in a coherent, logically *possible* way.”

      1. Actually, it works both ways! There is a certain Calvinistic physicist who pops by occasionally who argues that miracles are possible because God does not do things which break the laws of physics, he simply operates “outside” of them! 😀

  11. I have noticed recently among atheists a trend to label themselves “agnostic atheists” The argument is that “we can’t prove a negative (agnostic) but they believe there is no god “atheist”.

    However this seems to me pedantic and unnecessary because it is not parsimonious. They would have to be agnostic about any assertion without evidence, no matter how ridiculous.

    Understanding that claims without evidence can be dismissed without evidence probably renders the label of “agnostic” unnecessary.

    To me it is problematic because agnostic can also be misunderstood as “unsure” and the religious will pounce on that in any conversation.

    1. Yes some news papers announced shock horror that Dawkins secretly admitted to being an agnostic a while back. Perhaps they missed the chapter in the God delusion prior to that.
      If pushed for strict position I would go agnostic atheist on the grounds that a completely hands off god cant be disproved but is also irrelevant for the purposes of living. So in general discussion its atheist, especially since most religions have a god which is anything but hands off apart from when decent recording equipment is available (strange how the supernatural are put off by that, maybe its the modern equivalent of hanging an iron horseshoe on the door)

  12. ” “you can’t prove nonexistence” claim is fatuous when the evidence should be there. ”

    I think this statement sums it up. If the evidence ‘should’ be there then proving a negative is trivial, so long as there is no evidence. This is a normative claim, one that ascribes how things ought to be if we were rational beings. Many people are afraid to prove a negative. It is easy, just say there is no evidence. There is always tomorrow…and maybe there will be evidence for the dragon in the garage or Nike in Athena’s right hand.

      1. Looking at that picture made me kind of sad. I was thinking how it would be received today if someone were to try to build a statue of Athena in the capitol. I know this building is in Austria, but there are similar things in the U.S (e.g at the Library of Congress: http://www.vanderbilt.edu/AnS/Classics/classical_traditions/). I just have a vision of someone trying to build some of these things now and being loudly opposed by the religious right who would see it as the installation of false gods, or undue reverence to the homosexual greeks, or some other such nonsense. I just can’t picture contemporary political leaders having any appreciation for these things and less so that the public would tolerate it and that’s kind of sad.

    1. You don’t need any evidence to prove that round triangles don’t exist anywhere in reality. This is crucial to note, because some conceptions of God seem to be logically impossible. For example, a totally omniscient God who also has free will appears to be logically incoherent. Thus, such a God can be proven — in a logical sense — to not exist. When God isn’t understood in any logically problematic way, though, evidence is indeed what matters. And when it comes to evidence, certainty is ultimately not attainable. Practically, though, this doesn’t matter one bit when the probability of, say, Richard Dawkins existing is virtually 100% and insensible unicorns is virtually 0%.

        1. That’s exactly the point I was trying to make in a more general fashion in a comment below. All knowledge is relative to certain premises. You only have to give up all pretense to absolute truth—which is a backdoor to authoritarianism in any case, so that would actually be a very good thing indeed.

  13. There’s a classic response to that, by philosophy professor Steven Hales. It’s programmatically titled “You Can Prove a Negative”. He makes two good, and cricial, points, about how any statement can be formulated as a negative and about how the demand for proof of every premise used in an argument must lead to an infinite regress and thus make any proof at all impossible. (Just don’t be too trusting of his second part, on induction, which he gets mostly wrong.)

  14. By FAR the best response to the “can’t prove a negative” myth is by Steven Hales. Professor of philosophy at Bloomberg University:


    To paraphrase him, he makes the following 3 points:
    1. You can always prove a negative as easily as you can prove a positive.
    2. Deductive arguments are the only things that can prove anything 100% (if you accept their premises). There are countless examples (like in logic, or mathematics) of proven negatives.
    3. Inductive arguments (the usual sort employed in science) can’t prove anything to 100%. But this again is as true of positive as negative statements.

    It really is an excellent little article – I can’t recommend it highly enough.

      1. In 1965 Prof Dijkstra proved that semaphores would allow asynchronous processes to safely update a common resource – c’mon WordPress . . .

    1. » Jeremy Nel:
      Deductive arguments are the only things that can prove anything 100% (if you accept their premises).

      That doesn’t make a whole lot of sense, I’m afraid. You simply cannot say that deduction is capable of proving things absolutely (which “100 %” will be taken to mean) and in the same breath acknowledge that even a deductive argument can only be relative to its premises. Only the second part is true.

      Deductive arguments are the only kind of argument that transfer the truth-value of their premises to their conclusions. (And, mutatis mutandis, the falsity of their conclusions to their premises.) That is all they do: they tell us what would be true if their premises were true. In the words of the logician Mark Notturno (from Science and the Open Society:

      [Logic] cannot force us to accept the truth of any belief. But it can force us, if we want to avoid contradicting ourselves, to reexamine our beliefs, and to choose between the truth of some beliefs and the falsity of others …

      Inductive arguments (the usual sort employed in science) can’t prove anything to 100%.

      Inductive arguments cannot, in fact, prove anything. And they’re not employed in science, either. First, the logic. So far, nobody has been able to meet Hume’s challenge to induction (from An Enquiry Concerning Human Understanding):

      It is only after a long course of uniform experiments in any kind, and we attain a firm reliance and security with regard to a particular event. Now where is that process of reasoning which, from one instance, draws a conclusion so different from that which it infers from a hundred instances that are nowise different from that single one? This question I propose as much for the sake of information, as with an intention of raising difficulties. I cannot find, I cannot imagine any such reasoning. But I keep my mind still open to instruction, if any one will vouchsafe to bestow it on me.

      … It is impossible, therefore, that any arguments from experience can prove this resemblance of the past to the future, since all these arguments are founded on the supposition of that resemblance.

      So there is no logical basis to call induction valid. And, as Hume says in his last sentence there, there is no empirical basis either, since an appeal of the “It works, bitches” kind blatantly begs the question, i.e. assumes that which it purports to prove.

      Second, science. As Popper explained at considerable length and in great detail in his The Logic of Scientific Discovery, scientific theories have the logical form of strictly universal statements, which cannot be verified (as strictly singular statements can) but can only be falsified:

      The theories of natural science, and especially what we call natural laws, have the form of strictly universal statements; thus they can be expressed in the form of negations of strictly existential statements, or, as we may say, in the form of non-existence statements (or ‘there-is-not’ statements). For example, the law of conservation of energy can be ex­pressed in the form: ‘There is no perpetual motion machine’, or the hypothesis of the electrical elementary charge in the form: ‘There is no electrical charge other than a multiple of the electrical elementary charge’.

      In this formulation we see that natural laws might be compared to ‘proscriptions’ or ‘prohibitions’. They do not assert that something exists or is the case; they deny it. They insist on the non-existence of certain things or states of affairs, proscribing or prohibiting, as it were, these things or states of affairs: they rule them out. And it is precisely because they do this that they are falsifiable. If we accept as true one singular statement which, as it were, infringes the prohibition by asserting the existence of a thing (or the occurrence of an event) ruled out by the law, then the law is refuted. (An instance would be, ‘In such-and-such a place, there is an apparatus which is a perpetual motion machine’.)

      1. Peter – thanks for the reply. I’m aware of both Hume and Popper’s arguments (and a big fan of both, actually). While I agree that no one has been able to meet Hume’s induction challenge, he himself didn’t advocate its abandonment, as I’m sure you’re aware. I think it’s important here to distinguish, as he did, between being able to justify induction on logical grounds (which he showed convincingly that you can’t do) and whether you therefore ought to jettison it (which he variously thought impossible, or unneccessary, or possibly even irrational).

        Regarding science’s process, again, I’m sure that you’re aware of the criticisms of Popper’s ingenious solution to the problem of scientific induction. You may or may not find them compelling. But what I think is beyond doubt is that science DOES employ inductive arguments all the time. But arguments such as this are common: (1) the Drosophila in this lab speciate by mechanism X, (2) the genetics of Drosophila are similar to the genetics of the rest of the eukaryotes, (3) therefore, mechanism X is likely to be employed by many/most eukaryotes. I don’t think Jerry would be working on Drosophila for most of his career if he didn’t buy into induction!

        Of course, it is at this point (but not before) that Popper’s falsification becomes relevant.

        1. Jeremy, two things. Hume did indeed not abandon induction—but that is because he couldn’t conceive of an alternative. He said that humans, given the invalidity of induction, were forced to reason irrationally. But philosophy didn’t stop with Hume; crucially for this problem, Popper came along and offered a solution to Hume’s problem that presented a credible alternative.

          The problem was: How can we reason from finite observations to general statements about how nature works that are certain? The answer is: we can’t. But if we only give up the idea of certain knowledge, we can employ a deductive kind of reasoning that can at least tell us what kinds of ideas cannot be true and thus which ideas are better than others, given a particular problem-situation.

          And that was Popper’s goal: to find a convincing mechanism for the growth of knowledge. Now, what you said about Drosophila isn’t about how we increase or knowledge but (at best) how we arrive at new theories. And if you must, you can call that induction; in that sense, words are just labels, and as long as we are not confused by them, we can use them the way we want. But I think you are, in fact, confusing something here, namely the reasoning from a finite set of observed instances to actual knowledge—which is what Hume and Popper were talking about when they used the term “induction”—and some kind of inference that is merely inspired by observation. And as Popper pointed out, how you arrive at a theory is irrelevant to the problem of the growth of knowledge: what counts is a theory’s explanatory power and its being tested.

      2. I don’t want to defend the philosophical idea of induction, especially since I think we are saying the same thing.

        But “It works, bitches” is a valid map from the existence of physical laws, and induction is then a clumsy, but valid, embodiment (toy model) of such laws. (But only such laws.)

        [The real strength in induction is the pattern recognition of possible hypotheses.]

        And naive popperianism doesn’t map to exclusionary testing at large. [I’m quite frankly amazed that Popper stopped there, if he did.]

        As already the mythical but eminent empiricist Sherlock Holmes hypothesized: “It is an old maxim of mine that when you have excluded the impossible, whatever remains, however improbable, must be the truth.” Wikiquotes.

        Or as Jerry’s physics mentor Sean Carroll hypothesized: “The Laws Underlying The Physics of Everyday Life Are Completely Understood”.

        Whenever prediction beats down uncertainty in a process under 1 % (and we have excluded possible competitive theories), it doesn’t matter whether we consider past or future events. We can trust the result to the given quality figure.

        I trust that the Earth will continue to rotate at 24 hours/day for the practical foreseeable future. I don’t want to call that ‘induction’ as much as “trust in the processes used to derive this knowledge”.

        1. » Torbjärn Larsson:
          But “It works, bitches” is a valid map from the existence of physical laws, and induction is then a clumsy, but valid, embodiment (toy model) of such laws. (But only such laws.)

          I’m afraid I don’t know what that is supposed to mean.

          [The real strength in induction is the pattern recognition of possible hypotheses.]

          That sounds like you are using “induction” in the same sense as Jeremy Nel was. I already answered that above: you are confusing two entirely separate meanings of the word.

          And naive popperianism doesn’t map to exclusionary testing at large. [I’m quite frankly amazed that Popper stopped there, if he did.]

          So you don’t actually know whether he stopped there, but you are still amazed? How about you read something that was actually written by the man before coming to conclusions about his ideas? You could do worse than start with this excerpt about falsifiability from Popper’s Conjectures and Refutations.

  15. We cannot even agree – or the religions cannot – what a god actually IS or what it precisely means. That means the idea of a god is incoherent – so what are we looking for in terms of evidence? We discussed on WEIT a while ago what it would take for people to be convinced that there was a god. For some reason a 200 foot tall Jesus was something mentioned but there were those of us who said that might be just evidence of ‘super-beings’ not of a god/s. This is another reason for me NOT to believe in a god, even as a sensible concept.

    By the way – “Can you disprove the notion that fairies live in my garden” – I concluded from the view that you live in an apartment!

  16. Just as an aside, the claim that you can’t prove a negative is simply the claim that there is no logically sound way of demonstrating, logically, a negative existential judgement. We can prove, within systems of mathematics, where the axioms (or basic assumptions) are given, that 2+2 does not equal 5, which is proving a negative. So, the general claim that you can’t prove a negative is false. But, with existential judgements, the issue is otherwise, since to prove a negative existential judgement you need evidence, and it is a matter of the logic of induction that you cannot produce what would have to be both contingent and a priori evidence for a negative claim of existence. That’s the point of “you can’t prove a negative.”

    As you say, Jerry, we hold it “provisionally” that fairies don’t exist. It’s provisional because there is no logically sound way of showing that the claim that fairies exist is a contradiction, but we simply find no evidence that they do. It’s not a very sound basis, therefore, for claiming that something, like fairies, for instance, exist.

    The same goes (in spades) for a supernatural entity called a god. Stoics and Epicureans (if I have my history right) believed that there were gods; but held that they were indifferent to human beings, and did not intervene in the world for good or ill. Such gods are as good as non-existent, because the only reason for positing supernatural entities is either because (i) there is a theoretical basis for such a posit, or (ii) because they are held (on their merits in evidence) to play some role in human life and worldly affairs. Neither assumption seems able to be established. A prime mover, in Aristotle’s case, was simply — for Aristotle, because motion depended on a prior cause — a necessary theoretical assumption at the end to a chain of reasoning based on causation, given his understanding of motion, and his presupposition that an infinite regress is absurd. Too many ifs in that to be much good to religion, for such a being has no religious point, and there is no evidence that a god intervenes in human affairs to give it such point.

    However, it is important to note that contemporary theology in its more radical variety accepts that there is no existent being called god. Tillich would say that that is not what we mean by the word ‘god’ anyway, because God is not amongst the things that exist. God is, instead, the ground of existence, the Ground of Being. God transcends existence altogether. This is basically Aquinas’ argument. Other theologians say that God is neither an existent nor a ground of being, but a human symbol referring to a human ideal. Don Cupitt is amongst those who suggest this, and he goes on to suggest, especially in his later writings, that this must tend to make religion a individual pursuit which is realised in the constant quest for understanding, based on an ongoing theological discussion. We can use traditional religious categories if we like, so long as we bear in mind that these are shorthand ways of talking about the depth grammar, if you like, of human life, where our fundamental beliefs in the beauty and wonder and orderliness of nature, the importance of moral value and the ethical life, and other very fundamental beliefs that give our lives a sense of rational order, get to be expressed. Of course, atheists may turn around and say that this is just hypocrisy, but it is not clear to me that this is a fair accusation to direct at people who are trying to make sense of their deepest values in a naturalistic world. But if the word ‘god’ is used in this context, then the question of God’s existence is neither here nor there, for such people are not making existential claims. They are just saying that religious language still makes sense in our world, because it gives shape and order to the living of a responsible human life This is Philip Kitcher’s point in response to what he calls Militant Modern Atheism, and as time goes on, I think that Kitcher may be on to something.

    1. But, with existential judgements, the issue is otherwise, since to prove a negative existential judgement you need evidence, and it is a matter of the logic of induction that you cannot produce what would have to be both contingent and a priori evidence for a negative claim of existence. That’s the point of “you can’t prove a negative.”

      That’s not entirely accurate. Logically incosistent things cannot exist. You can prove that there is no simple polyhedron with seven edges, without having to produce any “evidence”. In logic and math one can also prove the existence of things for which, provably, there exists no evidence.

    2. Religion as practiced by the large majority of believers bears little resemblance to religion as described by Kitcher. It seems ridiculous for Kitcher to criticize atheists for being critical of religious belief based on his conception of what religion has become, or is evolving into, or might evolve into, when his conception accurately describes, at best, a tiny fraction of believers.

      “But if the word ‘god’ is used in this context, then the question of God’s existence is neither here nor there, for such people are not making existential claims.”

      I’m not sure what the intent of this statement is. But I would like to point out two things. One, it is fairly obvious that the OP is not addressing such nebulous god claims, even without the update explicitly stating so. Two, just as Kitcher’s conception of religion, the large majority of believers do not fit that description.

      Regarding nebulous conceptions of religion such as Kitcher’s, why continue to reference god, jesus, allah, or any particulars of any traditional religion? If you don’t believe in any of that anymore, why continue to pay lip service to it? Or even continue to call it religion?

      1. What is worse is when theists flip-flop between these views. When we atheists argue against God’s existence, He’s Kitcher’s abstraction; at other times they regard Him as being far more concrete. PZ Myers characterised this as GAWD v. GAWR.


    3. Eric MacDonald wrote:

      Of course, atheists may turn around and say that this is just hypocrisy …

      No, I think we would describe the stance of the radical contemporary theologians not as “hypocrisy,” but as “atheism.” The hypocritical part comes in when they start to rail on about something called “Militant Modern Atheism.”

      It’s not “militant:” it’s just clear.

  17. The “can’t prove a negative” response has a certain traction within deism (or when the god in question is so vaguely defined that it’s unfalsifiable). But theists shoot themselves in the foot with this move. The adage is invalidated by specificity. And theists are quite specific about the attributes of their god.

    1. If they define “God” so vaguely that they leave out every single aspect of Mind — God doesn’t think, it isn’t aware, it has no goals, it has no values, it has no emotions, it has no interest in us, it has no interest in anything, it’s not in any way involved in compassion or creativity or justice or morals or “being fair” — then I think they’ve just defined “God” right into atheism. Congratulations.

      They can’t do it. No matter how much they wave their arms about with incoherent bafflegab and obscure blather-about-nothing, sooner or later you’ll see them sneaking in some mental attribute. They can’t help it because that’s the whole point of God being meaningful.

      Then you can catch them with the scientific improbability of mind/brain dualism and the theory that reality is mind-like and the material world came second. It’s no longer a probability, a 50/50 proposition, or even a reasonable option. They’re left with “hope.” Which translates into the refusal to let go of something you want to be true. Which translates not into humility, but arrogance.

      1. But on technical, philosophical grounds, I’d say it’s trickier to dispose of deism. My pet peeve is when theologians (who are clearly theist/Christian) appropriate deistic arguments to buttress rather specific dogmas. Essentially: “you can’t prove a negative, therefore Jesus walked on water.” This bait-and-switch maneuver should be called out more often than it is. I think it was Hitchens who said that the Christian/Muslim/Jew who manages to justify deism still has all his work ahead of him.

        1. Deism is a vague concept imo. I have never heard of a coherent definition of a deistic god and the deistic disposition is often used as a copout because the deist knows that he/she can’t in any way whatsoever prove that there is a god. I see deism as a religious equivalent of non-religious agnosticism.

          In other words; Deism ( and agnosticism for that matter ) is lazy thinking/dishonesty in action.

          1. Deism was understandable during its 18th century heyday. There was not scientific explanation for a universe with life and an earth that sustained such life. So a divine watchmaker was a reasonable idea. Less defensibly, but still understandable, certain commonalities to the moral teachings hinted at a single source of inspiration.

            As rational explanations have arisen for all of these, Deism has been undercut, and really serves no purpose any longer.

        2. I agree that religions with more details (like Christianity) make more claims and are therefore easier to dispose of. So you’re right about that.

          But my own personal pet peeve is when people who are “spiritual but not religious” put forth their vague, wishy-washy, transcendental but clear enough supernatural bullshit and then act like atheists should just fall all over them in gratitude, approval, and forbearance. There can’t be any objection to their half-blown unorganized theology, after all, since they don’t condemn people to Hell or have an actual, technical, religion with a lot of rules — just a love for our spiritual reality and metaphysical ways of knowing.

          Sorry, no. That’s still a religion and it’s still wrong and I’m still going to say so — on technical, philosophical grounds.

  18. From what I’ve seen a large part of theology is devoted to explaining away the lack of evidence for God. Most of the rest deals with Gods personality and various pet peeves and the almost Brazilian bureaucracy that exists in heaven.

    1. Has a nice ring to it, Brazilian bureaucracy. Not that I know jack-nuthin’ about Brazil’s governmental structure. Still. Skootch over and make a little room, Byzantine labyrinth.

  19. Everybody reading these words can prove to themselves that they are not currently being broiled alive by the flames from ten thousand Norwegian Ridgeback dragons.

    And Epicurus, centuries before the invention of Christianity, proved that there are no powerful beings looking out for humanity, using the exact same method. You can confirm his results by flipping the local newspaper to the police blotter and looking for evidence that your favorite deity called 9-1-1 in time to stop the crime in question.



  20. As I always say, “proof is for math and whiskey.” 😉

    The concept of proof as an absolute is of course only going to apply to analytical systems where WE define and set the parameters and definitions and then, like gods, demonstrate that what is on one side of the equation is the same as what is on the other side of the equation. Math is a language with no ambiguity describing worlds we control. In that sense, direct, unmitigated, uninterpreted experience is analytical as well: if you feel your head hurting then you feel your head hurting. True by definition.

    But we do not create the empirical world outside ourselves — so absolute certainty is not possible. The minute you bring in interpretation and other people with possible differences in interpretation you lose definitive “proof.” Which is a good thing. That humbles us.

    But that’s not really the kind of proof they’re talking about with religion, is it? As you point out, they’re playing around with a deepity. (They’re also leaving out the interpretive part of “a direct experience of God.”)

    The “you can’t prove a negative” argument is wrong. You can prove a negative, which means disproving a positive (i.e., God exists)—if you construe the word “disprove” as meaning “showing that the existence of a phenomenon is so unlikely that one would have to be blinkered or perverse to still believe it.” And that is the case for God.

    Agree — and I think this is where religious faith does its dirty work. It defines “disprove” this way:

    “showing that the existence of a phenomenon is so unlikely that one would have to be MOTIVATED to still believe it.”

    And then they screw around with what it means to be “motivated” when it comes to religious and/or spiritual claims. If you believe them this is because you have hope, trust, love in your heart and an open mind. If you choose to NOT believe them then your motivations are dark: you treat the supernatural like any other hypothesis because you don’t have hope, trust, love in your heart or an open mind.

    Or, in other words, the nonbeliever isn’t motivated in the right direction because they are perverse. They don’t want it to be true. It has nothing to do with the line of reasoning from evidence. Place the focus elsewhere so that it’s all flipped around.

    Neat little trick there, isn’t it?

    In a spiritual framework one works backwards from a forgone conclusion, as from the position of a god who has already defined the system. A wise person approaches belief in God as if they were approaching belief in their values. As if they were reaching out to their best friend. As if they were committing themselves to following a virtuous path. They don’t need any “proof” that good is better than evil. They’ll make that leap.

    And the reverse of that screws our rational arguments.

    It seems to me that category confusion is once again right at the heart of this “you can’t prove a negative” bullshit. They’re really saying “you can’t prevent me from believing something if I am motivated enough — and being ‘motivated enough’ is a good thing.”

  21. I think there is plenty of evidence for God if you define him as Jim Holt does in his excellent book Why Does The World Exist?

    God is a hundred per cent malevolent but only eighty per cent effective.


  22. As time has gone on, the gods of religions get more and more vague. The claims of how a god does this and is that melt away under the light of reality.

    Theists whine that their gods exist but even they realize that their gods cannot exist as their predecessors have claims. Their religions depend on constant heresy as they must adjust their nonsense.

  23. “In the case of God, then, the absence of evidence is indeed evidence for His absence. “

    In all cases this is true. Somehow (possibly due to a misprint or misstatement by Carl Sagan?), there is a very common misapprehension that “Absence of evidence is NOT evidence of absence” which is absurd.

    Absence of evidence is indeed evidence of absence, but it is not PROOF of absence.

  24. Professor Ceiling-Cat should be charged with aiding the enemy! It works this way. Everytime you use the word ‘God’ you are subscribing to, and endorsing, any number of religious positions; bringing them into existence, in effect.

    • That there is a widely agreed concept of ‘God’
    • That the idea of God is within the realms of experiential knowledge.
    • That his existence is perfectly acceptable.
    • That there are no glaring inconsistencies in the notion of ‘God’
    • That it is reasonable to discuss God
    • That discussions about God are similar to discussions about other high probability concepts.
    • That discussions about God assume the intellectual probity and reliability of believers joined to the discussion.
    • That there are no enormous prior logical objections to block discussions upon the existence of God.
    • That discussions about God in no way draw upon serious emotional needs underpinning their adherence to positive views about the existence of God.
    • Etc., etc…

    It seems to me far better to refer to ‘their gods’!

    Referring to ‘their gods’ in this way strips away the appearance of widespread public acceptance of the proposition that their gods exist. Referring to ‘their gods’ incrementally undermines their faith and their belief by positing powerful objections to their belief in putting the importance of evidence firmly at the feet of believers. By always referring to ‘their gods’ you take the logical upper hand, and set back their assumption of faith to a hopeless position; almost as if they have to start religion all over again. It is a small but corrosive antidote to religious belief.

    Adapted from ‘Origins of Belief and Behaviour’ – two thousand pages of evidence that religion is a brain disorder!

    1. “… bringing them into existence…”

      Beetlejuice! Beetlejuice! Beetlejuice!

      Oh, and: “• That discussions about God are similar to discussions about other high-probability concepts.” FIFY


      1. Sorry, Ant, but you’ve lost me. What does it mean when you repeatedly shout ‘Beetlejuice’? And what could FIFY possibly mean? It makes this website sound like just another gathering point for a contemporary cult where the initiates develop an arcane language. Best stick to English, eh? Especially for French speakers.

  25. I’ve started to take on this same position myself. If you apply the scientific method to determining if gods exist, you get your answer. Therefore science has killed god. Or as Fred said:

    God is dead. God remains dead. And we have killed him. Yet his shadow still looms.

  26. The most aggravating thing about this, to me, is that pretty much every educated human being on the planet is comfortable using “prove” in the vernacular when speaking about emprical things. We all know and understand the speaker is talking about high confidence, not philosophical certainty. Except when it comes to the subject of God.

    Our culture has a double standard/carved out an exception for God-talk. I can say “I’m wearing a red shirt” and nobody challenges the lack of caveats in that empirical claim. Nobody thinks I’m claiming absolute philosophical certainty about my (knowledge of) shirt color. They’d probably think that interpretation of my statement was just absurd. But if I say “there is no God,” suddenly its completely reversed. Suddenly, they assume I’m making a claim of absolute philosophical certainty.

    That’s a double standard. It bothers me greatly that it exists. Shirts or God, you should interpret any uncaveated empirical claim basically the same way. And the natural way to interpret such statements is as high-confidence yet provisional claims.

    1. Great post.

      On the other hand, God is purportedly intelligent and powerful and could therefore hide all of evidence of its existence/intervention intentionally, a possibility not open to most other empirical phenomena.

      To me this just moves the burden of proof back to Theology where it belongs. Why is God so determined to hide Itself, and yet wants Its existinance to be acknowledged. Why is the “source of all goodness” such a lying, deceitful bastard?

      1. If you really examine the claims of those who agree that God is definitely “hidden” when it comes to evidence and reason, you often find an implicit claim for ESP. God is hiding Itself so that people can demonstrate their capacity to love by tapping into their extra-sensory perception and finding God.

        Those that don’t use their ESP abilities to find, recognize, and relate to God are refusing to use them because they’re too arrogant, opinionated, and stuck on worldly values.

        I think you can rephrase that into the language of most religions. They seldom use the term “ESP.” But the concept is there.

  27. An issue with this topic is that people who bring up “You can’t prove a negative” are often a bit smug about having acquired this bit of pseudo-sophisticated wisdom about the world that they can’t even properly consider that it might be false.

    I brought up something similar to the dragon in the garage at a dinner party the other night, and the disputants just laughed and stated in a louder voice “You can’t prove a negative.” Very irritating.

  28. If you posit a deistic God who doesn’t do anything, or some nebulous apophatic “ground of being” God, then of course you can’t disprove it.

    Thinking about theism vs. deism is useful in that it brings up another concept related to evidence: the concept of boundaries or limits. I may not be absolutely certain that F=ma exactly. Maybe its F=(m+k)a, where k is just really, really close to zero. But using empirical data, I can tell you just how close to zero it has to be, to be consistent with observation.

    Likewise for God; empiricism cannot rule out all possiblities, but it can put limits on how theistic (vs deistic) a God can be and still be consistent with observation. Like the k example above, modern emprical observation tells us that the “theisticness” of God must be pretty close to zero. If a theistic God exists, he is observationally indistinguishable from a deist God.

  29. I don’t get theists arguing for the god of the gaps.

    For the sake of argument, assume god exists, and is consistent with all observations made thus far. How the heck is that the god of any of the major religions?

    1. Because the biggest leap has been made. You’re starting out knowing that God exists and it’s perfectly reasonable to believe in dualism, magic, and miracles. Now all you have to do is quibble over which god is real.

      It’s equivalent to believing in psychic powers but admitting there are a lot of fakers out there. It’s not that hard to show why your favorite is the real deal.

      And if nothing else, the argument is reduced — and placed in their territory. They can even bring in “faith” to supplement a general knowledge based on reason.

  30. God is real, as real as real
    if I stay inside my head.

    It’s only when I step outside
    I realize that he’s dead.

  31. Beyond the obvious arguments of absence of evidence for miracles, for me there are four main arguments that argue against gods of the type proposed by the major religions (some of which have been mentioned):

    1-Can thought exist outside of time?
    2-Can consciousness exist without physical structure?
    3-Can physical structure exist without energy consumption?
    4-Can actions be taken without leaving an energy signature of some kind?

    1. If one assumes a Matrix-style simulation or Alice’s Red King’s dream, the answers to all of those questions are, “yes.”

      However, whilst it is provably logically impossible to rule out the possibility that our reality is simulated, it is equally logically impossible for those running any hypothetical simulation to also rule out the possibility of their own simulation. That is, even if we are inside the Matrix, the programmers of the Matrix have no way of knowing that they themselves aren’t mere pawns in the Red King’s dream. As such, any claims about any sort of “ultimate” reality are incoherent nonsense.

      It’s also worth noting that, whatever its “ultimate” nature, reality as we perceive it is “real” to scales incomprehensible to the human mind. If we ever were to discover that we’re being simulated, that would, in all reality, be no different from Einstein discovering a “deeper” layer to mechanics than Newton’s.



      1. Thx Ben, but it can be argued that those scenarios are really just variations of the god hypothesis, particularly the cosmic dream. They just pass the metaphysical bits off to the matrix or dream instead of a god.

        What they don’t address is thought absent time. I think it can be argued that thought requires some organizational structure anchored in time. How can any entity cognitively act or react to its environment without knowledge constructs consisting of experiences, learning, etc.

        Imagine this thought experiment: all knowledge is flat – there is no time differential amongst any data. Now imagine how any of the data could be processed, accessed, or acted upon without any time differential. How would such a consciousness know whether it was coming or going. All knowledge which we actually know to exist, is cumulative.

        1. You miss my point, methinks.

          Of what sense does it make to describe as divine an entity that isn’t even theoretically capable of knowing whether or not there are other, even more powerful / knowledgable / creative entities? The best you can do is suggest that a god is an entity more powerful than you…in which case we are all gods, and the term is again rendered meaningless.

          Or, if Space Jesus somehow fits the definition of a god, then were James “The Amazing” Randi to set himself up as a god to some back-bush tribe, Randi would fit the definition every bit as much as Space Jesus. Since Randi clearly wouldn’t be a god, no matter what the locals thought, it’s equally clear that no entity, no matter how powerful, can really be a god.

          As to the question of time…remember that it’s just another dimension. I’m sure a clever enough theoretical physicist could devise a self-consistent universe without a time dimension. I’m neither a theoretical physicist nor that clever, so I won’t speculate beyond that.



          1. As to the question of time…remember that it’s just another dimension. I’m sure a clever enough theoretical physicist could devise a self-consistent universe without a time dimension.

            Yes, but our understanding of thinking is still based on time. That’s the anomaly.

            It’s like saying God is pink — but not “pink” in the way we understand the color pink. Oh no, not in a lowly, perception-dependent sense! God is pink in a way that transcends all our limited experience of pink in the universe.

            So why the hell would anyone call God “pink?” Why use a familiar term at all if we’re going to just rip it out of the only context in which it makes sense to us? I think that’s Pliny’s point. Thought is a process.

          2. I would put the emphasis on understanding, not on “thinking.”

            What I’m driving at…well, for example, photons do not experience time, and time also gets similarly fuzzy in other parts of the quantum realm. Without going deeper into quantum woo territory, I’m suggesting that somebody versed in the field should be able to construct a model universe with the capability for cognition but without time.

            As another analogy, as we normally understand it, a triangle can never have more than two right angles. But, if you make your geometry non-Euclidean, you can have two or even three right angles in your triangle. I’m suggesting something similar, only with math and geometry way above my pay grade.



          3. Understanding vs thinking. Interesting. I’m not able to model understanding that doesn’t involve process. Admittedly, a possible limitation of my operating system. 😉

            But in some ways what you are proposing fits with Christian faith – since faith is knowledge (or understanding or belief)absent evidence it could be thought of as understanding absent process. I’ll have to ponder that one.

            As for cognition without time, it would require a method of knowing that is completely different from any of our current models and is untestable based on our understanding of any known thought processes and underlying structure. Absent any evidence that such a thing is possible, aren’t we back to to where we started – a model that may not be testable, but which clearly is contrary to anything we do know?

            Thanks for stimulating the old noodle.

          4. I don’t think what I’m suggesting means there wouldn’t be evidence, just that this hypothetical universe wouldn’t have a temporal dimension.

            Think of the descriptions in popular culture of how some quantum computers work, for perhaps a better example.

            Now, I should hasten to add: I don’t in any way think that this hypothetical actually represents the real state of reality as we understand it. Quite the contrary — it clearly has no bearing on reality whatsoever.

            I’m just suggesting that arguments from geometry are only applicable within systems that actually have the geometry of the system being described. It’s generally not difficult (for certain definitions of “difficult”) to construct a self-consistent geometry in which what is impossible in the one is trivial in another.

            And time is most emphatically a geometrical property of an Einsteinian space….



          5. Sorry, Ben, I did catch the nesting dolls (hall of mirrors?) analogy, which is one with which I also agree. I should have so stated. The god of the old testament fits this bill, I my opinion. Just because he was more powerful than the people of Israel doesn’t mean he created the universe, which is your point, I believe.

            My follow-up was about cognition absent time. While it may or may not be possible to construct a universe absent time, I do not think it possible to create a model of cognition in such a place.

            Admittedly I am biased in this regard. My area of work is in machine intelligence and virtual simulation. Despite the incredible speed (I know, it’s a time-related construct) with which current machines perform, we still have to partition and prioritize data using time-based relationships to process information into useful machine knowledge. Trying to process a mass of data absent time stamps for example removes a critical frame of reference that renders the output useless.

            Interestingly, we do exploit the ‘dream state’ concept in our work. One of the ways we test the AI systems is to disconnect them from whatever ‘real’ system they are operating with and reconnecting them to a simulator that uses the same machine interface (the full extent of senses that can be processed by the machine). Since the AI can’t know if it’s connected to the sim or the real world, it’s a perfect testing environment.

            Perhaps I spend too much time with machines within an artificial matrix 😉

      2. whilst it is provably logically impossible to rule out the possibility that our reality is simulated

        That only applies for logic applied to classical physics. Quantum physics tells us that its logic is “no hidden variables”, so when we stress a causal system for computational resources it will eventually tell us: “hey, I’m, not a whole observable universe, you know!” Or in other words, the simulation will break down at some point, or quantum physics is invalid.

        As I have suggested before, one way to do it may be to ask for increased precision in particle physics. We should know well before the Planck scale whether or not the universe is a “Matrix”.

        But I realize from listening to Susskind’s cosmology lectures that we already have proof. Large scale structure formation is predicted by quantum fluctuations in inflation. No hidden Matrix could have been responsible for the quantum fluctuations in the inflaton field.

        1. Once again, you’re assuming that the simulation is honest.

          ‘Twould be trivial (again, for certain definitions of “trivial”) to put you, say, in a Star Trek-style holodeck where you could run the observations any way you liked, and the instruments would always show you exactly the results the desired measurements, regardless of whatever the real measurements might actually be.

          What we’re really discussing here is Turing’s Halting Problem, and said problem is famous because it can’t be solved. While you’ve outlined all the reasons why we can have great confidence in the “no simulation” hypothesis, I guarantee you that you can’t absolutely rule it out — or else you’d incidentally have a solution to the Halting Problem as well.

          And a solution to the Halting Problem would very likely enable the construction of a perpetual motion machine….



  32. Jerry writes:

    Now mathematicians can indeed “prove” things, for their domain is not the real world but the consequences of a series of axioms. Mathematicans can prove that, in Euclidean geometry, the square of the length of the hypotenuse of a right triangle is equal to the sum of the squares of the other two sides. [my emphasis]

    And that is exactly the same thing as in science. We prove things relative to certain premises. Just as there is no absolute kind of proof in science, there isn’t in mathematics either. The only thing logic can do is to force us to choose to accept certain conclusions or to reject certains premises. In that central fact, there is no difference whatsoever between science and maths.

    1. With math and other deductive systems, the conclusion must be true given a set of premises and operations (like “and”). The parallel is not the case with the empirical sciences; all our observations could be ‘true’ and we could still arrive at the ‘wrong’ theory.

      The “absolute proof” JAC is referring to is, IIRC, the absoluteness of the connection between premises and conclusions, rather than the absoluteness of the premises themselves. The conclusion is a logical consequence of applying the rules to the claims. In induction, not so much. Applying the rules to the claims can eliminate some options, but still always leaves you with a hypothetically infinite number of acceptable conclusions.

      To put it another way: if induction were like deduction, nobody would argue over best fits to the data. There would only be “it fits” or “it doesn’t fit.”

      1. » eric:
        With math and other deductive systems, the conclusion must be true given a set of premises and operations (like “and”). The parallel is not the case with the empirical sciences; all our observations could be ‘true’ and we could still arrive at the ‘wrong’ theory.

        eric, please see below where I explain that science is not, and cannot be, an inductive system, as you imply. What we do in science is, in fact, using deductive logic.

  33. Sort of a corollary to all this from Jonathan Swift:
    “It is useless to attempt to reason a man out of a thing he was never reasoned into.”

  34. ” he’s supposed to be omnibenevolent, omnipotent, and omniscient.”
    – although Atheists do very well at proving that the connecting human-to-human complex relationships can create veil of complication, or work – in ever present ability – to transform it. Can this not also go to say that what we perceive as ever-potential, is strongly blinded/prevented in ability by us, at least to some large degree. Just because we can prove the huge influence of people to people, and difficulty to change, does not mean that change is not always present, or that lack of allows more potential for more harm. Maybe some of the question is not always, is there a God that will break into this strong social webbing, but if there is a God how would the “good-present-now” find ability to work into it all more survivable for sake of preserving?

    1. So what I get from your word salad is that god doesn’t show himself because we don’t believe?

      You’re just spouting the rationalizations that Dr. Coyne takes to task in this post.

  35. On the occassions that I still discuss these things with theists, I ask what would be different in the Universe if their God suddenly disappeared. What effects would we see?

    Usually, I don’t get much in the way of an great response. It then puts me in a position to say that my lack of belief in God is completely rational, as my God-free conception of the Universe seems to be entirely compatible with how it actually is. Why add unnecessary elements like a spooky invisible person?

      1. Actually, your interlocutors may not know their own religion’s official doctrine (depending on their religion, of course). Catholicism and many Protestant versions, as well as Islam, seem to hold that if god were to disappear, the universe would too. That this creates a tremendously nasty version of the problem of evil is usually unnoticed.

  36. “A benevolent God would not kill off humanity in 5 billion years.”

    I think we will be able to avoid that fate in any of a number of ways, if we are still around at that time. Migrating to other stars should be possible. Of course we would run out of stars eventually.

  37. “Why this largesse of uninhabitable space if God created Earth for humans? Even if life exists elsewhere, it can’t be common, and the trillions of uninhabited stars serve no purpose.”

    We have as yet no reason to think that life isn’t common. If stars are widely enough spaced to avoid disrupting each other’s Oort clouds, with unpleasant consequences, there has to be a lot of empty space.

    1. Mathematically speaking there might be a high probability that life is common( how do we define common in this respect btw? ), but this is where we need to insist on empirical evidence to back up the math.

      Imo ,we have as of yet no empirical reason to believe that life as we know it is common.

      1. I think you go much too far in your assertion that there is no empirical reason.

        Indeed, we have every empirical reason to suspect the commonness of at least simple life, save for actual observation of it. And that failure to observe is entirely a result of the limitations of our technology, not because we looked and didn’t find any.


        1. We only have one confirmed planet with life as we know it at the present. There might be an abundance of lifeforms out there, but we still don’t know the exact process of how chemistry became biology. It might as well just be that the conditions required for life to form is very rare and very shortlived. In other words life could be a barely detectable blip on the ever expanding radar.

          One occurrence of carbon-based life in our small corner of the galaxy isn’t enough in my book to earn the label common.

          But I do hope I’ll be proven wrong during my lifetime.

          1. While we don’t know the exact chemical pathway that life on Earth took, and we might never know for the same reason that we’ll likely never know what form the last common ancestor between vertebrates and invertebrates took, we’ve known for generations that the basic foundations of the chemistry of life are about as common and as inevitable as a precipitate forming from a solution.

            What would be remarkable would be to be able to make a thorough analysis of an environment that’s been in the basic ballpark for an hundred million years and not find life. The reason we haven’t found life is because there’s nowhere we can reach that fits that description. The places we can reach aren’t conducive to life, and the places we know about that are conducive are unreachable.



          2. Hopefully we’ll develop the capacity to detect and confirm life relatively soon so the question of life on other planets can be completely resolved and put to rest.

            Until that day arrives though, I’ll stick to the cynical conclusion that thus far, we are one of a kind.

    2. Every planet around every start could be habitable to humans, and the percent of the volume of the universe which is habitable would still be negligible. We would still have that ‘largesse of uninhabitable space.’

      If you don’t believe me, calculate the volume of the solar system and then the volume of habitable earth space in that volume. I roughed it out a few years back and just going out to Pluto – not even the Oort cloud – its on the order of 1E-28%. 0.0000000000000000000000000001%. And that’s not counting the space between solar systems.

  38. Perfect!

    On the empirical evidence I would add:

    – There is no magical action inside closed systems, which there should be if nature isn’t exclusively non-magic. We know enough such systems to make a binomial test @ 3 sigma.

    – There is no magical interaction with the QED sector, which like closed systems there should be because the particle vacuum only excludes interaction specifically forbidden.

    – There is no measurable magical action involved in the universe or its constituents, which there should be if nature isn’t exclusively non-magic. Both the size of the local universe* and its structure formation are caused by spontaneous quantum fluctuations during inflation.

    * Not equal with the observable universe, see Susskind’s cosmology lectures on youtube.

    “You can’t prove a negative”

    Oy! Basic physics proves that there isn’t such a thing as “can’t prove a negative”:

    The universal speed limit excludes that anything moves locally relatively faster.

    Fermion statistics excludes that any other particle occupy the same quantum state at any given time.

    … and so on.

    Proving negatives is just harder, but not forbidden by any law of nature.

    That predicts why excluding magic actions wasn’t the first consequence of applying science. But people started to get there in the 19th century (thermodynamics of closed systems).

    1. Missing piece: “… because the particle vacuum only excludes interaction specifically forbidden and the LHC has now completed the Standard Model that makes all such exclusions.”

      Missing link: Fermion statistics.

    2. Another missing piece: “… no magical action inside closed systems according to thermodynamics, which …”

  39. OMG NO!

    Now even this blessed website has become infected by the dreaded “like” buttons!
    And it’s wreaking some havoc with page loading too.


    Is there no corner of the internet untouched by this scourge?


  40. Oh Great Ceiling Cat, unleash the necessary awesome power I hope to hell you possess to instantly and permanently remove these doggamned ‘like this’ buttons.
    afeline & hallelujah to ya

  41. Happened a few days ago:

    – There’s not one only observation i made that make me think of god existence

    – Well, is proven that people possessed by the devil starts speaking in aramaic or other languages…so what?

    How could you continue such dialogues?

    1. Depends on the context.

      You might be able to ask the person if perhaps one should abandon such childish fantasies after puberty sets in. In more situations, you should be able to suggest as much without being quite so blunt about it.

      And, in all situations, you can, after the fact, and perhaps referring to the person without any other identifying features, suggest that said person really should grow up.



  42. The idea that one cannot prove a negative is a strange one – what’s the difference between “there is an elephant in my fridge” and “there isn’t an elephant in my fridge” in terms of hypothesises? The same evidence will demonstrate the truth of one proposition and the falsity of the other.

    In terms of God, I think the general idea is that one cannot rule out the possibility of God, even though no evidence has turned up from God. This would be significant if the case for God were deductive, but since God is a defeasible concept, then possibility doesn’t really mean much. It’s possible that all evidence for evolution is illusory and that God created the world in 6 days, some 6000 years ago. Defeasibly speaking, however, young earth creationism is long since disproved*. That is to say that the empirical evidence doesn’t support such a proposition, but does support it’s negation.

    *Creationists could try to save this by speaking to the Omphalos Hypothesis – that the world was created with the illusion of age. While this move would “save the phenomena” in that it makes young earth creationism consistent with the evidence, it does so by sacrificing being falsifiable and becomes not even wrong. At that point, it’s dismissed on philosophical grounds as being vacuous, rather than on empirical grounds as being wrong.

  43. In these arguments about whether or not some god or gods exist, I would like to see more use made of the “null hypothesis.” In science, the null hypothesis (the simple negation of any hypothesis) is the default position. Therefore, if you can’t “prove” (i.e, produce reasonable evidence) that the hypothesis is, at least possibly and, preferably, probably true, then the null hypothesis holds. This is the equivalent of saying “the hypothesis” is false (e.g., “one or more gods exist”). Now, the burden of proof rests solidly on the shoulders of those who claim the hypothesis is true. As a skeptic, I don’t have to “prove” that the hypothesis is false. It’s up to the believers to “prove” the hypothesis is true.

    1. Perhaps one problem with some religionists is that they simply cannot imagine that God doesn’t exist, so deeply inculcated is the idea that He does, and so just cannot countenance “our” null hypothesis, even for “the sake of argument”.


Leave a Comment

Your email address will not be published. Required fields are marked *