“You can’t prove a negative”

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.


  1. Morg
    Posted October 14, 2013 at 2:10 pm | Permalink

    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?

    • Posted October 14, 2013 at 2:15 pm | Permalink

      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.



  2. kelskye
    Posted October 14, 2013 at 2:19 pm | Permalink

    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.

  3. John R. Vokey
    Posted October 14, 2013 at 2:59 pm | Permalink

    On the oft-used canard that one can’t prove a negative, you might want to read the philosopher Tony Pasquarello from the Skeptical Inquirer (1984). Here is the link:

  4. Dale Franzwa
    Posted October 14, 2013 at 10:19 pm | Permalink

    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.

    • Posted October 14, 2013 at 10:47 pm | Permalink

      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”.


  5. Diane G.
    Posted October 14, 2013 at 11:57 pm | Permalink


One Trackback/Pingback

  1. […] have a lot of time to do my usual vomit of atheist thoughts. So instead, I wanted to link something that Jerry Coyne brought up a few days ago on his awesome blog, Carl Sagan’s Parable of The Dragon In My […]

Post a Comment

Required fields are marked *


Get every new post delivered to your Inbox.

Join 27,051 other followers

%d bloggers like this: