Letting Go of the Crutch: Probability and physics are on the atheists’ side

A few days ago I was talking to a blogger called Physics and Whiskey. He wrote two posts, one called The atheist’s crutch: Misusing probability and another called The atheist’s other crutch: Breaking the laws of physics. In these posts he makes two clear arguments. In the first he argues that to rebut certain parts of the Bible atheists use an outright wrong type of statistical analysis. The example Physics and Whiskey uses is the resurrection, and the argument he says atheists use is simplified to this: because a resurrection has never been confirmed, the probability of a resurrection is zero. He rebuts this by saying that everything that has happened has happened a first time, and before that had never happened. If we applied this type of analysis we would never be able to accept any claim of anything happening. The prior probability of something that has never happened before happening is not zero is it unknown (or “undefined”).

The second post is about the nature of miracles. Physics and Whiskey separates miracles from criticisms that they break the laws of physics. He argues that because the cause of any miracle is nonphysical (he doesn’t not explain what this means or give examples) it is not bound by the laws of physics. Therefore miracles do not break the laws of physics because they are not bound by them because they are not physical.

The example used to tie these together is an alien landing: what are the chances that an alien spaceship had landed on Earth? No spaceship landing on Earth has ever been confirmed, but it does not mean it is impossible (in fact, we have landed people on the moon and technology on Mars, Venus, Titan and in the gas giant, Jupiter. Voyager I has left the solar system!). So the probability is no zero it is unknown. And there is no way of knowing. There isn’t even anything stopping it from happening—as evidenced by the fact that we have landed in other places.

“There isn’t even anything stopping it”.

That’s important, because when you shift focus from alien landings to Jesus’ resurrection there is something stopping it from happening: physics. When things are dead they don’t come back to life, which is a law in biology that we are very confident of. We are even more confidence of entropy, and resurrection violates that too: once a body is dead is becomes an isolated system and an isolated system is the context where entropy happens. It violated entropy for a closed system to become more ordered, so it violates entropy for a dead system to become an alive system.*

*for those that think this claim would lay waste to abiogenesis you are wrong for two clear reasons: abiogenesis happened in an open system and the level of entropy was wildly different in the “primordial soup” than it is in a human body.

A little revision of statistics for you: the sum of the probabilities of all outcomes equals 1. This means that if there are only two options, and you know the probability of one then the probability of the other option is 1 takeaway whatever the probability you know is. If I have a weighted coin and we play ‘heads or tails’ with it, and the chances of getting heads is 0.7 then the chances of getting tails is 1 – 0.7 = 0.3.

0.7 + 0.3 = 1.

This means that if we can quantify the probability of entropy governing a physical body (like the corpse of Jesus Christ) we can equally quantify the probability of its violation. And, as our confidence in entropy has good reason to tend towards 1, the probability of Jesus’ resurrection tends towards 0. If you don’t like numbers, look at this this way: if we know entropy is true, we know the resurrection didn’t happen. And that is what we mean when we talk about extraordinary claims and extraordinary evidence.

Given that the probability is exceedingly low, and the nature of the event would overturn our understanding of the universe, the evidence better be pretty good. Physics and Whiskey offers none. But he does ask us what we would accept as evidence. I feel that misunderstands the duties in this conversation: what is your evidence? As I said, no evidence is presented. Instead an unsupported conceptual loophole is offered:

It’s important to specify exactly what it means for such a claim to violate the laws of physics. A purported perpetual motion device constitutes a claim that physical objects can be arranged so as to function in a manner contrary to how they are observed to function. It requires belief that tens of thousands of rigorous experiments have all been fundamentally flawed.

That’s important because a miracle claim is completely different. A miracle claim may involve events which are not physically possible, but that’s the point; the appeal is not to physical causes. A miracle does not require that nature operate in a manner contrary to observations, because nature is not what’s doing the operation. Miracles do not challenge our observations about the world. Physical laws describe what we know about physical reality; by definition, a nonphysical cause cannot break physical laws because it is not described by them.

– Physics and Whiskey

The quote that gives away the issue here is this: “the appeal is not to physical causes”. Appealing to physical causes, when the physical world is well documented and evidenced, is not a fallacy. Physics and Whiskey is appealing to something; something nonphysical. This nonphysical world, unbound by the laws of physics, is not evidenced. Therefore an appeal to it is a fallacy. It’s a particular fallacy when what you are doing is begging the question on God. And let’s be sincere, that’s exactly what he’s doing.

If I am mistaken and he is not begging the question on God (by appealing to non-evidenced nonphysical causes to explain unsubstantiated miracles) then this is arguing ignorance: you can’t prove me wrong so my claim is reasonable. The truth is no has demonstrated a miracle to have ever happened, so the claim of the nonphysical is meant to explain… what, exactly?


19 thoughts on “Letting Go of the Crutch: Probability and physics are on the atheists’ side”

  1. Masterful, well-written. I look forward to Physics and Whiskey responding, but I do feel, if my statement is worth anything, that your point-of-view is ironclad.

    1. Thank you for your confidence, it’s very sweet. Physics and Whiskey certainly did reply, and I gave another riposte. Enjoy the read.

  2. Hey, thanks for writing this!

    At the outset, I need to point out that the two different posts addressed two different arguments (the “crutches” that were referenced). Both arguments are unique in that they don’t directly address evidence and probabilities. Rather, they both allege that miracles cannot be considered probable. So, right off the bat, we aren’t dealing with the question of probability (which would require us to investigate evidence and so on), but rather the question of possibility.

    It’s one thing to question the evidence that a particular event happened. It’s another thing entirely to make the positive claim that a particular event could not happen. Like I said, both of the arguments I addressed fall into the latter camp: they are arguments which claim miracles are essentially impossible. So I don’t have to prove that miracles are probable; I just have to prove that they are not impossible.

    The first argument was this: “Miracles cannot ever be considered probable because miraculous events have a prior probability of zero.” This is why I used the example of an alien landing; though we have certainly landed spacecraft on other worlds, we have no prior probability for the existence of alien life, much less an alien landing on Earth. Yet this doesn’t restrict us from considering the possibility (however slim), so an unknown prior probability does NOT make something impossible. I would call this the Fallacy of the Missing Prior.

    The second argument is different: “Miracles cannot be considered probable because they break physical laws.” This is intensely circular. A violation of physical law IS, by definition, a miracle. So all this says is that miracles can’t happen because miracles can’t happen, which doesn’t tell us anything at all.

    The common addendum is to say “Well, we’ve never seen a violation of physical law, so it has a prior probability of zero.” In other words, the “missing prior” business may not work for most events, but it DOES work for events that break physical laws. Of course, this is an unsubstantiated assertion; it needs evidence just like any other claim. Remember, this claim says that miracles cannot be possible; logical impossibility has a really high standard of proof, and it only needs one counterexample to be overturned.

    My counterexample to this argument (which you didn’t mention in your post) was time travel. Time travel is not possible according to the laws of physics as we know them; we have never witnessed a violation of temporal continuity. Temporal continuity has the same history of reliability as any other observed physical law. But this doesn’t prevent us from evaluating the possibility of time travel, and I gave an example in which this would be necessary.

    It’s not question-begging for me to define the cause of miracles as nonphysical, because miracles (again, by definition) are the violation of the physical. If miracles exist, they are nonphysical. That has nothing to do with evidence; that’s just a statement of fact. I was using it to rebut the assertion that miracles are not physically possible.

    Saying that miracles are not physically possible is tautological. If something is physically possible, it can’t be a miracle, can it? It’s like saying you can’t taste light or see sounds: true, but utterly unhelpful.

    Let me return to one particular quote:

    Given that…the nature of the event would overturn our understanding of the universe….

    The use of the word “overturn” here is extremely misleading. Does the nature of miracles add to our understanding of the universe? Absolutely. But nothing is overturned, because nothing about our understanding of the universe is wrong. It’s like peeople who are born deaf and given cochlear implants: their understanding of the universe wasn’t wrong, just incomplete. The existence of miracles merely says that there is another facet of reality that we have not yet explored; it doesn’t say that the portions we HAVE explored need to be thrown out.

    One more quote, and I’m done.

    the evidence better be pretty good. Physics and Whiskey offers none. But he does ask us what we would accept as evidence. I feel that misunderstands the duties in this conversation: what is your evidence?

    Understand that asking “what would you accept as evidence” doesn’t defer the burden of proof. It merely attempts to characterize the bounds of the debate. In my experience, atheists quite often maintain unfalsifiable presuppositions; determining what sort of evidence would be sufficient is a necessary step in testing those presuppositions.

    1. Let me just jump in here before Allallt picks up the bat. You believe Jesus performed miracles, correct? Could you perhaps explain then why this rather talented individual didn’t know that Moses ever existed?

      1. Seems like a bit of a derailment, given that the topic is standards of evidence and questions of possibility, but that’s okay.

        I’m puzzled by your question. My younger brother is an incredibly skilled guitarist and musician in general, but he is not particularly familiar with the physics of audible acoustic waves. Nor is he familiar with the mathematics of harmonic frequencies. Should I see a contradiction in this?

        1. I don’t want to distract from the topic, so i won’t engage too deeply, but you claim your Jesus could perform miracles which would mean he was superhuman. My point being, wouldn’t a superhuman also know history…. particularly his “own” history of revelation? Question being: can you reconcile Jesus not knowing his own history?

    2. I appreciate the time you’ve taken to reply.
      To address the points in no particular order, miracles do overturn our understanding of the universe: where are certain inviolate rules have given us an idea of what is not possible, the demonstration of miracles would mean that all things are possible.
      If I can, I’d like to try and offer a definition of a miracle that we can work with, as we’ve been working with fragments at a time. Given the fragments we have been using, are you happy with the definition “when a nonphysical cause effects a phenomenon or phenomena that violates the laws of physics”?
      Also, as the content of the laws of physics have changed significantly over time, are you happy with the definition of the laws of physics as “a set of laws that best represent our understanding of the universe by repeatable observation”? (My reasoning being that saying the “laws of physics” are the list of laws we know now means we are arguing largely in a darkened room of ignorance, and any novel or expansive observation would be a miracle). I have given some ground on this definition of the laws of physics. The laws of physics could be defined as a set of principles (or—given the elusive Grand Unified Theory—the single principle) that govern the universe. That definition defines God and miracles right out of the picture and the actual content of those principles is beyond our understanding. I hope you appreciate the concession.
      We may also have to define “nonphysical”, because it would be easy to interpret light or even the laws of physics as nonphysical. Although we both appear to be agreeing to it being a cause of pure will, like a mind, neither of us have actually formalised it. Without doing this, I’m still not sure what it is we’re talking about. If you are talking about will or a disembodied mind then our conversation has to take a brief digression about whether freewill is a coherent idea, unbound by the laws of physics. (https://allallt.wordpress.com/2012/11/18/freewill-the-neurological-weather-pattern)
      A miracle, if you accept these two definitions, would be an event where study of the event would give us no repeatable natural or physical laws to work with (which separates them from learning new laws, which would then be incorporated). Miracles would also be a violation of those laws.
      My overall point is that your two posts complement each other in content, but then ignore each other. We have good reason, from Bayesian probability, to place a high prior probability on the laws of physics being inviolate. It follows then that we place a low prior probability on the laws of physics being violate (as their total must be 1—I do not see that this commits the fallacy of the missing prior). The latter would fit a part of the definition of a miracle: “phenomena or phenomenon that violates the laws of physics”. To make this possible, we assume the rest of the definition “a nonphysical cause”. (That said, if two boulders collide and their collision sprays Skittles™ over the surrounding villages, I’m not sure we can assume a nonphysical cause although the event certainly looks to be a miracle.)
      If the probability of the laws of physics being inviolate is 1, then the probability of miracles is 0.
      I cannot find any defence of the idea that the laws of physics are inviolate other than they’re consistently un-violated. So our confidence definitely tends far towards 1. I think you need to (and would appreciate it if you did) provide some sort of idea of what a nonphysical cause behaving independent of the laws of physics is. When I have a better idea of what it is, I may be able to articulate an idea about what would constitute evidence for that. Because short of time travel to go and examine Jesus’ body (including full blood work… just in case that bitter wine is implicated in this somewhere) and then observe his tomb for the three days until it is founds empty, I think proper analysis of the events is possible (which is the nature of the discipline of history: (https://allallt.wordpress.com/2013/02/15/history-cannot-prove-jesus-resurrection)
      I may address the time machine point at some time—both in evidence and probability/possibility terms (such a post certainly isn’t scheduled). My plan is to try and research whether time travel is prohibited by the known laws of physics (and how reliable those laws are) or whether it is simply not obvious whether any part of the laws of physics permits time travel (which is more a symptom of an incomplete picture of physics). But this post snuck over the 1000-word count, and I don’t like that for my posts.
      Thank you for getting this far, and sorry that my comment is nearly as long as the post itself.

  3. Yet this is what the skeptical atheist must maintain. Why? Well, we don’t have any past instances of alien landings on Earth, so we have no way of assessing the probability of the alien option. Clearly, then, there’s no reason to consider it a live option. No matter how strong the evidence for an alien landing was, the stubborn misapplication of prior probability would lead the atheist to insist that it was not an alien landing.

    1. You have missed the nuance of my post. To say that we have no evidence that a thing has happened before, therefore the probability of it ever happening is 0, is wrong. It does what my conversational partner, Physics and Whiskey, has called the “fallacy of the missing prior”. Statistics do not work that way.
      My post is about whether we can establish the “missing prior” by establishing other probabilities first. An alien landing doesn’t, so far that I can see, violate anything–so there is no other related probability to establish.
      Whereas miracles do violate something: the laws of physics (there is a good conversation about that in the comments above). Therefore, if we can establish some sort of prior probability for the laws of physics being inviolate, we can establish the related probability of them not being…

  4. To draw a correlation between alien landings and resurrection of a dead body is disingenuous.
    One is a distinct possibility as it remains firmly in the physical realm whereas coming back from the dead violates every known natural law.

    This is Physics and Whiskey doing the classical “Let’s see if I can fool them with pseudo intellectual bullshit”
    Politicians do it all the time, and so do the likes of William Lane Craig. By the time they have finished one is reaching for the aspirin and wondering what the hell was the question you asked in the first place, and such twits retreat with a ”Yes I am a smug bastard” look on their face.
    Allowing P&W the air space to yank everyone’s chain in this manner merely panders to his infantile ego.
    Much more intellectually honest if he just came and said “I am a Christian” end of story.
    Until then, he is just a dickhead.

    1. His initial point, that he has coined “the fallacy of the missing prior” is a genuinely good point when it comes to Bayesian probability. But you are right in saying that there is an extremely high prior probability that the laws of physics remain constant. This post is about expressing the difference (and calling out the bullshit for what it is, which is important because it slowly arms the audience against the claims that we are stubbornly refusing arguments).
      You are right that Phys Whisk’s argument looks like desperate clawing at excuses, and compared to his normal good work, it is.

      1. It is a constant frustration to read material from an obvious erudite and intelligent person such as P&W who then clings to religion as a child might cling to a security blanket.
        If one eschews religious fundamentalism what has mainstream or liberal theology got to offer? And in this case Christianity. Its core tenets are the same whether one is a raving fundie or a laissez faire believer.
        Jesus is the god,God.
        It is truly baffling.

  5. Just spotted P&W’s blog and appreciate your response to him. I think the difference in prior probabilities between a miracle and a UFO crash is a red herring though. No matter how low the prior, an equally low probability of the evidence occurring if our hypothesis were false will still give us a 50% posterior probability of the hypothesis being true.

    So, I think we’re getting dragged into a debate over the minutia of whether a miracle is more unlikely than a UFO, when really it doesn’t matter (except as a fun mental exercise for us geek types). They’re both incredibly, ridiculously unlikely, and in both cases we have nothing like the equally ridiculously unlikely evidence that P&W describes in his story (the unknown writing, ancient records of advanced science, etc…). If we did, scientists would be taking the idea of miracles and UFO crashes seriously.

    1. A Bayesian statistical analysis can give you an idea of the probability of a thing happening. Part of what you need to know to do the maths for Bayesian analysis is the past-nature of like-events happening. If a like-event has never happened, however, the maths does not get get a 0, it gets an undefined value.
      The alien landing is a good example: if you want to find the probability of aliens landing you’ll come up against no evidence and therefore you’ll get this ‘undefined value’.
      The problem is very different for a resurrection; you come up against a mountain of evidence to the contrary. Resurrections violate laws we very have exceptionally good reason to believe are air tight (entropy). We therefore have exceptionally good reason to believe resurrections don’t happen. Therefore the prior probability does tend strongly towards 0. In fact, you have to doubt entropy in order to make room for potential belief in the resurrection.

      In terms of the overall conversation of whether Jesus was real and resurrected, you are right to say that this conversation is a red herring. (I have better discussions about it on my blog: https://allallt.wordpress.com/2013/02/15/history-cannot-prove-jesus-resurrection and https://allallt.wordpress.com/2013/08/14/jesus-resurrection-best-explanation-of-the-evidence.) However, the conversation itself is worth having. It is not just about probability, but possibility. And that can be very revealing.

      1. We’re in agreement on pretty much all points. The only thing I’m not sure about is how you’re getting at the idea that the prior probability of a UFO crash is undefined (maybe we’re talking about observed frequencies as opposed to hypothetical frequencies?). I definitely agree the prior probability of a miracle is far lower than that of a UFO crash though. I’m just not sure most people would see that as immediately obvious since, in both cases, we’re talking about numbers that are far too big for people to wrap their heads around without using some math (e.g. “what’s the difference between 10^300 and 10^301?”).

        Arguing about the impossibly low prior for a miracle sounds a bit like saying that a miracle is impossible (which most people would object to, for basically the same reasons P&W did). Whereas pointing out that low priors can be overcome by equally low consequents IMO sounds more like saying that miracles are really unlikely, and to prove one you need equally unlikely evidence (which most people would readily agree with).

        Looking at it from that perspective P&W’s argument is actually an argument *against* miracles (since most everyone would agree that miracles are at least as unlikely as UFO crashes, and it should be obvious that we don’t have nearly as much evidence for miracles as we would for UFO crashes in the situation that P&W describes).

        1. You raise good points.
          The point I’m making is that UFO crashes are more likely than miracles.
          For a UFO we have to simply claim that we have no idea. Our closest alien society may have invented travel last week and be en route–meaning a landing is due and could become frequent. Or they may not exist. The data simply isn’t in.
          For a miracle we have to claim that inviolate laws have to be violated. It’s not just that that has never happened, but that there is a truth that inhibits it.

  6. I agree with nearly everything you’re saying, but I’m also saying it is possible to calculate the prior probability of a UFO landing or a miracle, despite the fact that we have zero observations of either event.

    To think about it another way, consider that the sun rising tomorrow is an event that has never been observed before (specifically that event, the sun rising tomorrow, not some other day). But intuitively we know that the probability of the sun rising tomorrow is not zero, nor is it undefined. We sense that the event “the sun will rise tomorrow” is extremely probable, even though that event has never been observed before (nor will it ever be observed again).

    Another example would be if you opened a board game and removed one of the dice. Before you roll the die, you would say the event “I roll a 6” is about 17% probable, even though the observed frequency of that particular die coming up 6 is 0 (it’s never been rolled before). But your intuition isn’t going off the observed frequency. You’re going off the hypothetical frequency. Your intuition is right, and that could be demonstrated empirically. We could roll a bunch of never-before-rolled-dice, and test the results against our hypothetical frequencies versus our observed frequencies to see which is more accurate. Also, as Richard Carrier pointed out in Proving History, if we averaged enough observed frequencies they would start to converge on the hypothetical frequency.

    In the case of the UFO, the hypothetical frequencies are far less obvious than they are with a simple die, but we could still deduce them (Drake’s equation gets us started in the right direction, even if it doesn’t give us anything close to a useful answer).

    Likewise, there has to be a hypothetical frequency for miracles (assuming here, that we mean a suspension of the laws of physics), and that hypothetical frequency will be more accurate than the observed frequency. I agree with you though, that the prior probability we calculate for a miracle is going to be less than that for a UFO landing (far, far, far less), and for all the reasons you’ve already provided.

    [ note: I actually decided to do a brief experiment using this [http://www.wizards.com/dnd/dice/dice.htm], because I’m waiting for my project to render and I’m bored.

    I did 6 runs of 6 rolls each, resulting in the following:


    Which if we assume the observed frequencies are the true probabilities of rolling a 6, we get 5 different answers.


    I know the real probability of rolling a 6 is 1/6. This means the observed frequency only gives me a correct answer 1 out of 6 times, or roughly 17% of the time. If I average all the observed frequencies together, I get a revised probability estimate of .83/6 or roughly 14%, which is actually much closer to what I know is the real probability of rolling a 6.

    I then did 6 more runs of 6 rolls each, resulting in the following:


    Which yields the following probabilities (again, if we assume the observed frequency is the probability)


    So, once more, the observed frequency was only the correct probability 1/6 or roughly 17% of the time. If I average the results from both runs, I get a revised probability of estimate of 1.83/12 or roughly 15%.

    So, the more observations I made, the more the observed frequencies converged on the hypothetical frequency. From this, I think we can conclude that the hypothetical frequency is a more accurate measure of the true probability, but also that any scientific field where we have overwhelming amounts of empirical data (i.e. observations), the difference between the hypothetical frequency and the observed frequency will approach zero. This means there will be close to no probabilistic “wiggle room” in something like physics. Trillions and trillions of observations entails the probability of a well-established physical law being correct approaches 100%. And that confirms your conclusion. If physics says Jesus couldn’t have resurrected, then we are justified (infinitely close to 100% justified) in being confident that Jesus was not resurrected. ]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s