Friday, November 27, 2015

Atheism, Theism, And the Burden of Proof

I'm going to say something that should be uncontroversial among people who take pride in being rational: one's beliefs must conform with reasons for belief. Quite simply, this means that rationality itself places a burden of proof on everyone to have reasons for their beliefs. But a strange (and frankly, embarrassing) thing happens to many atheists when they enter into a dispute about the existence of God. Suddenly, atheists who would normally agree with what I wrote above, claim that the burden of proof lies squarely and only with theists. Apparently, according to these atheists, as soon as their belief is in question, the burden is entirely on the other side.

How do these folks support the claim that the theist has the sole burden? Their argument goes something like this:

Premise 1: The burden of proof is always on the claimant
Premise 2: Theists are making the claim that God does or probably does exist
Premise 3: Atheists merely lack belief in God's existence and, as such, are making no claim about it.
Conclusion 1: following from (1) & (2), theists have the burden of proof regarding the existence of God.
Conclusion 2: following from (1) & (3), atheists have no such burden.

This would be a valid and sound argument, and a very dandy one for atheists, except for one problem: the claim that atheists just lack belief in the existence of God is often misleading. It fails to accurately describe what most atheists usually think about the probability of God's existence, namely, that it's unlikely. Here's the rub: these atheists think that God probably doesn't exist, and that's a claim about the existence of God. It therefore follows from Premise 1 that they do indeed, have a burden of proof.

So who has no burden?
The only person with no burden regarding a claim is the person who hasn't had enough of a chance to think about the truth or falsehood of the claim and formulate a belief either way. If you haven't had a chance to really think about whether taking in Syrian refugees right now is a good idea, you could say that you "don't believe" that taking them in is a good idea, just as you could say that you "don't believe" that not taking them in is a good idea. You really don't know what to think. When in the psychological state of not knowing what to think about a claim, you really have no burden because you really are making no claim. Please notice that this position poses no challenge to a claim.

The atheists I'm addressing in this post often make the preposterous announcement that they lack belief in God the way a baby or a dog does. While it's true that babies and dogs haven't thought about the probability of God's existence, and therefore make no claim and assume no burden, I can hardly see why atheists would want to bring agents that lack the cognitive abilities to even weigh in on the matter into the fold. I encourage folks who talk like this to think carefully about claiming that they are like opinion-less infants.

There are only 2 ways to challenge a belief
As an atheist, in order to say that you don't believe in God and have that mean anything, you really have to have thought about it and decided that ascent to the claim would be intellectually wrong. You can fail to accept a claim by reaching one of two conclusions:

1. It's probably false. Believing that a claim is more likely to be false than true is a very good reason to not believe the claim. Arguing that a claim is false is known as making a de facto objection. The idea is that there is a fact of the matter regarding the claim and that fact is that it's false, or at least more likely to be false than true.

2. It's unjustified, or irrational. This is a different sort of objection that has nothing whatsoever to do with the veracity of the claim. Rather the objection is that, whether it's true or false, it's unjustified or irrational to believe. This is what's known as a de jure objection.

Here's an example. Say that I produce a lunch box and tell you that there is a hockey puck in it. You could rightly ask why I believe that. Did I look in it and see a puck? Did a reliable source tell me that there is a puck in it? Did I X-ray it and find a puck? If I answer in the negative to all of these sorts of queries, you would rightly question why I claimed that a puck is in the box in the first place. The objection here is that while there could be anything in the box including a puck, the reasons for believing that there is a puck in the box fail to justify that belief, ie. it's irrational.

So atheists have to decide what type of objection to theistic belief they have, and then they have a burden to defend that position. If an atheist thinks that the probability of God's existence is roughly 50/50, they can still advance the damning de jure objection that theistic belief is irrational, and I have absolutely no problem with that, even though most people (myself included) would probably call such a person an agnostic, rather than an atheist. But what atheists shouldn't do is believe in and make de facto objections to theism and then, when challenged, shift the burden of proof onto theists by retreating to a de jure objection, or even worse, to the preposterous position of claiming to have no belief whatsoever regarding the question of God's existence, like a newborn baby. To anybody looking upon this debate with fair eyes, these two moves look lame and shifty because they are.

The guys and gals who behave in the way I am spotlighting are very real and very strongly committed to their fallacious position. In it's defence, you'll hear them say things like:

"The burden of proof is always on the one making a positive claim."
Since the claim that God probably doesn't exist is a negative claim, they are relieved of their burden, or so they assert. But a little reflection reveals that this just isn't true. Imagine that I say that you should take an umbrella to work tomorrow because it's probably going to rain (a positive claim) and you say that you shouldn't because it probably isn't (a negative claim). Am I really the only one among us who has to have a reason for my particular belief? Would you automatically be rational despite having no reasons whatsoever for believing that it probably won't rain? That's nonsense. And besides, negative claims can always be rephrased as positive claims and it's absurd to think that merely rephrasing the same idea suddenly imposes a burden to support it. Here, watch:

Me: "It's not going to be a dry day tomorrow (negative claim), so you better bring your umbrella"

You: "Oh yes it is (positive claim). I'll leave my umbrella at home, thank-you."

By just rephrasing the claim, your negative claim is now a positive claim. Are we really supposed to believe that this rephrasing has suddenly switched the burden from me onto you?

So much for Modified Premise 1. You've got a claim (positive or negative)? You've got a burden.

"But You Can't Prove a Negative!"
One can prove a negative by finding a logical inconsistency in it. For example, I can prove that married bachelors don't exist because they can't. The logical problem of evil represents an attempt to prove that God doesn't exist because of the logical inconsistency posed by an omnipotent and morally perfect God in the face of evil and suffering in the world. Whether this logical argument is successful is another matter, but you get the idea.

One can also argue a negative evidentially. When evidence of a certain kind is expected given a particular claim, the absence of that evidence makes the claim less likely to be true. In other words, sometimes, absence of evidence really is evidence of absence.

For instance, if, whenever you bake a cake, the kitchen smells of baking, then it's reasonable to conclude that you probably didn't bake a cake whenever the kitchen lacks the scent of baking. There are several excellent evidential arguments against the existence of God (I explain a particularly powerful one here) and I encourage atheists to use them.

"No, you're wrong, and Russell's Teapot settles it"
In 1952, Bertrand Russell, a Nobel prize winner and influential philosopher of the 20th century, wrote a paper entitled “Is there a God?” wherein he outlined why he doesn’t believe. Included in the paper is this famous quote:
"Many orthodox people speak as though it were the business of skeptics to disprove received dogmas rather than of dogmatists to prove them. This is, of course, a mistake. If I were to suggest that between the Earth and Mars there is a china teapot revolving about the sun in an elliptical orbit, nobody would be able to disprove my assertion provided I were careful to add that the teapot is too small to be revealed even by our most powerful telescopes. But if I were to go on to say that, since my assertion cannot be disproved, it is intolerable presumption on the part of human reason to doubt it, I should rightly be thought to be talking nonsense. If, however, the existence of such a teapot were affirmed in ancient books, taught as the sacred truth every Sunday, and instilled into the minds of children at school, hesitation to believe in its existence would become a mark of eccentricity and entitle the doubter to the attentions of the psychiatrist in an enlightened age or of the Inquisitor in an earlier time."
Don't you just love that picture? That's the look I envision on his face when he reminds theists in that quote that it is not enough that theistic belief is widespread and that they therefore have a burden of proof. That's not in dispute. What is in dispute is whether the notion that theism is probably false also has a burden, and nowhere does Russell suggest that it does not.

Russell's teapot is meant to prevent theists from employing the fallacy of "shifting the burden of proof" onto atheists, and that is all. Ironically, atheists who think that God probably doesn't exist and then claim that the burden is entirely on theists to argue that he does are as guilty of shifting the burden of proof as the theists that Russell was scolding. A case of the teapot being as black as the kettle, perhaps?

"But ... The Legal Burden of Proof!"
It is true that in criminal cases, the one charging another with breaking the law (ie. the state) has the burden of proof. The defendant has no burden to prove that she is innocent because she is presumed to be (which doesn't mean that she is, of course). At a minimum, all that the defence must do is show that the prosecution's evidence is weak and therefore raise a reasonable doubt about the accused's guilt. Things are this way to prevent the state from abusing its power. Can you imagine if the state could accuse somebody of a crime and punish them unless they could prove themselves innocent? No person would be free of the threat of that kind of tyranny.

It is noteworthy that the verdict in a criminal case is either 'guilty' or 'not guilty'. Courts never reach a verdict of 'innocent' beyond a reasonable doubt and it would be disingenious for the defence to claim that the defendant was innocent if all they argued was that she was not guilty. Why? Because they would not have met their burden for *that* claim. OJ Simpson was found not guilty, but that clearly didn't mean that he was innocent. Similarly, atheists who believe and claim that theism is probably false are disingenuous if all they do is argue that theistic belief is unjustified or that they lack belief the way a baby does. Why? Because they aren't meeting their burden for the claim that theism is probably false, which is what they really believe.

Outside a criminal court, in discussions about the existence of God, or who would make a good President, or whether the minimum wage should be raised, etc. the burden of proof that matters is the "Philosophical Burden of Proof" and it applies equally to both sides of a claim for we don't presume that a claim is true or false. Accordingly, both sides can be guilty of shifting their burden fallaciously when they try to avoid it.

So there is a reason that criminal courts have an asymmetric burden of proof, but they nevertheless do not reach conclusions beyond what is successfully argued for. If atheists want to avoid their burden of arguing that theism is probably false, they too should not reach or hold conclusions beyond what they are prepared to argue for.

"Extraordinary Claims Require Extraordinary Evidence"
This quote from Carl Sagan is true, to be sure,  but it's also a word game that atheists who want to avoid their burden of proof like to use and it's really quite easy for the theist to dismiss. All the theist must do is ask what makes God's existence an 'extraordinary' claim.

The answer can’t simply be that the evidence and arguments in support of God's existence are insufficient. That only permits one to say that belief in God is unjustified, and while the atheists I'm addressing in this post do think that belief in God's existence is unjustified, they also believe much more than that. They also believe that God's existence is unlikely, and that's a claim that they have to be prepared to back up. You see, saying that a claim is extraordinary is just a dramatic way of saying that it's probably false. The low probability of it being true is what makes the claim that it nevertheless is true, extraordinary. Switching in the word "extraordinary" for "unlikely" doesn't magically make the burden of showing God's existence to be unlikely disappear. And so, when the atheist explains why they think that God's existence is "extraordinary", they will then be addressing the burden of that claim.

"I Don't Have a Burden Because I'm Not Trying to Change Anybody's Mind"
There are 2 responses to this. The first is that if you are involved in a discussion about who has the burden of proof, you are necessarily talking about situations where at least two people disagree and are trying to convince the other of their error. The second is that you have a burden to yourself, in order for your beliefs to be rational, to make sure that they are justified.

"I'm Not Making a Claim to Knowledge"
Here's some important news for anybody tempted to use this canard: if your reasoning is evidence-based, you are always dealing with probabilistic beliefs. Whether you know it or not, when you update your beliefs based on new evidence, you're employing Bayesian reasoning, and the result of Bayesian reasoning is always probabilistic. The difference between a hunch, a belief, and knowledge is just a matter of the probability you assign to the truth of the claim. Your evidence-based reasoning doesn't suddenly acquire a burden of proof when the probability you assign to the truth of a claim reaches whatever threshold you have for calling it knowledge. If you have a hunch, you have a burden. If you have a weak belief, you have a burden. If you have a strong belief, you have a burden. If you think that you "know" something, you have a burden. These burdens are not all equally heavy, to be sure, but they are all very real and shouldn't ever be ignored.

"If I Have To Disprove God, then I Have to Disprove Everything"
This is an obvious non-sequitur. It just doesn't follow from the fact that you must have reasons for your beliefs that you're obliged to anticipate every potential claim and disprove it. You just have to have reasons to support what you've come to believe.

"Prove that unicorns that ride rainbows and fart glitter don't exist"
This was an actual response that I received when I merely suggested that atheists who think that theism is probably false carry a burden of proof. This challenge seems to be a combination of "You can't prove a negative" and "If I have to disprove God, then I have to disprove everything" rolled into one. I suspect that the reasoning was that if one isn't capable of meeting this burden, then it's completely unfair to tell any atheists that they have a burden, too. Here was my response:
"Unicorns are large horse-like animals. Humans have pretty much searched all of the possible habitats for large horse-like animals and reliable evidence of the existence of unicorns, including unicorn remains, unaltered photographs, caged unicorns, etc. have never been produced. Since we could very reasonably expect to find such evidence of the existence of unicorns if they actually existed, the fact that we have not is very powerful evidence against the existence of unicorns. Since we have very powerful evidence against the existence of unicorns, we also have very powerful evidence against the existence of unicorns that ride rainbows and fart glitter. The existence of those specific unicorns, with no evidence in the tree of life of any other animals with such capabilities nor any reason why those capabilities would've evolved, is astronomically unlikely. Furthermore, the notion of  a material being riding a rainbow seems incoherent. The extremely low probability of the existence of such unicorns constitutes what I would consider to be proof that they don't exist. As an epistemic fallibilist, I always leave room for the possibility of being wrong, but the matter has been established to my satisfaction: unicorns that ride rainbows and fart glitter don't exist."
There are some folks who claim to be agnostic about God the same way that they are agnostic about unicorns. I hope it's clear that we don't have to be agnostic about unicorns.

A Way Forward
Because the term 'atheist' is vague and fails to identify whether one has de jure or de facto objections to theism of varying strengths, I propose that people just state what they believe regarding the particular God(s) in question, and how strongly they believe it. In his book, The God Delusion, Richard Dawkins rightfully acknowledged that our beliefs are held with varying degrees of confidence and provided a seven-point scale of belief (more on that here). Notice that the only real "default position", which isn't a position at all, but rather, is the psychological state that exists when someone hasn't yet formulated a belief, isn't even on the scale. Once one forms an opinion about the existence of particular God(s), then atheists and theists alike should figure out where on the scale they sit, and then defend that position. An atheist who's only prepared to claim that theism is unjustified should probably identify as a 4 on that scale. Such atheists really don't have a burden to prove that theism is more likely false than true because they're what most people would call agnostic and are therefore making no such claim. If an atheist identifies between 4.1 and 7, then they have a burden to explain why they think that theism is more likely to be false than true, just as a theist who identifies as a 1-3.9 has the opposite burden. There is no special burden of proof that one side has that the other doesn't, and perpetuating this lie creates a toxic situation where both sides try to shift their burden onto the other. I never want to see my fellow atheists trying to shift their burden onto theists by advancing the lie that they're opinion-less infants, or that their only burden is to reject the evidence and arguments in favour of theism. If you think that God's existence is more likely false than true, you obviously have to do all of that plus more. My request is one that no good atheist should ever resist, for all I'm asking is that you defend what you really believe.

If you want to become familiar with arguments, including evidential arguments, against theism, I suggest following Justin Schieber at Real Atheology, and on Twitter and Facebook.


  1. Yorgo,
    Excellent post. I completely agree and share your disappointment in those who deny the need to defend their position. Your response to the glitter farting unicorn challenge was spot on. And I was glad to see you bring Bayesian reasoning into the discussion, since it seems to me that a failure to recognize this lies at the root of the problem. Which brings me to my primary reason for commenting: when you made the distinction between a de facto objection and a de jure objection, my immediate thought was that these are really the same thing from a Bayesian perspective. In other words, the assertion that there is a hockey puck in the lunch box is irrational because all of our background knowledge tells us that, in the absence of other evidence, lunch boxes very rarely contain hockey pucks - so "irrational" and "probably false" are one in the same. Perhaps the distinction is one of relative background knowledge? If you grew up in a town where lunch boxes were only ever used to transport hockey pucks and had no knowledge that this was unusual elsewhere, then I would suggest that you are actually being rational by believing that a lunch box contains a hockey puck even though it would seem to be an irrational claim if I were to have made it. So I guess the question is whether you think there is really any useful distinction between a de facto objection and a de jure objection once a Bayesian perspective is employed?

    I'm also interested in your thoughts on the role of Ockham's Razor in this kind of discussion. Bayesian reasoning is largely dependent on induction. That is, we can assign non-uniform probability distributions only because past experience tells us how frequently certain things occur. When it comes to the question of God however, the proposition describes an entity which generally defies induction and lacks the objective, empirical grounding that enables us to compare opposing views. The first example of this comes in the form of Plantinga's claim that God belief is "properly basic". There is a claim to some sort of subjective experience or feeling which is attributed to God and then used as the foundation for profession of a divine realm. Another example is the "invisible dragon" of the dualist, who insists that the mental must be distinct from the physical no matter how strong the correlation with the physical because the 1st person and 3rd person perspectives of neural states are so different. Here again we have a foundation for an immaterial realm in which God can be placed, making him immune to the kind of reasoning you employed with the glitter farting unicorn. In either case, the claimant then has the background in place to support the rationality of their position (just like the person who grew up in a town with a predilection for transporting hockey pucks in lunch boxes). My take, however, is that this doesn't put the claimant on equal footing with the pure physicalist because, assuming they aren't pure idealists, the non-physicalist has multiplied entities; and Ockham's razor seems to be a type of Bayesian reasoning - which I believe is formally described by Solomonoff Induction (though I must admit I haven't scrutinized this enough to be confident in that assertion). Have you given much thought to the role of Ockham's Razor in this kind of situation?

  2. Hi Travis. You consistently provide the best questions and discussion over here, so thanks, as always, for your interest.

    From a Bayesian perspective, I think it’s important to recognize that ascent to a belief is only rational when the probability that it’s true > the probability that it’s false. So a de jure objection to belief would occur when the probabilities of truth and falsehood are equal or inestimable. De facto objections of varying strengths would occur when the probability of falsehood exceeds 0.5. If we were just to state our probabilites, share evidence, and update accordingly, always talking about probabilities, we could avoid using the terms ‘de facto’ and de ‘jure’, but people just don’t talk like that. We seem to agree that it would be useful for more people to recognize that they nevertheless reason like that.

    I think that Plantinga’s reformed epistemology is a clever way to shift the burden of proof. Notice that all that it claims is that it is rational to believe – not that what is believed is true. But RE basically says that it’s rational to believe Christian creedal specific beliefs as long as there are no defeaters for those beliefs. See what’s happening here? It’s rational to believe one’s Christian beliefs on the basis of a warm fuzzy until somebody proves otherwise. This clearly places the burden of proof on the other side. It would be unfair to characterize this as Plantinga placing the burden of proof on non-believers because he makes it clear that believers themselves have to seek out and defend their beliefs against defeaters. But it does place the burden of proof on the other side.

    Plantinga accomplishes this by fiat. He says that doxastic communities can decide for themselves what belief types they think are basic and then develop a framework for basicality around those. His community believes that Christian beliefs count, and his A/C model provides a framework for them to be warranted, and that’s it. This isn’t to minimize the philosophical task he has accomplished, but it is to say that he and his doxastic community have just decided that their creedal beliefs are rational, and then shown that they could be if true, leaving the burden entirely on the other side. Since they are rational by declaration, de jure objections basically can’t succeed, and the only way to object is by making de facto objections, ie. arguing that the beliefs are false. That’s one amazing way to shift the burden of proof!

    Plantinga would argue that if we can do this with beliefs about the external world, other minds, etc., then it’s ok for him to do with his Christian beliefs.

    I’ll give some more thought to your question about Occam’s Razor.

  3. Travis, you may be interested in what I wrote about Occam's razor earlier.

  4. Yorgo,
    From your response I'm gathering that a de jure objection can only be of the form "you should actually be agnostic on that" when operating in the context of probabilistic arguments. I know it's just semantics, but I'm trying to understand the implications of such a distinction.

    With regard to Ockham's Razor, your parsimony article went in exactly the direction I was thinking. The discussion on information content ties directly into the formalization found in Solomonoff Induction (as I understand it). So when you say that "Plantinga would argue that if we can do this with beliefs about the external world, other minds, etc., then it’s ok for him to do with his Christian beliefs" I take this kind of rationalization to be a demonstration of the problem. Just as you explained your parsimony article when discussing solipsism vs realism, the problem is that those Christian beliefs entail a multiplication of information beyond what is necessary to explain them. The acceptance of the external world and other minds already yields the simplest framework to sufficiently explain our experience - including warm fuzzies - without positing the supernatural. I'm sure the religious minded would protest that naturalism can't account for those subjective experiences, just as they do with morality. Regardless, if we apply parsimony to the equation then it seems to me that acceptance of Plantinga's position should entail acceptance of Berkeley's idealism, where everything exists only in the mind of God. This would seem to have the lowest information content if one insists on including the supernatural in their foundations because it includes everything without the extra act of God creating a separate entity (the universe). As far as I can tell, however, Plantinga rejects anti-realism along with most everybody else. It seems like there is also a parallel argument with respect to dualism - that the dualist should favor panpsychism or idealism over realism for the same reason. Does this make sense?

  5. "From your response I'm gathering that a de jure objection can only be of the form "you should actually be agnostic on that" when operating in the context of probabilistic arguments. I know it's just semantics, but I'm trying to understand the implications of such a distinction."


    I guess the question you may be asking is whether an abductive argument based upon parsimony could represent a defeater for his beliefs. It's an interesting question. I suspect that you may be on to something there. Such an argument, if it succeeded, would permit you to conclude that the existence of God is less likely than his non-existence, which should defeat his theistic beliefs *all else being equal*. I suspect that Plantinga would find ways to argue that all else is not equal, even if the kind of argument you seem to be proposing succeeded.

    I'll think some more on it.

    1. More specifically, I'm proposing that Plantinga's assertion that theistic belief is properly basic is incompatible with his ontological realism regarding the external world when viewed under the scrutiny of parsimony. As I understand it, the claim for properly basic beliefs is that "all else is equal" with respect to those beliefs, so there aren't any mitigating factors that can overrule the probabilistic advantage that pure realism and pure idealism would have due to their lower information content, relative to a dualistic worldview.