The other day I was picnicking in Central Park with my 13-year old daughter, when the issue of burden of proof came up. Yeah, I know, what sort of a geeky daughter and overly-intellectual father are we talking about? Nevertheless, Caley (my daughter) knows very well that I am both a skeptic and an atheist, and we occasionally have good humored conversations about these issues.
Specifically, the discussion was about how do I know that there are no ghosts (or other paranormal phenomena). I explained that I don’t know in the sense of certain knowledge, but that there are two major reasons I don’t believe in ghosts: a) Their existence would be at odds with much of what we know from science; and b) I haven’t seen any positive evidence (that has withstood scrutiny) that ghosts exists.
Caley’s response focused on (b) above: “yes, but you also don’t have evidence that ghosts don’t exist, so there.” This is, of course, a classic example of shifting the burden of proof, which is a well known logical fallacy usually described under the heading of “appeal to ignorance.” The idea, as I tried to explain to my daughter, is that the statement “I don’t believe in ghosts because there is no evidence” and the statement “I believe in ghosts because there is no evidence that they don’t exist” are not epistemologically equivalent, so that it is much more reasonable not to believe in ghosts (given the lack of positive evidence).
Although Caley listened to what I was saying, and she is a smart girl, it was clear that I wasn’t getting through. It just seemed logical to her, that it is okay to believe in ghosts if there is no proof that they don’t exist. I then switched tactics and used an example that she could follow more easily. I reminded her of what happens in courts of law (of non fascist countries): the burden of proof is always on the part making a positive claim, not on the one making a negative one. Most especially, one is always presumed innocent unless proven guilty (beyond reasonable doubt, another concept that also nicely fits with skepticism). It would be grossly unfair if we went around presuming people to be guilty of crimes with no other “evidence” than the fact that they can’t prove that they didn’t do it.
Caley paused, clearly understanding the example. But just when I thought I scored one for critical thinking, she went back to her original position: “well, you are making a claim too, that ghosts don’t exist, so don’t you have the burden of proof either?” No, I don’t , but by then it was really time to slice the bread and make our salami and provolone sandwiches, so that was the end of the conversation.
However, I have in fact had similar discussions over and over again — with adults — and have experienced the very same difficulty getting through regarding this concept . Why is it that few people seem to have problems with the burden of proof when it comes to the innocence or guilt of a murder suspect, but then cannot apply the same exact logic to more esoteric issues, such as the existence of ghosts, gods, and the like?
I suspect part of the answer is that in the latter cases most people are attracted to some intuitive (and wrong) notion of epistemic fairness: well, you are making one claim, the other guy is making another claim, the two of you are therefore on equal footing. This seems also to be the underlying philosophy — if there is any — behind the infamous practice by much of the media to “present the two sides,” regardless of how inane one of the two sides actually is. (Incidentally, the latter case is an example of an additional logical fallacy, the false dilemma: why, exactly, would we expect there to be two sides, as opposed to a number of possible, more nuanced, alternatives? Remember “you are either with us or with the terrorists”?)
But if this explanation is correct, then why don’t people also apply the same logic to the legal system? I think that’s because it is our very concept of fairness that shifts in that case: apparently, what is epistemologically fair for impersonal situations is suddenly unfair when people are involved. This may be similar to the case of trolley dilemmas-type thought experiments, which show that people shift from a utilitarian to a deontological ethical system depending on whether they have to make relatively impersonal or very personal decisions about somebody else’s fate (apparently, it’s okay to throw a switch to save five people while killing an innocent bystander, but not okay if you have to physically throw the innocent person in front of the trolley).
The job of the skeptical critical thinker is to convince people that these seemingly different situations are logically equivalent, and that it is therefore not rational to believe in ghosts without evidence at the same time that one wouldn’t dream of convicting a person of a crime just on the basis that she cannot prove her innocence. But as is often the case, human psychology gets in the way of rational thinking.
I've always thought these examples highlight the idea of why burden of proof should fall on the claimant rather than the skeptic.
The problem in justice is different from the epistemic problem, because of the implication.ReplyDelete
You are free to think that someone is guilty, but you won't put him in jail if you don't have any evidence. But you can still think he is guilty... So belief and justice are not really the same because they don't have the same implication.
I would say that the main argument for the burden of proof is that if you do not accept it, then you will have to believe in a lot of things - just as Matt said, the invisible pink licorn, and the teapot, but also thousands other things - there is really no limit. That's why this position is absurd, wheras believing only in things you can find evidences for will never be an absurd position , but a rational one.
The use of a switch is not the main difference between the two trolley scenarios. The second scenario can be modified to use a switch to open a trap door beneath the fat guy, and people still won't find it acceptable.ReplyDelete
I think that "innocent until proven guilty" in the legal context is justified on ethical rather than epistemic grounds: we would rather let several guilty people go free rather than imprison one innocent man. This is a fact about the relative costs/benefits of imprisoning the innocent versus freeing the guilty, not a fact about our belief in innocence or guilt.ReplyDelete
If 51% of the population had participated in some crime, but it was not known which 51%, then we would still want to presume innocence in the absence of evidence, despite realising that any one person is more likely than not to have participated. There are cases where identical twins have escaped punishment on such grounds.
In contrast, the reason for not believing in ghosts is epistemic: a claim so specific as "there is a sentient but invisible being responsible for the strange noises in my house" is just apriori extremely unlikely. And there it remains unless evidence arises to support it.
I think this is what you were getting at re positive/negative claims, but as your daughter showed this framing can be misleading since it is not always clear which side should be considered positive or negative (add a double-negation to the "positive" claim and the two claims would appear to swap places).
Incidentally, I would count our scientific knowledge of physics, life, and intelligence as direct evidence against ghosts.
Courts of law present both sides, regardless of how inane one of the two sides actually is.ReplyDelete
The most important court in the nation allows, in fact supports, equivalency between accusation and truth.ReplyDelete
The amalgamation of pop-media and public opinion.
So ignorant, so ruthless.
Where would lead stories be without the erroneously condemned having to prove the non-existence of guilt?
Actually, I would contend that your analogy with the law court setting can be turned on its head.ReplyDelete
Generally speaking, people shift the burden of proof to whichever party is trying to take something valuable away. In the case of the law court, taking away someone's life or liberty is a very serious matter. The burden of proof, therefore, should be on the one claiming that the defendant is guilty, and therefore deserves to have life or liberty taken away.
Likewise, if the belief in the supernatural, perhaps God or ghosts or whatever, is a perceived value in society, then the burden of proof shifts to the one claiming that the supernatural doesn't exist. It might seem surprising that an intangible thing like belief in the supernatural could be perceive as a cultural good in the same way that human life is, yet that is precisely the case. Therefore, it is not surprising that the burden of proof is often shifted to those who are attempting to strip away that belief.
Maybe you can just leave "burden of proof" out of the discussion and explain why it's unreasonable to believe in unobservable entities unless observations seem to require you to posit such an entity. Invisible Pink Unicorns and analogues are perfect examples of why you should go for sparse ontologies.ReplyDelete
And another possibly educational strategy: what would evidence that ghosts don't exist look like?ReplyDelete
Jameson: "It might seem surprising that an intangible thing like belief in the supernatural could be perceived as a cultural good..."ReplyDelete
It's not at all surprising. (Indeed, it's quite evident.) But is it rationally defensible?
After all, one is free (at least in some countries) to espouse any belief that one chooses, however rationally indefensible or fallacious it may be. If enough people share a fallacious belief, is that a sound argument in favor of giving it equal treatment with those beliefs that rest on firmer ground?
Politically, perhaps. And inasmuch as this is the case in real-world law courts, then I suspect that Massimo would agree: so much the worse for the courts (and for justice, in general).
With apologies for my lack of sophistication, the example I use in this situation is the existence of Santa Claus, which most American children have needed to consider. If another child believes in Santa, and Caley can't prove his non-existence, does that mean that Santa might be real? If Caley feels obliged to state that Santa might be real, it makes for an interesting discussion.ReplyDelete
On a different note, I have always enjoyed reading the Yuletide reprinting of the classic "Yes, Virginia, there is a Santa Claus" and substituting the concept of God for Santa.
Good stuff pugowner. Of course Santa exists for those who believe in Santa, and have no other reason to believe otherwise. It is more about existence as a subjective thing, and less about what is real'. The child, having approached age 6 or 7, whatever, has not only read and sung about Santa, but seen a living apparition in the mall and onstage at school. As adulthood is reached, Santa's existence for the believer has been mitigated a tad, but not deleted, all assertions of "they lied to me" notwithstanding.ReplyDelete
I think AlexF nails the issue. Belief acquisition and revision is about two things -- the evidence in favor of particular beliefs and the costs of being wrong. To paraphrase Popper, it is one of the advantages of rational thought that it permits us to separate these out, so we can use the best evidence re: what to believe, but act on the basis of cost-benefit analysis.ReplyDelete
While the Scottish verdict of "not proven" emerges from an odd historical quirk, rather than a careful reflection on policy, it is, I think, conceptually quite useful. One could imagine a system in which the three verdicts had different standards of evidence and different consequences. So for example one might make it so that "guilty" should only be returned if the state proves its case "beyond a reasonable doubt," BUT not guilty would require the jury to believe that the "preponderance of evidence" (50+%) was that the person was in fact *not* guilty, and "not proven" would cover everything else (the area between 'more likely than not to be guilty' and proven "beyond a reasonable doubt"). In terms of consequences, you could change them, as well -- one ideas would be for "guilty" and "not guilty" to basically the same, but for "not proven" to permit the state to bring another trial on the same charge if they could meet some reasonable standard of new evidence (e.g. finding new evidence of guilt that the state could not have reasonably found or presented at the first trial).
So -- the confusion with ghosts, etc., seems perhaps to be people's mistaken views about the costs / benefits of beliefs versus the reasonableness of them. As several commentators (and Massimo) have noted, given all we know about how the world works, ghosts are exceedingly unlikely beasts (even less likely than pink non-magical unicorns). So our "prior" belief in them should be quite low, and the "evidence" presented in favor of them is quite *weak* so there is no good reason to revise to our beliefs to accept them.
If, though, you think that a belief in ghosts is not just "harmless" but for example "comforting," then if you are inclined to confuse the reasons for accepting a belief with the reasons for acting on it (when action matters) you might be tempted to believe in ghosts despite the bad evidence. This is especially true if it is belief itself that does the work, rather than an action taken on the basis of the belief.
So for example, I might believe that the preponderance of evidence is that chemical x, used let's imagine in some but not all available utensils, is harmless, but I might reason that the cost to me of avoiding it is very low, and the potential benefit reasonably high, so I should avoid it, despite thinking that it is *probably* harmless (but where there is some positive reason for thinking it might not be). The problem is that I don't see how one can separate out beliefs from actions where the beliefs themselves are doing the work. But of course, once one realizes that one is adopting a belief not because one has good evidence for it, but because one thinks that believing it will itself bring benefits, one probably has to give up the process on pains of a kind of irrationality.
(An interesting problem from decision theory is whether one can commit to plans that it would be irrational to actually carry out, if so committing is overall rational; the problem of adopting beliefs one has reason to think are false, in order to secure important gains, is probably a sub-set of this problem...)
I think it's really cool that you're willing to get into discussions like this with your daughter, and I'm glad you posted about it.
I'd like your take on the following question though. You said that the burden of proof lies on the person making a positive claim. When you speak about making a "positive claim," what do you mean? I'm guessing you mean a positive existential claim; but if so, then my intuition doesn't always assign burden of proof that way.
Take the following example:
There does not exist a single human on earth who worships an orange tree as a deity.
and its negation:
There exists at least one human on earth who worships an orange tree as a deity.
My intuition is that both of these claims are subject to the burden of proof. On one hand, there must not be very many people who worship an orange tree as a deity, or I probably would have heard about it. On the other hand, given that there are 6.6 billion people on earth, it seems fairly plausible that one or two of them might worship an orange tree as a deity. I'm not really inclined to give credence to either claim without some kind of proof.
What do you think? Is my intuition wrong? Or am I misunderstanding your point about positive claims?
Scott, if I may cut in on Massimo...ReplyDelete
Your example reminds me of Carl Sagan's credo: "Extraordinary claims demand extraordinary evidence." Of course, different people may define different phenomena as "extraordinary." But I find neither of your propositions to be particularly extraordinary, and both to be about equally plausible, such that neither begs for evidence any more than the other.
That said, the two propositions are unequal in methodological terms, or should you be pressed to fact-check them.
For the latter, your work is done as soon you find a single example, whereas in the former, you have to either survey the entire world population, or else settle for a statistical inference, which is weaker.
I'm still not clear on how to determine which statement is considered "positive." The Wikipedia article on the burden of proof gives some interesting pointers but I don't see what is meant by "ontologically positive" (unless you try to argue the ontological parsimony route). The same article has an interesting example dealing with fairies that shows the burden of proof is much higher for positions positive for the existence of fairies than any position negative for it. The article also says that something is ontologically positive if it requires the addition of something to the existing body of knowledge. But if someone believes in fairies, doesn't the position that they don't exist (potentially) add something to their body of knowledge?ReplyDelete
I like to apply the following test: "If I'm willing to believe THAT, what WOULDN'T I believe?" By analyzing propositions in this way it becomes apparent that, roughly speaking, propositions come in classes (a continuum, really) based on the quantity and quality of the supporting evidence. So a fair policy of belief would require that were I to believe "God exists," I would also have to believe "the Flying Spaghetti Monster exists," "Werewolves exists," "Witches exist," etc. To do otherwise would make me an intellectual hypocrite. It also shows that simply by applying a fair standard of evidence, vast numbers of propositions are kept at bay by virtue of their mutual incompatibility. The result is that one cultivates an "inner sanctum" of maximum certainty (never absolute) that only admits propositions that are supported by the most empirically compelling evidence and which are in conceptual harmony with their propositional peers.ReplyDelete
I consider the above process to be far more relevant to personal belief than notions of "burden of proof." In fact, though the thought only just hit me (and isn't developed), I will go so far as to say that the whole notion of "burden of proof" really doesn't apply outside of an institutional setting, such as a courtroom, where proceduralized debate leads to some sort of verdict.
Yes I was thinking of that phrase as I wrote and wondering if it captures my intuition more effectively than the "positive claim" rule.
It strikes me that there might be a probabilistic way of thinking about that credo: the claim that requires proof is the claim that is more improbable given prior knowledge.
The challenge, of course, would then be to calculate the relevant probabilities the right way.
That does seem to jibe nicely with the way burden of proof is determined in legal contexts (setting aside the additional complication of consequences that others have pointed out). In many (most? all?) cases, without further information, it's unlikely that a randomly chosen individual is guilty of a given crime, so such an accusation won't stick without proof.
I think many people have problems understanding the differences between methodological naturalism and metaphysical naturalism with regard to Russell and rational thought. Many materialists seem to dogmatically cling to Russell and justification theory under epistemology beyond its rational limits within the realm of scientific claims and empiricism.
Like it or not, there are many ontological and Platonic foundations of the sciences. Take views of mathematical realism for example and the existence of abstract entities distinct from "physical nature".
We cannot prove there is an actual infinity. Yet we assert in the sciences that infinity exists and is limitless, and we use it as a foundation for many of the sciences. However, this presents a paradox because the statement contradicts itself. Specifically, it imposes the limit on itself that there are no limits to infinity as a entity. Because of this irrational self-contradictory and self-referential statement we can interpret infinities and paradoxes any way we choose and none can prove us wrong. We can accept them at face value as apparently irrational and inexplicable, deny they are irrational, or simply ignore them.
There are a number of metaphysical interpretations from this.
Much of this divisiveness revolves around difficulties with coherentism and views of "reality" (all that exists whether it is observable or comprehensible) vs. "nature", and the fact that there is no obvious way in which a coherent system relates to anything that might exist outside of it.
With regard to burden of proof for metaphysical arguments beyond the limits of positivism does having a belief always require burden of proof if the counter-claim that the belief is unjustified is equally subject to similar burden of proof?
example: "God created infinity, and man, unable to understand infinity, created finite sets."
I've seen Professor Pigliucci gallantly defend philosophy against those who insist it is dead and worthless. Yet they don't understand why many of us in the sciences fear that form of dogma.
I shouldn't wish people dogmatically to believe any philosophy, not even mine.
Existence is relatively easy to prove with a single example, but hard to falsify. A single captured ghost would prove that ghosts exist.ReplyDelete
Universal laws are relatively easy to falsify with a single contradiction, but hard to prove. A single perpetual motion machine would falsify the laws of thermodynamics.
“well, you are making a claim too, that ghosts don’t exist, so don’t you have the burden of proof either?”ReplyDelete
You are not making the claim that ghosts don't exist. You are making the claim that there is no positive evidence that ghosts do exist. In such a case the most logical and parsimonious conclusion is to believe that ghosts do not exist. Until presented with good positive evidence for the existence of ghosts there is no reason to believe in ghosts.
Great post and discussion, but..ReplyDelete
Would it be innapporpriate to add the comment that I really dislike the new "wall paper" background of the site? One, its very distracting. And two, who would ever have such a large library with no books having an author or title on the spine? It looks silly.
I suggest keeping it all very simple. Maybe even try a white background.
yeah, I'm tweaking the layout by making use of Blogger's new capabilities. We are also having internal discussions about the best format. Right now, the "books" are gone...
Laneman, totally agree. I think this is a helpful angle to take. I am not claiming that ghosts don't exist, I am claiming there is no intelligent reason to believe in ghosts.ReplyDelete
Re: trolley dilemmas. Our moral notions are inseparable from our intuitions about norms of social cooperation. It's very different to sanction the decision making authority necessary to throw someone in front of a trolley than it is to sanction the decision making authority necessary to throw the switch.
Max: I take it you think existence is objective, in other words something can be said to either exist or not exist, rather than existence depending on the observer.ReplyDelete
Can you share an example of something that you think exists?
When confronted with a valid belief statement, that is one that can be stipulated as true for the believer, existence issues for the referent of the belief are simply data points on the usefulness and possibly the reality of the referent which may or may not influence the believer.ReplyDelete
In any event both parties have the same burden of proof. If you claim non-existence for the referent that is an affirmative claim and must be affirmatively supported.
A far more useful "attack" on belief claims is on the usefulness issue. That is, what has the referent actually done for you. That is the reason I am an atheist. Stipulating the existence of God either as a "real" mythical being, or even as an believed to be supernatural omnipotent alpha humanoid, I have yet to find one that is even useful, let alone worthy of any of my time or money.
Didier, I'm not convinced that a defense of philosophy as alive and full of worth is dogmatic. I'm familiar with a specific reference here, but it doesn't seem like your quote applies.ReplyDelete
I agree that bearing the burden of proof for non-belief in the face of zero positive evidence is functionally impossible. We cannot believe in all possible beliefs for which we do not have direct evidence against, and for which there is no positive evidence. We could sit around all day every day conjuring new things to believe in, and still fail to believe in all possible beliefs.
Interesting post, and thread. Unfortunately, I bet you'd be surprised by how often this presumption is simply discarded, and the state simply claims publicly, time after time, that the parties are guilty, before the trial has taken place. In this case, they attempt to shift the burden of proof back on the suspects.
I think the problem can usefully be illustrated via Kant's diagnosis of the failure of the ontological argument. The only difference between the real, and the imaginary dollar bill is that in the case of the real one, perception precedes conception. With claims about, e.g. transcendent entities and many other things that are not directly observable, many people simply have no grounds upon which to judge questions of evidence--for the simple reason that there's no obvious appeal to perception to be made. The only "evidence" that is accessible is the group of claims made by parties to the dispute, and in such cases, the natural (and in many cases highly advantageous), though epistemically naive, assumption is that the frequency of a belief is a guide to its truth. Since (by far) more people accept some (undefined) deity, this is conceived as the default assumption, and the burden is shifted to those attempting to *disprove* it.
I think the responsibility for the burden of proof is an interesting debate, but fundamentally, as a sceptic, everyone has to justify their position, whether they're making a positive claim or a negative one.ReplyDelete
If I say I believe in God, I should be prepared to provide my basis for holding that belief for the scrutiny of others I may wish to convince. If I say I don't believe in God, I should be prepared to give my reasons for doing so, and defend them against those who don't agree. (even if the reason for holding a position is that there is no evidence to give cause to hold an alternative position, at least that gives a starting point for discussion or a basis to evaluate whether the sceptic is being reasonable or just a denier)
Simply declaring that 'the burdon of proof' lies with the other person is a very weak argument, and usually used mainly by those who are not confident enough to stand up for their own position.
In global warming discourse, the 'sceptics' will often say "there is no evidence the world is warming" or some variation. This gives the rationalists the opportunity to introduce evidence, whereby the true sceptic is immediately seperated from the denialist by how they react to the introduction of evidence.
Deniers as opposed to skeptics tend to constantly demand more and more evidence and demand that the burdon of proof is on those who make the claim and that they're merely sceptics who are unconvinced by the evidence, but there reaches a stage when evidence is presented where the burdon shifts to the sceptic to provide a valid critism of the evidence provided and this usually involves making their own claims that can be tested to see if it is valid criticism.
@Austin - re "We cannot believe in all possible beliefs for which we do not have direct evidence against...." we can, we can.ReplyDelete
We can do using the subjective nature of 'evidence'.. What constitutes evidence in scientific circles does not constitute evidence in the minds of the non-scientific.
And so the burden of proof can be shifted not to those who claim the gods or other-worldly entities exist, but to those who claim that we and perceived things in our lives do indeed exist, and are not extensions of our imaginations, that for whatever reason can be counted on to appear and behave the same way every time.
@DaveS: "We cannot believe in all possible beliefs for which we do not have direct evidence against, and for which there is no positive evidence."ReplyDelete
Let me change that "and" to an "or." Can there be a state of suspended belief? I cannot believe in something I have no evidence for, I cannot disbelieve in something I have no evidence against.
I have argued this with Mintman re: a previous post. He replied that the guy who claimed that alien's built Egypt should be "laughed out of the room." I agreed. I think I do here, too.
Belief holders must have some epistemological standard for evidentiary requirements for belief, otherwise we believe everything, i.e. "Alpha Centauri is populated by pink unicorns" and at the same time "Alpha Centauri is not populated by pink unicorns." This is not logical. I would again assert that it is impossible to believe all things for which there is no evidence, therefore evidence is required for belief, or disbelief. This is stipulated to. Mintman agrees and says this is uninteresting. We have an coherent, evidence based account of Egyptian history that does not make room for aliens. We have a coherent, evidence based account of a universe that does not make room for god.
>Can you share an example of something that you think exists?
I can try: Myself, Earth, moon, shoelaces...
It sounds arrogant when skeptics say, "There is NO convincing evidence of ghosts, God, alien visitations, etc." Nobody could've reviewed ALL the evidence. It's more honest to say, "Of all the evidence I've seen, none was convincing."ReplyDelete
But I agree with Akrasia that merely dismissing evidence without offering alternatives sounds weak. The believer's reaction is, "Well if it wasn't a ghost, then what was it?"
"I don't know" is a valid answer, but it won't turn any believers into skeptics.
@Max - Thanks. Apart from your 1st example, all other answers are examples of things which exist for you, which make sense to you and so forth. They are also things one would call 'real' or 'tangible', as opposed to intangibles like friendship, boredom, tension, etc.... The reason I think there is no difference between the 'real' list and the intangibles is that both are things you perceive, nothing more nothing less. The perception is fueled by your 5 senses, public opinion, and your proclivity for siding with any given stream of opinion, and perceptions harden into beliefs over time.ReplyDelete
Take those shoelaces, you say they exist. I say they exist with lots of conditionals that bound the terms of their so-called existence. They exist for you and me, because I too know what they are, I even have a few pair. Speed up a shoelace's clock, and each shoelace is popping in and out of existence faster than can be measured, making it irrelevant whether it is real or not. You may say the shoelaces obey the laws of science, but these laws, too, are bounded by time and other conditions allowing them to appear to you.
Things do not exist in vacuums, that's why you can't say something exists, and then walk away.
I'm with your daughter there. But we could argue endlessly over that. As an aside, India is a democratic country, ruled by a democratically created Constitution that lays down the laws of our land. I am not sure if it is the same elsewhere, but did you know that, in India, the burden of proof, in the particular case of crimes against women, lies on the person being accused? Interesting fact, just FYI.ReplyDelete
The reason (reasonable people believe that) ghosts don't exist is that a model of the universe without ghosts is more parsimonious than one with ghosts. "x has the burden of proof" is just another way of saying "x's model is the less parsimonious one given the data we currently have, so x needs to present new data".ReplyDelete