One of the most confusing topics regarding argumentation and rationality is what we call the “burden of proof.” What is it? Who has a burden of proof? I will argue that there are two kinds of burden of proof—(1) a principle of debate and (2) a principle of rationality. These two principles are similar but there are important differences. As a principle of debate, the burden of proof determines who needs to prove their assertions. As a principle of rationality, it determines what beliefs are irrational without further evidence in their favor.
One idea behind the burden of proof is that someone has to prove something or we have no reason to agree with the claim being made. We could then say that person has a “burden of proof.” The person should prove (or argue for) some assertion they have made. If I tell you that fairies exist, then you could argue that I have the “burden of proof” because the existence of fairies is far-fetched or you simply know of no reason to agree with me.
Imagine that I say that fairies exist, and you disagree. I might then say that you have no evidence that fairies don’t exist, so you should believe in them. My reply would seem to irrationally reverse the burden of proof. It seems like I should give you a reason to believe in fairies and that you have no burden to give evidence that they don’t exist.
The burden of proof is related to Carl Sagan’s saying that “extraordinary claims require extraordinary evidence.” If we find out that fairies exist, then our worldview would be significantly changed, so it is seen as an “extraordinary claim.” We aren’t going to believe in fairies unless there is better reason to believe in fairies than there is to continue to maintain our current understanding of the world.
A principle of debate
In a debate, people are required to give us reasons to agree with their controversial claims. It is a fact that those in a debate (or those in the audience who watch a debate) probably think fairies don’t exist, so anyone in a debate who asserts (or requires us to believe) that fairies exist has a burden of proof.
Both sides in a debate start with a burden of proof.
People in a debate disagree about some fact (such as the existence of God). One person in the debate is then required to argue why the fact exists, and the opponent is required to argue why the fact doesn’t exist. The burden of proof is not specifically against only one side of a debate because both sides need to prove that their claim is true.
Arguments require premises—certain beliefs must be agreed-upon or we will not agree with the conclusions. Consider someone who argues that “Hypatia is a woman; all women are mortal; therefore, Hypatia is mortal.” In this case everyone is likely to already agree with the premises, so it’s a good argument in a debate.
However, imagine that someone argues that, “Hypatia is a fairy, all fairies are mortal; therefore, Hypatia is mortal.” In this case no one is likely to agree with the premises, so the debater has the burden of proof to tell us why we should agree with them.
The burden of proof in a debate can shift.
Although both sides start with a burden of proof in a debate, the person who has the burden of proof can shift. When two sides debate and they both argue for their conclusion, we might not know which side’s conclusion is true (and which arguments are unsound). At that point both sides will have a burden of proof to explain why the opposing argument is unjustified or inferior. Any side that fails to meet this burden of proof will continue to have one. For example, one person might argue that abortion should be legal because women have a right to choose and another can argue that abortion should be illegal because the fetus is a person. At this point we might not have any idea which argument is reasonable (if either of them are), and we will not be able to know which conclusion is true as a result.
For this reason, each side will not only argue for their conclusion, but they will object to the arguments given by the opposing side. The opposing side then has a burden to reply to the objections (and they will continue to have a burden of proof until they reply to the objections). Perhaps the pro-life advocate argues that women don’t have a right to choose because human life has value and the fetus is a human life. At that point the pro-choice advocate would have the burden of proof to explain why this argument is unconvincing, and that burden of proof will continue to exist until the explanation is given. (We could say that the pro-life advocate has the burden of proof to prove that abortion should be legal at that point.)
There are some exceptions to the general rules discussed here. Some arguments might be too unreasonable to be worth responding to. If an argument is obviously unreasonable, then it will not shift the burden of proof.
Do all assertions we make in a debate require the burden of proof?
Not all assertions we make in a debate necessarily have a burden of proof because those we debate with (and the audience) will agree with certain points. People already agree that murder is morally wrong, that other people have minds, that there is an external world, that we can legitimately generalize certain data, and so on. These beliefs do not need to be proven in a debate.
If we did have to prove all our premises to be true in a debate, then it would lead to a circular argument or an infinite regress. Imagine that Samantha argues that “Hypatia is a woman; all women are mortal; therefore, she’s a mortal.” I think this argument is perfectly good, but Johnny could then object, “How do you know all women are mortal?” Hypatia could then respond, “All women we know about died in the past before reaching two-hundred years of age.” Johnny could then object, “But maybe immortal women keep it a secret.”
This line of questioning can go on and on forever. For that reason people in a debate should agree not to question such plausible premises and they should provide as much evidence for their own position as they demand the opposing side to provide.
What about the supernatural?
Some people argue that assertions of the supernatural are “extraordinary” and require extraordinary evidence for that reason. This is only true within a debate when people don’t already agree with the supernatural assertions and demand those assertions to be justified.
A principle of rationality
Some people argue that it really is irrational to believe in fairies and rational to disbelieve in them because of the “burden of proof” (in or out of a debate context). This concerns how much we should be justifying our beliefs on our own, and what beliefs we are rationally required to have. Many people want to say that we should believe that fairies don’t exist and that fairy-believers have a burden of proof to justify their belief. If there’s no reason to believe in fairies, then it is supposedly irrational to believe in them.
The rational principle of the burden of proof is that some beliefs are initially irrational or less plausible than the alternatives prior to debate or evidence. Some beliefs should be rejected unless we have sufficient reason to agree with them. We would have a “burden of proof” for any belief that’s irrational without a needed argument.
The rational principle of the “burden of proof” is not entirely separable from the “burden of proof” in debate because people who debate also care about rational standards and often conclude that people should agree with them, or that people are irrational for disagreeing with their conclusions.
When are beliefs rational or irrational?
Rational beliefs (i.e. rationally permissible beliefs) are generally said to be beliefs we can have without being irrational. When is a belief rational? This is a controversial subject, but a relatively permissive view would say that beliefs are rational as long as we have absolutely no reason to reject them. This belief seems to be shared with Karl Popper, who argued that scientists (and people generally) invent hypotheses that are consistent with their understanding of the world and they believe those hypotheses are true until we have a sufficient reason to reject them. (Or, at least we know they should reject their hypotheses once we have a sufficient reason to reject them.)
The relatively permissive view does not say that we need to prove all our beliefs to be true, and that view seems to be implausible. For example, it would seem to lead to circularity or an infinite regress, just like when someone who questions us in a debate over and over forever. See “Beliefs are Innocent Until Proven Guilty” for more information.
A belief is irrational when we have a sufficient reason to reject it. We have a sufficient reason to believe that “1+1=3” is false, so it is irrational.
If a belief has a “burden of proof” (as a principle of rationality), then the belief is irrational unless the burden of proof can be met. (There should be a good argument as to why the belief is not irrational.) No belief has a burden of proof (or is irrational) in absolute isolation. However, beliefs can have a burden of proof (or be irrational) in consideration of our background knowledge (our current understanding of the world). At one point the belief that the Earth revolves around the Sun could have seemed irrational and perhaps we would need evidence in support of such a belief in order to justify it. That is no longer the case—now it would be shocking to find out that the Earth doesn’t revolve around the Sun.
Should we believe in fairies?
No, we shouldn’t believe in fairies. It is rational to believe fairies don’t exist, and it is irrational to believe fairies do exist.
Why shouldn’t we believe in fairies?
There are at least two main factors that determine the fact that we shouldn’t believe in fairies (and that those who believe in fairies have a “burden of proof”): (1) Occam’s razor and (2) our background knowledge. Beliefs come with a burden of proof whenever the belief requires us to believe something that conflicts with our current understanding of the world for no good reason. I will describe how Occam’s razor and our background knowledge determine that those who believe in fairies have a burden of proof.
Occam’s razor – Occam’s razor is a principle of justification that helps us determine if a belief is sufficiently rational or not. There are at least two important formulations of Occam’s razor. One, “don’t multiply entities beyond necessity.” (i.e. Don’t posit the existence of entities unless it’s necessary.) Two, “if two explanations are equally good in every other respect, then the simplest explanation is better than the other one.” Occam’s razor should encourage us to question whether the existence of fairies would be multiplying entities beyond necessity or if they are used as an explanation when simpler explanations are available.
People generally believe in fairies based on various experiences—stories they’ve heard and objects that seem to move around on their own. The problem with these experiences is that there are simpler explanations as to why such stories exist and why objects seem to move on their own other than the existence of fairies. For example, people could have just made the stories up, we might have faulty memories concerning where we leave objects, and humans might have a tendency to anthopomorphize (imagine that people-like entities exist where they don’t exist).
If everything we know about the world would remain unchanged without assuming fairies exist, then we have no reason to believe in them, and we have at least one reason to reject them (i.e. Occam’s razor). That should be the case for everyone. We should know how to fully understand the world without positing the existence of fairies.
Our background knowledge – Occam’s razor doesn’t help us in isolation. We need to know whether or not we are multiplying entities beyond necessity (or whether or not the simplest explanation is just as good as the alternative explanations in other respects). Consider fairies. They significantly conflict with our background knowledge (because, for example, (a) we would have expected to know about them by now, (b) they require us to believe in magic, and (c) we thought people just made them up). And if they don’t exist, then nothing we know about the world is changed. Therefore, fairies seem entirely extraneous and it doesn’t seem necessary to believe in them.
On the other hand the existence of pain is perfectly compatible and perhaps necessary based on our experiences, and the positing the existence of a neighbor being home seems reasonable when a light is turned on next door. In these cases we are positing the existence of substantial entities, but they seem compatible and perhaps even necessary for us in order to adequately explain and/or understand the world around us.
What about extraordinary claims?
There is a sort of “extraordinary claim” that is relevant to the burden of proof as a rational principle. In this case a claim is “extraordinary” when it significantly conflicts with our background knowledge. The claim that “extraordinary claims require extraordinary evidence” is relevant because the more a belief conflicts with our background knowledge, the more evidence will be needed before we can rationally accept the belief. We often say these claims are “ambitious” or “far-fetched.”
What about supernatural entities?
There is a question concerning whether or not anything supernatural exists whatsoever. Many people claim that every supernatural claim is “extraordinary” and conflicts with our worldview. It is true that supernatural claims do conflict with certain people’s worldview, but not necessarily everyone’s. For example, some people might think that psychological phenomena is supernatural, but there’s nothing extraordinary about psychological phenomena.
Even so, people should be consistent about how they treat supernatural entities. People reject supernatural entities in general. There is no religion I know about that agrees that all supernatural entities exist. Do fairies, vampires, werewolves, zombies, ghosts, demons, demigods, angels, or gods exist? Does Thor, Zeus, or Isis exist? Given any particular supernatural entity, it seems reasonable to expect people to be skeptical. Anyone who wants to reject all supernatural entities except a few of their favorites could be inconsistent in how they apply the burden of proof or principles of rationality in general.
Although people who discuss burdens of proof seem to be talking about the burden of proof as it exists within debate, I think what we are more interested in are burdens of proof as they exist in rationality. We want to know what we should believe, what beliefs are rationally permissible, what beliefs are rationally required, and what beliefs are irrational.