We gain information about the outside world through observations. There is also another kind of knowledge, abstract knowledge, which can be attained by logical deduction, such as knowledge of mathematical truths. What always holds is that there must exist some justification for beliefs in order that they may be considered even likely true.
But some things seem impossible to know — some claims are impossible to either prove or to disprove. What, then, should we think about these? Is the only alternative available in all such cases to acknowledge that it’s impossible to say anything about the matter, or that any answer must be based on faith instead of knowledge?
Surprisingly enough, something can still be said in some cases that might appear insoluble this way.
We can form any beliefs whatsoever about the world, really; anything that can be imagined. Beliefs can correspond to the world, at least roughly (for example, I can know that I am sitting on a chair without knowing the exact molecular structure of the chair and myself and so on), but they don’t have to. If I believe that some particular box has a rabbit in it, I am wrong if the box contains nothing but air. Such a simple matter as this is usually easy to settle as well; I can open the box and see whether I see a rabbit in it. If I were to only see a closed box and not so much as try to pick it up to gain a hint about its contents, I of course have no idea about whether there is a rabbit in it, unless I have previously gained information on the matter through some other means, such as a reliable witness who says they saw a rabbit being put in. If all the information I have is that I am seeing a closed (and opaque, immobile, odourless, quiet, large enough to hold a rabbit…) box, I really can’t know whether there is a rabbit inside or not. It takes evidence to ascertain the truth or falsity of beliefs.
The degree to which evidence supports a hypothesis varies along with the likelihood that the evidence gives for the hypothesis being true. If the box were to start jumping up and down as if a rabbit inside was trying to get out, I might perhaps guess that there might be a rabbit inside, but it would not be certain, especially when I probably wouldn’t be able to tell whether it would bounce just like that if there was specifically a rabbit inside. If I opened the box and saw a rabbit, the evidence would bring near certainty — it might still be a kitten disguised as a rabbit or a hologram or something, but such possibilities would be very unlikely.
In some cases, two or more possible beliefs have been brought up, and the choice must be made between them based on evidence. (Note, though, that it isn’t the case that the alternatives that have actually been brought up are the only possible ones. This is a simplification.) A real rabbit or a hologram? Bigfoot footprints or fakes? Which historical person was telling the truth when they both wrote about the same event, or did neither? In the first example, one can try touching the rabbit, and if it feel soft and fuzzy and/or bites, it gets all the more likely that it’s not a hologram, because holograms so interactive as that only appear in science fiction. (Although who knows what someone in Japan has invented by now…) In the second case, well, it could go either way depending on the quality of the footprint and other details, but apparently those found so far have been proven to be fakes. In the case of the historical accounts, on the other hand, there may of course be further evidence that confirms or refutes one of the stories, but it’s also easy to imagine how in such a case the evidence would be so incomplete that there would be equally much for both, and it would be simply impossible to say anything else than that both are equally likely to be true, or they could both be untrue. Here, we see an example of a case where it’s impossible to choose between different possible beliefs based on evidence and there really is no sensible reason to say anything about the matter other than that it could be either way.
By the way, I want to ask one thing at this point: How often, when you see a closed box, you think there might be a rabbit inside? If not every time when you can’t prove that there isn’t, then why not?
Now, we can move on to another type of case: those in which the claim is by definition impossible to prove or disprove. A definitive example is the existence of God. Now, of course many people do claim there is evidence for it. This is not the place to evaluate these claims, and I’m only using the existence of God as an example here, so for the sake of example I’m going to accept the notion by many modern and also past Christian (and probably other) theologians that it is purely a matter of faith. Everything can be explained either naturally or assuming God is behind everything. A believer can see anything in the world as an act of God, a nonbeliever need not see anything as such. Many other things called supernatural are similar; scientific study shows that the reasons people have for believing in them do not in fact necessitate the belief, but that doesn’t stop the people believing in them going on believing anyway, since it could still be true.
Well, there you have it. Unless you say that it can’t be known whether for example God exists, then you have to be taking all on faith; either faith that God exists, or equally baseless faith that God does not exist. Doesn’t this follow from the foregoing?
As I said at the beginning, we can form any beliefs whatsoever about the world. Here’s one possible belief: there is a rabbit skull with the words “Quidquid latine dictum sit, altum videtur” written on it in red ink hidden inside the floor of the room you are in. Suppose I were to assert that this claim is true. I don’t know who’s going to read this and where. But I’m asserting that for each of you who ever read this atop a floor, there is such a skull hidden inside the floor. Do you believe this claim?
Now, it would be possible in principle to disprove this claim. You can dig up your floor to check, though I suspect not too many of you will consider that worth the effort. But you haven’t done it yet, unless one of you is really quick, so you can’t yet know the fact of the matter. And suppose I were to say that the skull is invisible on all wavelengths and disintegrates into nothing as soon as anyone touches it. It does have an odour, but that is the exact same as the odour of its surroundings. So there’s no way to detect it. If you can think of some way still, I can already tell you that won’t work either.
So, this apparently means that if any one you now believes that there is no such skull inside your floor, you’re basing this on faith the same way as someone who believes in God based on faith. Surely you all now think that now that the reliance on faith of such propositions has been proven, the odds are precisely equal that the skull exists or not.
To make the point explicit: since we can come up with an infinity of beliefs, beliefs unsupported by any evidence are so unlikely that they can be rejected outright even though they can’t technically be shown to be wrong in a strict deductive sense. Equally well as God creating the world, it might have been created by five-dimensional supernatural birds building their nest, or as a result of a car race in another universe, or by a law of nature that states that at any moment there must exist at least five universes that contain a copy of Michael Jackson, or by a God that is exactly as we imagine but also has three glowing material hands that float in space and scratch each other. The idea of God may seem the obviously simplest one of all of these, even though it in fact carries an awful lot of extra cultural and anthropocentric baggage, but insofar as it’s possible to explain everything naturally, God is already an unnecessary addition, and as such the same by nature as all of these others. Since these hypotheses cannot be distinguished based on evidence, how could we know which is the correct one?
By the way, what’s the difference between God and the rabbit skull I mentioned? It is that the concept of God is culturally accepted and people are used to it. In addition, it’s probably more satisfying psychologically. But epistemologically there is really no more reason to believe in God (again assuming the proofs aren’t valid after all) than the rabbit skull. There is only psychological motivation to believe. People start out believing, and then are ready to justify it by how it can’t be proven wrong either, whereas in the case of the rabbit skull, they never started to believe, because nothing outside the evidence motivates it, and neither does the evidence itself.
Therefore, if there is no evidence whatsoever of a thing, it is in no way irrational not to believe in it. This is why the agnostic idea that there is no point in believing anything about the transcendent because such things can’t possibly be known by humans is missing something if it does not lead to atheism. If we can know nothing positive of such things, that does mean that we can know people’s guesses about them aren’t be right.
This can be compared to lottery numbers. If you think up an arbitrary number, it is extremely unlikely that it will win the jackpot on the next round. But if you try to come up with one that won’t win, well, that’s virtually certain to succeed. Again, it’s presumably not equally irrational to believe that a particular number is practically certain to lose than it would be to believe that it will win.
Note the difference to the earlier example where it really made sense not to come down on either side. In that one, some evidence pointed to one thing, some to another. Each idea had evidence, they just countered each other. If there is evidence of something, but it’s not strong enough to come to even a likely conclusion, it’s not the same as there being no evidence.
To summarise: If something is extremely likely, there is reason believe it to be true. If something is extremely unlikely, its not being true is extremely likely. Anything of which there is no evidence is extremely unlikely, because the variety of such notions is constrained only by the limitless human imagination, not what there actually is. For this reason, there is reason to disbelieve things for which there is no evidence.
The following well-known ideas can all be derived from this one way or another:
- Occam’s razor: A principle according to which, when trying to explain something, one should settle for the simplest explanation that accounts for the observed facts. (So not just the simplest explanation if it doesn’t explain them as well as some other.) If, for example, the universe can be explained in a particular way based on the laws of nature, the laws of nature will be necessary for explaining it; but if you also add “and God is behind all this,” then you are adding something superfluous and, considering that the laws of nature already explain everything, there is in fact no evidence for God since all the evidence can be explained entirely without it.
- The burden of proof: The burden of proof lies with the one who makes a claim. This person must produce some evidence, which also must not be cut off by Occam’s razor, for the above reasons; and not demand that others prove her wrong and then pretend that if it can’t be done, a reason has appeared from somewhere to believe her claim. There must always be a reason in the first place.
- Russell’s teapot: A famous example about the burden of proof, similar to the rabbit skull example above. This one is about a teapot orbiting the Sun between the orbits of Earth and Mars for no reason, the existence of which it would also be impossible to disprove.
- The falsifiability of scientific theories: This principle is a crucial part of why modern science can be trusted at all. Put in a simplified way, it means that science only accepts theories that contain the possibility of experimentation or observation showing that they are wrong. Experiments are even set up so that a prediction is made up based on the theory and then, as it were, scientists try to prove it wrong. If it turns out to be correct, the theory gains some verification. If it turns out to be incorrect, something’s wrong somewhere, and they need to find out what — and take into account the possibility that it may be the theory itself. If, on the other hand, there are theories (incidentally, I’m using the word “theory” pretty loosely here) that allow you to always make up a new excuse when their predictions go wrong, or that no-one even tries to make predictions based on, they get us nowhere. Because people can develop any beliefs whatsoever, one must always start with the assumption that they may be wrong, and assume otherwise only once the evidence starts coming in. Theories that “explain” all possible occurrences explain nothing; they don’t increase our understanding of matters, they don’t help predict any events, and since they really have no link with the contents of evidence, they’re just the ones that it’s impossible to prove or disprove in principle, and therefore there cannot be any reason to believe in them. For this reason, it’s also no use saying that a typical idea of God explains everything and is therefore spared by Occam’s razor. It’s not any sort of explanation to say about everything: oh, so now God decided to do that for some reason that we can’t understand — or for some reason in hindsight that could not have been reliably predicted.
The same principle also applies in cases where something is possible in the sense that such states of affairs are known to occur. For example, consider the claim that someone is lying. If you have no particular reason to assume that this particular person is lying about this particular thing — she doesn’t usually lie, the claim isn’t less than believable for some reason, and so on — we still know that people lie sometimes, so it is a possibility that it might be worthwhile to note, unlike an idea with no evidence whatsoever. One should still remember, however, that if there is no evidence of lying in this particular case, there’s reason to consider it unlikely until proven otherwise. (Unless one believes that people lie so often that it’s always likely until proven otherwise. I do not believe that.) What I want to say here is that you also shouldn’t jump between degrees of surety from “that’s possible too” to “it may well be true.” Even if something is not supernatural or otherwise extremely unlikely, one should still require evidence and apply, for example, Occam’s razor.