Against the supernatural as a profound idea--Part 2

Definition Jumping

People who claim that the supernatural exists claim it can be defined meaningfully. If it means anything at all, this can only be interpreted as a claim that the boundary that I just discussed can be meaningfully described. People who do this typically indulge in what I will call definition jumping. They claim to define the boundary in one way that is hopelessly vague, to the point of being meaningless, and when challenged they merely offer a different, similarly vague, definition of the boundary. If challenged again they may offer a further, vague definition.

For example, someone might claim that supernatural things are "non-physical," which amounts to an attempt to say that the boundary has physical things on one side and non-physical things on the other. If you ask what this means you do not get a proper definition of "non-physical." Instead you are told that non-physical things are "immaterial," which amounts to a claim that the boundary has "material" things on one side and "immaterial" things on the other. You may then be told that this means that natural things are "scientific" and supernatural things are "beyond science." None of these definitions say anything, because they do not describe the boundary completely and attempts to define them merely result in a new definition that does not describe the boundary either. Just one of them would be enough if it described the boundary.

The Supernatural as a Semantic Island

Suppose we replaced the word "dog" with a nonsense word like "znud." We would still be able to explain what a "znud" is and make people realize that it means the same as "dog." We could tell them that znuds are mammals, that a greyhound is a type of znud, that znuds are related to wolves, that some znuds are trained for police work, and so on. Even if we replaced a number of other words that we might use to explain what "znud" means we could still explain "znud" because it is part of our meaningful language and has a rich set of connections with other concepts that we can exploit. Unless a huge number of words were turned into nonsense words we would always be able to explain "znud."

People may think that terms such as "supernatural," "natural," "physical" and "non-physical" mean something, but we could see how much they mean by replacing them with nonsense words too. Suppose a relatively small number of words used to describe the boundary between the "natural" and the "supernatural" were replaced with nonsense words. For example, "natural" could be become "vuf," "supernatural" could become "super-vuf," "physical" could become "zlup," "non-physical" could become "non-zlup," "material" could become "znurgibop," "immaterial" could become "non-znurgibop" and so on. I suggest that by replacing words in this way, it would be practically impossible to explain what the new words mean using the remaining language. Someone trying to explain "super-vuf" might try telling people that it is "non-zlup," for example, but that would hardly help. Because all these supernatural definitions are only defined with reference to each other there would be nothing to use to define the nonsense words apart from more nonsense words. You may argue that I could make anything impossible to explain by turning enough words in human language into nonsense words -- if nothing else, I could change every word in the language into a nonsense word and prohibit any communication -- but what is important here is that I am suggesting that you could make the concept of the "supernatural" impossible to communicate in this way by changing a relatively small group of words. This should be seen as a serious problem, because if it is correct it means that the words relating to the "supernatural" form a "semantic island," cut off from all the meanings in the rest of human language. This suggests that the words are meaningless.

If people want to be pedantic, there is one way in which someone might try to define the supernatural if we replaced all the words associated with it by nonsense words like this. They might reference someone's knowledge about people and institutions that believe, or promote belief, in the supernatural, or they might reference the previous making of claims for the supernatural, or people known to have discussed it. For example, they could say:

"Paul Almond placed an article saying that vuf is an incoherent idea on his website in November 2008."

"Religions tend to promote belief in vuf, but vuf is not the same as God. Vuf means an entirely different sort of thing."

To make it impossible to express these statements I would have to replace even my own name and words like "religion" with nonsense words. Does this invalidate what I just said? I think that the idea still stands. These sorts of explanations would be a kind of cheating because they would not really explain the "supernatural" to you: they would merely associate the nonsense word "vuf" with some previous memory you had of someone discussing the word "supernatural." You would not be finding anything out. This is in contrast with the example of replacing "dog" with "znud." The concept of "dog" has a rich relationship with lots of things that allows us to you use many ways of explaining what "znud" means without having to depend on descriptions of people who have various beliefs about dogs, or people who have written books about dogs, etc. The "supernatural" and related terms may not be on a perfect "semantic island," but the only bridge between them and the mainland of the rest of human language is a very tenuous one of references to people's prior memories of beliefs and statements about it, which is worthless for meaningfully linking the concept to the rest of our language.

Some readers, of course, will insist that terminology associated with the "supernatural" is meaningful. They only think this because they have become so used to labelling things as "supernatural" and "natural" that they have become unable to see they are using a vague, incoherent boundary. To show this I will now discuss some of these ways of trying to define the boundary in more detail.

Next Page >