One of my main criticisms of methodological naturalism is that the word “naturalism” seems to not have any real content. It merely means “whatever the author wants to include in science” and non-naturalism means “whatever the author doesn’t want to include in science”. This paper by Halvorson is a case in point.
One of the more frustrating parts of talking about methodological naturalism is coming up with a suitable definition of “naturalism”. I discuss this in-depth in this paper. Essentially, most definitions of naturalism are either extremely vague or extremely self-serving. For instance, “testability” is one definition that is often put forward. But, if supernaturalism is true, there is no reason that supernaturalism couldn’t at least in theory be testable. Observability is another, but that is just another way for saying testable.
Hans Halvorson is a philosopher of science who has (accidentally, I believe) shown just how problematic the word “naturalism” is. He wrote a chapter for the Blackwell Companion to Naturalism called “Why Methodological Naturalism?” (free version available here). In the introduction to his paper, he says,
One striking feature of these theories is that they are naturalistic: they don’t mention gods, demons, or any other supernatural beings.
So, his initial mentioning of naturalism is that it excludes all supernatural beings. However, as he goes further, and develops his definition of “naturalism”, he says that,
an entity x is natural just in case x was created by God
So, for Halvorson, “naturalism” applies to anything that isn’t God. This seems to be problematic for his first discussion of naturalism, and what most people consider naturalistic. Are souls naturalistic? Angels? Demons? Well, in case you think I’m reading too much into his words, Halvorson provides a footnote, which says,
Theists might worry that this definition would classify angels as natural entities. My response: so what? As far as I can tell, theism doesn’t need to classify angels as supernatural.
So, while the original point of his paper was to say that methodological naturalism excludes spiritual beings, his actual definition, by his own words, includes them. So what is even the point of having the word “naturalism” if its meaning is so broad? Wouldn’t it be easier to simply say, “science can’t include God”? In that case, ID would be classified under methodological naturalism (Halvorson attempts to counter this, but that will have to wait for another post). In any case, marking one specific entity as something that science can’t touch is also special pleading.
My own definition of “naturalistic” vs “non-naturalistic” is this: <if a process can be modeled with a Turing machine, it is naturalistic, otherwise it is non-naturalistic. In other words, “naturalism” is computable. My point for making the cut at this point is not that everyone agrees with it, but rather that it is at least a workable definition. It is operational – we can test to see if a phenomenon can be modeled by a Turing machine (it is a well-known model with well-known characteristics). It matches the intuition – one of the things we expect from spiritual beings is that they have creativity and freedom, while mechanical things do not. It is held by individuals on both sides of the ball (Stephen Wolfram and Iris van Rooij on the naturalistic side, and myself and Eric Holloway on the non-naturalistic side).
My challenge to naturalists who don’t like my definition is this – come up with one which is similarly operational, matching the intuition, held by individuals on both sides of the divide, and isn’t guilty of special pleading. It’s harder than you think.
So, coming back to Halvorson, we can see that if a philosopher of science, in a paper about naturalism, can’t figure out if angels and demons are naturalistic, then we have a problem with defining naturalism. If we can’t define it, maybe it isn’t actually doing anything for us?