Naturalness for properties is usually identified by what it does for us – the natural properties are those which are reference-magnets, play roles in scientific or causal explanation, are potentially ontologically fundamental, and so on. How much is being smuggled in by the use of the term ‘natural’? I can think of at least three different available readings for the word ‘natural’ in the phrase ‘natural property’. I suspect people are differentially motivated by these different readings.
First, a property could be natural by not being ‘supernatural’. Just about any mundane everyday property fits this criterion equally well – being a table is just as non-supernatural as being an electron. This reading seems not to do justice to most uses of the notion of natural property, since it doesn’t distinguish grades of relative naturalness, and allows (for example) gerrymandered distributional properties to be as natural as intrinsic fundamental properties. Grue comes out perfectly natural. This can’t be the notion of naturalness at stake here.
Second, a property could be natural by not being artificial. The motivating idea for this is that natural properties are discovered, non-natural properties are invented. But this reading seems to change the subject somewhat, bringing in the realism/nominalism debate where we might have hoped to avoid it. It also makes it difficult to allow for degrees of naturalness; unless complex properties can be part-discovered, part-invented?
Third, a property could be natural in the sense of being a natural choice for application in our theories – this is nicely neutral between realism and nominalism. This reading does seem to get a handle on degrees of naturalness: if we’re looking to describe or manipulate reality, green is definitely a more natural choice than grue in the vast majority of circumstances. However, this form of naturalness seems not to be intrinsic to the property – prima facie, it is a complex relation between a theorizer, their circumstances, and a property. It’s not clear whether this sense of naturalness is unacceptably anthropocentric, though I’m inclined to think it isn’t. It’s also not clear that we can’t pose a further demand for explanation of this notion of naturalness – what makes a particular property natural in this sense? The worry is that a notion of meta-naturalness will be needed, leading to regress.
Fourth, a property could be natural in the sense of actually being part of naturalistic inquiry. ‘Naturalized epistemology’ is born of the metaphor of the philosopher becoming naturalized into the scientific community, as refugees are naturalized into the country which receives them. On this more deflationary view, a natural property is one which has widespread application in our mature scientific theories. The interesting thing about this is that a property could be natural in this sense without being ‘ontologically fundamental’ – maybe science, despite its success and maturity, gets some details wrong.
Someone must have written about this contrast already – anyone any idea where? I’m interested mainly in the latter two readings, but I get the feeling that the first two readings are often projected on to the natural property debate, leading to actual or at least potential confusion.