Let’s say you are inclined to go along with me so far: you agree that there is something troubling about metaphysical thinking, and you think that naturalism in particular should tread as lightly as possible where metaphysics is concerned.

We need a bit more, though, unless we want to announce that we have a metaphysical intuition that metaphysics is bad. (Actually, I am tempted to stop there.) Then, we get complications. It’s not entirely clear how to separate metaphysical from nonmetaphysical talk, and we can’t refer to any consensus on why exactly metaphysics is problematic.

Take statements like like “God is the creator of all,” or “God is an infinite person.” These might be among our defining examples of metaphysical claims—just the kind of theological monstrosities that the logical positivists wanted to single out as being meaningless. But here’s one immediate complication. Such statements, whatever they mean, do not always work the same way. In a classic metaphysical system such as, say, medieval Christian scholasticism, they function one way, but they don’t do exactly the same job in another, say, something more Neo-Platonic or Muslim. So it looks like we may have to consider metaphysical systems, rather than isolated statements. And then there is the way such statements work in a more ordinary religious context, where it’s hard to talk about church or mosque-goers depending upon any fancy metaphysical system at all.

Let’s say we want to take our suspicions of metaphysics in a neopositivist direction. Old-style logical positivism and naive verificationism is dead, and for good reason. But I think there is still merit in the basic positivist intuition: if something is supposed to be a fact, we should ask what difference that fact makes. There should be something in our encounters with the world that can count for or count against the fact claim in question. Such a view invites us to be sophisticated in our notions of reality-testing. We can, for example, recognize that investigating the world is not a theory-independent activity, and that we do not put isolated statements to the test. We can allow for context and history. And we can guard against introducing a hidden metaphysics in the name of excising metaphysics. Logical positivists, for example, had a simplistic view of observation, which could shade into idealism or dualism when discussing scientific observation. Sophisticated neopositivists can avoid such pitfalls.

This can be a problem, however, if, like the early positivists, we want some kind of philosophical or conceptual analysis to be our main tool for making the world safe from theism and other alleged metaphysical fantasies. Because now, the systems in which “God is the creator of all” and so forth find a role, plus the less systematic talk of ordinary religious people, almost invariably will make contact with reality tests. They won’t be obviously empty pronouncements. Sure, if armchair metaphysical reflection is the driving force behind a claim we encounter, we should expect plenty of outright nonsense as a result. Let conceptual analysis take the field! But this should be rare. More often we will end up in a gray area, where metaphysics takes a role in forming a picture that does, at least at its margins, allow for a degree of reality testing.

For me, this is not a source of worry. It tells me that to criticize theism and present a naturalistic view of the world, we should do something like what I attempted in The Ghost in the Universe. We should draw on philosophy, but also on the natural sciences, critical history, etc. etc.—with none of these disciplines having a privileged, supervisory role. We shouldn’t be doing metaphysics, in the sense that this is not another quest for what is really Real and rationally Necessary. Yet we shouldn’t expect too much from antimetaphysical philosophies in slaying our dragons either. A (broadly speaking) scientific approach will serve us better.

There is another advantage I perceive in such an approach. If we are encouraged to look for ways in which metaphysical systems might make contact with reality tests—rehabilitating what might at first seem to be bewildering claims that make no difference in the world one way or the other—this rehabilitation process should work even more easily with what is put forth as metaphysical naturalism. It might take no more than a minor restraint on our Platonic tendencies to rehabilitate metaphysical naturalism.

Saying this doesn’t end any debate. If, like I do, you want to lean heavily on a scientific approach, there will be an accusation of scientism to face. There are philosophical critiques of metaphysics that do not have a neopositivist flavor, and which might take naturalism in a different direction. And there is the question of whether my way of rehabilitating metaphysics actually transforms it, so that I end up not addressing a real challenge to naturalism.

I’ll leave these to further installments.