The Substantive Question Argument

Reductive accounts of mind and morality render real questions meaningless

What’s the best or most central reason to reject naturalism about mind and morality? Sometimes it’s suggested that normativity is simply “too different” from matter to be reducible to it. (Both Parfit and Enoch have argued in this vein.) But that seems too quick: plants and stars seem very different from atoms, yet they’re wholly constituted by atoms. Mind and morality may be even more different from matter—being non-concrete and all—but the same might be said of software (or other high-level patterns in reality). I think the non-naturalist can do better.

Clarifying Naturalism

Metaphysical naturalism is the view that all real properties ultimately reduce to natural ones: the kinds of properties studied by the empirical sciences. For example, metaethical naturalists identify the property of moral rightness with whatever natural property makes an act right (e.g. maximizing happiness). A physicalist about the mind holds that consciousness just is some physical or functional property of our brains. Non-naturalists and property dualists, by contrast, view these cases as involving correlated but distinct properties: the underlying natural ones give rise to some further (irreducibly moral or mental) feature.

These reductive views appeal to many. They’re ontologically parsimonious. They align (vibes-wise) with the extraordinary success of the scientific worldview. And they avoid “spooky” posits like irreducible moral facts or non-physical qualia.

One obvious cause for hesitancy: it’s extremely natural to hold that the natural right-making properties explain why an act is right, and that physical or functional properties explain why we are conscious. But things don’t explain themselves. The intuitive picture only makes sense if we have two distinct properties, one of which systematically gives rise to the other, in law-like fashion. Reductive accounts struggle to make sense of these natural thoughts.

That’s just a quick intuitive concern. A stronger objection can be found by exploring the deeper structural problem for reductionists: they can’t make sense of the fact that debates about the distribution of mental and moral properties are substantive. If physicalism is true, it’s a terminological question whether AIs and octopuses are conscious. If metaethical naturalism is true, fundamental moral disputes similarly bottom out in mere semantics. Let me explain.

Delineating the Sun

Imagine two microphysically-omniscient astronomers disagreeing about precisely which collection of atoms constitutes the Sun. Their proposals overwhelmingly overlap—they just differ slightly on where to draw the boundary. Maybe one includes a particular borderline atom in the Sun’s outermost fringes and the other doesn’t.

This is clearly a merely terminological disagreement. The two parties diverge on whether the label ‘Sun’ picks out the minimal collection S, or the slightly larger collection S*. There’s no further issue at stake about which they disagree. There’s no special further property of really being the Sun that one collection of atoms has and the other lacks. Once you know the physical facts—where the atoms are, how they’re behaving—there’s nothing more left to discover. The remaining question is just about how to use a word.

My key claim is that disagreements about the distribution of minds and moral properties are not like this. We’re trying to track independent reality, not just disputing empty labels.

Disputing Consciousness

Would you still be conscious if your neurons were replaced by functionally identical silicon chips? What if the functional role of your neurons were instead realized by billions of people communicating via walkie-talkies (as in Ned Block’s “Chinese Nation” thought experiment)?

These seem like substantive open questions. We know all the relevant physical facts: by stipulation, the silicon brain is functionally identical, just made of different stuff. If physicalism is true, and the physical facts are all the facts there are, then there’s nothing more to wonder about. The physicalist can ask the semantic question of whether ‘consciousness’ picks out functional property P₁ or biological property P₂. But given that we already know the physical setup (the silicon brain has P₁ but not P₂), there’s nothing in reality left open—just a terminological choice to make.