Wednesday, February 13, 2019

Anti-reductionism and supervenience

In the philosophy of mind, those who take anti-reductionism really seriously will also reject the supervenience of the mental on the non-mental. After all, if a mental property does not reduce to the non-mental, we should be able to apply a rearrangement principle to fix the non-mental properties but change the mental one, much as one can fix the shape of an object but change its electrical charge, precisely because charge doesn’t reduce to shape or shape to charge. There might be some necessary connections, of course. Perhaps some shapes are incompatible with some charges, and perhaps similarly some mental states are incompatible with some physical arrangement. But it would be surprising, in the absence of a reduction, if fixing physical arrangement were to fix the mental state.

Yet it seems that in metaethics, even the staunchest anti-reductionists tend to want to preserve the supervenience of the normative on the non-normative. That is surprising, I think. After all, the same kind of rearrangement reasoning should apply if the normative properties do not reduce to the non-normative ones or vice versa: we should be able to fix the non-normative ones and change the normative ones at least to some degree.

Here’s something in the vicinity I’ve just been thinking about. Suppose that A-type properties supervene on B-type properties, and consider an A-type property Q. Then consider the property QB of being such that the nexus of all B-type properties is logically compatible with having Q. For any Q and B, having QB is necessary for having Q. But if Q supervenes on B-type properties, then having QB is also sufficient for having Q. Moreover, QB seems to be a B-type property in our paradigmatic cases: if B is the physical properties, then QB is a physical property, and if B is the non-normative properties, then QB is a non-normative property. (Interestingly, it is a physical or non-normative property defined in terms of mental or normative properties.)

But now isn’t it just as weird for a staunch anti-reductionist to think that there is a non-normative property that is necessary and sufficient for, say, being obligated to dance as it is for a staunch anti-reductionist to think there is a physical property that is necessary and sufficient for feeling pain?


Brian Cutter said...

One relevant difference between the mind-body relation and the descriptive-normative relation is that it's a lot more plausible that descriptive-->normative conditionals are a priori than that physical-->phenomenal conditionals are. (For example, "If Joe tortures puppies for fun, then Joe has done something wrong" seems knowable a priori, but "If Joe has firing C-fibers, then Joe is in pain" is not a priori.) Typically, a priori truths are metaphysically necessary, at least where there are no twin-earthable terms involved.

Uncommon Sense said...

"those who take anti-reductionism really seriously will also reject the supervenience of the mental on the non-mental"

I thought (a) those that take non-reductionism of the mental seriously often accept supervenience, but note that (b) supervenience isn't reductive, since a variety of dualisms and non-physicalisms--eg. epiphenomenalism, emergent property dualism--satisfy the supervenience relation.

Alexander R Pruss said...


While *some* descriptive to normative conditionals are plausibly a priori, it does not seem to me to be that plausible that all of them (at least where the antecedents are sufficiently filled in) are. Consider Aquinas' observation that if we are to give an equal good to a parent or a stranger, we should prefer the parent, but if the good is much greater when given to the stranger, we should prefer the stranger. Somewhere there must be a cut-off line in how much more valuable the good must be for the stranger to be preferred (or a cut-off line as to where it becomes vague, etc.). But it does not seem all that plausible that it is a priori where that line lies: it seems the sort of thing that is as contingent as the constants in the laws of nature.

Moreover, some of our naturalist friends will think that some physical to phenomenal conditionals are a priori: e.g., where there is nothing physical, there is nothing phenomenal.

Uncommon Sense:

I should have inserted a "tend to" in the sentence. I am inclined to think that those who accept supervenience tend not to be seriously anti-reductionist.