Saturday, July 21, 2012

Information and Necessarily Coextensive Properties

In 'Why There Really Are No Irreducibly Normative Properties', Bart Streumer defends Frank Jackson's argument from the thesis (N) that necessarily coextensive predicates ascribe the same property, to the conclusion that there can't be irreducibly normative properties. After all, any normative predicate, e.g. "being right", will be necessarily coextensive with the vast disjunction D*: "being D1 or D2 or D3 ..." where each Di describes a right-making feature. But just because they pick out all the same actions across modal space, why should we thereby conclude that "being right" and D* ascribe the same property? The distinction between normative properties and their subvening natural properties seems like a straightforward counterexample to thesis (N).

In explaining away non-naturalists' persisting disagreement with him, Streumer suggests that they tend to "conflate properties with the meanings of the predicates that ascribe these properties" (28). But I think that non-naturalists are instead appreciating a deep and important connection that holds between properties and possible subject matters. Even Kripke/Putnam cases of the necessary a posteriori involve claims that are informative because they (indirectly) relate distinct properties:
[W]henever identity claims are informative, this is because they tell us that two distinct properties are co-instantiated by a single object. The cognitive significance of ‘water’, say, may be given by a certain complex functional property: roughly, being the clear drinkable liquid found in lakes and rivers around here. This differs from the cognitive significance of ‘H2O’, which is instead given by a certain chemical property. The claim ‘water is H2O’ is informative rather than trivial because it relates these two distinct properties. This is possible because the concept water is “gappy”: it refers to whatever actually fills the associated functional role. This functional role could, for all we know a priori, be filled by all manner of chemical substances. Hence it is informative to learn that H2O is the particular chemical property of the watery stuff.

The fundamental problem with normative naturalism is thus that it doesn't have enough properties to ground any new claims (or cognitive significance) beyond what could already be expressed in non-normative terms. But Open Question and Knowledge arguments establish that normative claims cannot be reduced to non-normative claims. When we call an act "right", we are saying something other than that it is "D1 or D2 or ..." But then what are we saying -- what new information are we conveying -- if not that the act possesses a distinctive further property (i.e., that of being right)?

Informative claims need to bottom-out in informative property ascriptions somehow, and it just doesn't seem that the naturalist has the resources to do this in the case of normative claims. (See also: 'Non-Physical Questions', for a straightforward application of this principle to refute physicalism in the philosophy of mind.)


  1. (1) I'm sure you know this, but Bart will agree that normative naturalism is mistaken. (He's an error theorist.)

    (2) Can't identity claims be informative because they tell us that two distinct predicates denote one and the same property? It can't be right to say that you need two properties to make an identity claim informative, because you need to *not* have two properties to make an identity claim true.

    1. Hi Alex,

      (1) Right, much of my post here is not directly addressing Streumer, but just using some of the things he says as a launching point. Though once we accept that any normative properties must be non-natural, then the error theorist can hardly appeal to thesis (N), because normative properties will constitute a straightforward counterexample (if there are any normative properties, which is precisely the question at issue).

      (2) As explained in the quoted section of the post, I do think you need two properties in order for the claim to be informative. Even for merely semantic information, we learn that the property of being denoted by the English word 'X' is coinstantiated with the distinct property of being denoted by the English word 'Y'. Of course, such merely semantic information is not very interesting. More worthwhile information may be found in the coinstantiation of the two distinct reference-fixing properties (as I described in the water = H2O case). But none of that would seem to help us ground the information given by normative claims.

  2. Richard,

    I'm usually sympathetic to your arguments against reductive naturalism. But let me play the devil's advocate for my own purposes of clarification.

    In your effort to defend normative non-naturalism, aren't you pulling too hard on naturalism at the same time? It seems like to prove your point you need reductive naturalism in general to be false. Granting this, I wonder if the informativeness of such claims as "water is H2O" can't be preserved as follows.

    The claim does indeed relate two properties, but those need not - and may not - be distinct intrinsic properties. The property "being the clear drinkable liquid found in lakes and rivers around here" may merely be a functional, second-order property of one and the same entity with a given first-order chemical property that realize it. But there in fact need not be two properties at all: why can't the functional property just be a second-order description, or designator, of the the very same thing after, say, some of its causal powers.

    Granted, in worlds with different laws, different chemical properties may realize the functional "property", and so the identity claim is informative insofar as this being the case under given laws is no trivial information. Still, in those different worlds, the different chemical properties could nonetheless be identical with the functional role they fulfill across worlds. Varying laws simply make for varying realizations of functional roles across possible worlds.

    So, the same may hold for normative "properties". They may well describe or designate first-order natural properties. And claims to the effect that "X is right" may still be informative insofar as, in other possible worlds, it could be that Y is right rather than X. Hence, even though normative properties would not be, on this account, *irreducibly* so, they may nonetheless be truly so, just as any second-order "property" can truly fulfill its functional role (e.g. biological, psychological), without it being necessary to add further properties to the facts. Thus, there would be nothing "over and above" natural properties, but talk of normative "properties" (with aforementioned qualifications) could still be warranted given the role they fulfill in our actual world, given, in particular, certain facts about us.

    What do you think?

    1. Hi Nicolas, I'm not sure I follow you. In particular, I don't think I "need reductive naturalism in general to be false". Functional properties are still natural properties, after all, so I took my account of water = H2O to be one which showed how even reductive naturalists need to (indirectly) relate distinct properties in order to make informative identity claims. I don't see anything inherently "non-reductive" about this. One could hold that being water and being H2O are one and the same property, whilst acknowledging that the predicate "is water" denotes this property in virtue of H2O having a (distinct, further) functional property: being a clear, drinkable liquid found in lakes, etc.

      I don't see how you can identify functional properties with chemical properties, given that they come apart in different possible worlds, whereas identity is non-contingent. (What you can do is use a functional property to fix the reference of a term like 'water', so that it rigidly designates whatever chemical property is coinstantiated with the functional property in the actual world. But that doesn't make the chemical and functional properties identical.)

      Finally, the supervenience of the normative on the non-normative would seem to preclude the possibility of the fundamental moral facts varying across possible worlds.

  3. Hi Richard,
    My point was that I took naturalists to be arguing that what you claim are distinct properties are in fact one and the same property, or that the functional property was a concept for the base property. And that it didn't seem unreasonable to expect such identity claims to be informative, if only for epistemological and explanatory purposes. So, if "being water" and "being H2O" are one and the same property, we're right to ask how supervenience could add properties to the world. My suspicion came from Kim's worries that talk of functional properties, if convenient, may not warrant acknowledging the existence of properties in addition (Mind in a physical world).

    Maybe functional properties and chemical properties cannot be identified if their relation varies across worlds. But just as pain may have multiple realizers across individuals, species, planets — and possible worlds —, I took it that water may have multiple realizers the same way. Identity remains non-contingent if it relates the functional property with either property in a disjunction D. (But it may be that your use of rigid designator is more appropriate to describe what I have in mind.)

    As for supervenience of the normative on natural properties, your point seems basically right, but only insofar as none of the right-making features among the base properties can be one of the laws of nature varying across worlds.


Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)

Note: only a member of this blog may post a comment.