Monday, March 24, 2014

Deflation of predicates

Some deflationary theories take some predicate, such as "is necessary" or "is true", and claim that there is really nothing in the predicate for philosophical investigation—the predicate is not in any way natural, but just attributes some messy, perhaps even infinite, combination of more natural properties.

But I know only three candidates for a way that we could come to grasp a meaningful predicate. One way is by ostension to a natural property. Here's a rough idea. The predicate "is circular" might be introduced as follows. We are shown a bunch of objects, A1,...,An, and told that each "is circular", and a bunch of other objects, B1,...,Bm, and told that each "is not circular." The predicate "is circular" is then grasped to indicate some property that all or almost all of the As have and all or almost all of the Bs lack. But there may be many (abundant) properties like that (for instance, being one of A1,...,An). Which one do we mean by "is circular"? Answer: The most natural of the bunch.

The second way depends on a non-natural view of mind. It could be that our minds, unlike language, can directly be in contact with some properties. And it may be that a predicate tends to be used in circumstances in which both speaker and listener are directly contemplating a particular property, and that makes the predicate mean that property.

The third way is by stipulation. I just say: "Say that x is frozzly if and only if x is frozen and green."

The predicates, like "is true" and "is necessary", that are the subjects of these deflationary theories are not introduced in the first way if the theories are correct to hold that the predicates do not correspond to a natural property. Are they introduced in the third way? That is very unlikely. I doubt there was a first user of "is necessary" who stood up and said: "I say that p is necessary if and only if...." That leaves the second way, the non-naturalistic way. Therefore:

  1. If these deflationary theories are correct, naturalism is false.
Which is interesting since the motivation for the theories is sometimes naturalistic (e.g., Hartry Field in the case of "is true"[note 1]).

But in any case, the following is very plausible. Any properties we are in direct non-natural cognitive contact with are either innately known or natural. So, the deflated predicates must refer to innately known predicates. I doubt, however, that we innately know any entirely non-natural predicates. And that leaves little room for these theories.

More generally, the above considerations make it difficult to see how we could have any genuine non-natural, non-stipulative predicates. Thus, if we have good reason to think that P does not indicate a natural property, and is not stipulative, we have good reason to have an error theory about P.

Concepts of artifacts appear to be a counterexample. "Is a chair" is neither natural nor stipulated. My inclination is to say that it is not really a predicate ("Bob is chair" expresses some sentence about Bobbish reality being chairwise arranged, or something like that), which makes for a kind of error theory.


Heath White said...

A milder version of the last option you consider would hold that e.g. artifactual concepts are not in error, exactly, but only vaguely or partially specified. The critique would then be that, if "true" or "necessary" were like that, they would be unsuited to the central theoretical role they need to play.

Heath White said...

BTW, what is the difference supposed to be between " a chair" and " arranged chairwise"? It seems to me that anything that fell under one would fall under the other?

Maybe the objection is that we can allegedly quantify over chairs but not over stuff arranged chairwise. But we had to quantify over "Bobbish reality" in order to make this point, right?

Lastly, surely we learn the content of "arranged chairwise" by learning to identify chairs, not vice versa? (Related to a point against sense-data theories: we learn to chair-like sense-data by learning to identify *chairs*.)

Alexander R Pruss said...

Technically, the difference is that "... is a chair" purports to take an individual as its subject while "... is arranged chairwise" at best a plurality.

But I see the point: the problem comes up with whatever reformulation one gives, and so I now endorse your first comment.