Anneli Jefferson and I on content holism

Content holism, to quote Gabriel Segal on this blog is the thesis that “the content of an individual’s concept depends on the totality of beliefs in which it features.”
The unattractive feature of content holism is that it loses a fantastically useful human invention that is central to rationality: error. We can see the problem clearly when we think about two types of disagreement : to use Quine’s terms, dogmatic disagreement and conceptual disagreement. Or less technically, genuine disagreements and disagreements that are merely verbal.
To illustrate with some examples:

DISAGREEMENT:

Alf: Zed is dead.
Beth: No he is not, I just saw him in the library.

MERELY VERBAL:

Alf: No not that Zed, I’m talking about Zed the astronaut, not Zed the librarian.

Here, what first appeared to be a disagreement turned out to be a misunderstanding. Of course this could have gone the other way.

ALF FALSE BELIEF

Alf: Really! That is amazing, I thought he was dead, I just heard that he had died on the radio. It must be a hoax.

Or

BETH FALSE BELIEF

Alf: No that can’t have been Zed you saw. I’ve just come from the hospital where I saw him die.

So the three options for a disagreement are 1. Alf has a false belief. 2. Beth has a false belief. 3. It is merely a verbal disagreement and neither have false beliefs.
The problem with any kind of holism is that everything falls into the third category and we end up never having false beliefs. This doesn’t amount to the fantastic news that we are always right, it is just absurd.
I guess a typical attractive way of avoiding this problem is by rejecting holism in favor of moleculism. This is to say that the content of an individual’s concept is dependent on a subset of all the beliefs that concept figures in. Anneli pointed out the problems with this. It is hard to come up with a principled way of saying which the meaning determining beliefs are. The Analytic/ Synthetic distinction won’t do because you end up with the same problem, namely which beliefs are analytically true. In the above dialogue, the sentence “Zed is Zed” may express an analytic belief or may be not. Nils’s suggestion that we make a dictionary met with derision for some unfathomable reason. So here’s my solution. Save holism by changing it to “dispositional holism”.

The content of an individual’s concept depends on the totality of situations in which the individual is disposed to have a belief in which the concept features.

So now I’ll flag up some advantages of this amendment;
Old Holism: Two people share a concept if and only if they have all the same beliefs in which the concept features. (absurd. Intuitively false)
New Holism: Two people only share the same concept if they are disposed to have the same beliefs given the same experiences. (Of course! This is true! Hooray!)

Disagreement problem solved: If Alf and Beth had a genuine disagreement then Beth must have had experiences such that, had Alf had those same experiences, then Alf would be disposed to believe as Beth does. This is the case. Had Alf seen Zed in the library, Alf wouldn’t have believed that Zed was dead. If Beth had watched Zed die in hospital, then Beth would believe that Zed was alive. The genuine disagreement depends on different experiences.

However, with the merely verbal disagreement this doesn’t happen. So even if Alf had just seen Zed (the librarian) in the library, he would not be disposed to believe that Zed (the astronaut) was still alive.

We can make of this a simple rule about belief content.
S grasps the content of p iff S is disposed to believe that p in situations where it is obvious that p.
Or a variation:
S grasps the content of p iff S is disposed to believe p in situations where the evidence warrants a belief that p.

This is supposed to be a mirror of truth conditional semantics.

This might frightened some linear thinkers since it is introducing epistemological terminology. But hey, it is these for these reasons I started studying epistemology. “Belief” is an epistemological term, and therefore so is “concept”.