Carnap vs Quine

Carnap

Frameworks

No metaphysics

Use whatever framework you wish for free

For every person $p$ there is a number $n$ that is the weight of that person in pounds. (Sortal quantifiers.)

Quine

* There is only one framework, the language we actually use

Realism

Ontological commitment

For every $x$, if $x$ is a person, then there is a $y$ such that $y$ is a number and $y$ is the weight of $x$ in pounds.

Price They agree on more than they disagree on. Metaphysics, as conceived earlier, is rejected by both. Quinean metaphysics is alive and well, but it is quite different from what was called metaphysics before.

Quine's devastating criticism of Carnap: there is no coherent distinction between "internal" and "external" questions, between questions within a framework and questions about a framework. Without the two factors, we are left with one language. Quine is generally acknowledged to have won.

Quine associated the internal–external distinction with the analytic–synthetic distinction, but it is not clear exactly how that goes. As Yablo makes clear, on a flat-footed reading of Quine's argument, it is obviously wrong. It is Carnap, in replying to Quine, rather than Quine himself, who gives a clear argument on Quine's behalf.

The flat-footed version: external=analytic and internal=synthetic. The cleaned up version (roughly as presented by Yablo): Adopting a framework is adopting a set of rules of language. Therefore, choosing a framework is choosing what the analytic truths are. If there is no analytic–synthetic distinction, then there is no separate procedure of choosing the analytic truths. "Analytic" and "synthetic" truths are not distinct, and elements allied to both appear in discovering/settling on any truth. The transition from "A whale is a fish" to "a whale is a mammal" is a familiar example of the point.

If we want to stop worrying and love strange (mathematical) entities, we need a neo-Carnapian reply to Quine.

Attempt 1. Price. Different truths are true in different ways. Within our one language, we have many different kinds of truth for different kinds of things. That is, in part, there are different rules for what makes a mathematical truth true and what makes a truth about medium-sized dry goods true. Those different kinds of truth can stand in for Carnap's frameworks, and so we don't need to worry about ontology. Note that Carnap thinks that there aren't any serious ontological questions, while Price only concludes that we needn't take ontological questions seriously. Given that we are working in a single language, Quine's "realism" is, for Price, tenable, just unimportant.

The view requires sustaining the claim that there are different kinds of truth and existence for different kinds of things. Quine's default position is that the reason we need, for example, different kinds of evidence for mathematical existence and existence of ordinary objects is that they are different kinds of objects—not different kinds of existence. So Price needs that the different kinds have different associated kinds of truth and existence. He claims that the burden of proof is on Quine's side. Quine endorses only very "thin" kinds of notions of truth and existence. The "thick" ones appropriate to the different domains are different.

The thin versions are disquotational. For medium-sized dry goods, the thick version of truth is correspondence.

I'm not sure what correspondence truth adds to disquotational truth, that is, I'm not sure what is thick about "thick truth": The correspondence account of "Brandy is a dog" is true is that "Brandy" refers to Brandy, "is a dog" refers to dogginess, and Brandy has dogginess. The first two of those clauses are themselves merely disquotational. The disquotational account says that ""Brandy is a dog" is true" if Brandy is a dog. The correspondence account says ""Brandy is a dog" is true" if "Brandy" refers to Brandy, "is a dog" refers to is a dog, and Brandy is a dog. All that it adds, it seems, is more disquotational facts. In addition, I don't know what any other "thick" account of truth (for example, for mathematics) would be like.

Price's attempt to rescue Carnap from Quine is based on an attack from the right: Quine says there is no framework–internal distinction. Price accepts that, but argues that there are distinct forms of truth and existence. He simply presents us with a variant distinction to fix up Carnap's position.

Yablo's attempt to rescue Carnap from Quine is based on an attack from the left: Quine says there is no framework–internal distinction. Yablo adds that there is no metaphorical–literal distinction.

We can certainly often explain things using metaphors that are hard to express literally. We're confronted with new situations all the time, and we often don't have any words that fit exactly. We use a metaphor. Often, over time, the metaphor becomes a literal usage. When we say that a politician's behavior is unsportsmanlike, are we using a metaphor, or saying something literal? The usefulness of a metaphor does not depend on its literal truth. Thus, I can use metaphors taken from fiction or myth to explain things. When I say, he was acting like Santa Claus, my meaning is neither more nor less clear because there is no Santa Claus.

Yablo says that, for example, when I use mathematics in physics, much the same thing is going on. Here is an example that I prefer to Yablo's, which are too simple by only employing counting: Newton's law that mechanical force in a direction is proportional to the second derivative of position as a function of time in that direction makes sense to us independently of what we think about the existence of functions and derivatives. The case is a lot like the Santa Claus case.

So far, all that follows is that when we present theories in part using metaphor, the theories aren't committed to the truth of the statements within the metaphor, and it isn't committed to the existence of any objects mentioned within the metaphor. For example, it can be informative, and either true or false to say that someone is acting like Santa Claus quite independently of the fact that there is no Santa Claus. It is important for Yablo's purpose that we could use the Santa Claus metaphor without knowing whether or not Santa is real. That means, if his analogy is correct, that the assertions of mathematical physics about the physical world just don't depend on whether or not mathematical objects exist.

That is all very well and good for metaphor, but, one wants to say, surely to know what someone's theory leads to commitments to, we need a serious, in particular, literal presentation of that theory. Quine happily admits that we do things for the sake of … without being committed to the existence of sakes, and similarly for behalfs. He talks about meanings quite freely though he argues words don't have meanings—that is more comfortable in English. Officially, in the end, one has to be more careful. To see what you are committed to, I have to see what you are willing to assert literally when being careful, not using shortcuts like sakes and metaphor. Now Yablo's criticism: just as there is no analytic—synthetic distinction in the sharp and well-defined sense Carnap requires, there is no metaphorical—literal distinction in the sharp and well-defined sense Quine requires.

Quine does not deny that "bachelors are unmarried" seems more analytic than does "there is a projector in the room." He rather argues that there is no principled distinction, only one of degree and habit. That is not enough to support a distinction between different criteria of existence for different frameworks. Similarly, Yablo only needs to argue that there is not a sharp, principled metaphorical—literal distinction, not that we don't have inclinations in one direction or the other about particular cases. As Yablo points out, lots of philosophers have thought that compared to the metaphorical—literal distinction, the analytic—synthetic distinction is a model of clarity. If the second isn't tenable, it is hard to see how the first could be.

What is the conclusion of Yablo's argument? Though he doesn't draw it, it is, it seems to me, that the question of what is real is itself untenable. If there is no way to distinguish literal claims from metaphorical ones, then there is no way to distinguish "real" existence claims from merely metaphorical ones for which the question of truth simply doesn't arise. Quine's historical opponents (even including Price) were trying to rescue a sense in which, for example, tables and chairs are real, but numbers, possible worlds, and moral responsibilities are not. Yablo's argument seems, instead, to obliterate any such distinction. In the end, I'm not sure which side Quine and Carnap would have taken Yablo to be on.

The expressed worry is But the background worry is

It is, to say the least, less than clear how they are related.

-- ShaughanLavine - 07 Apr 2008 - 09 Apr 2008 - 11 Apr 2008