Carnap and Quine

AssignedQuestionsCarnapandQuine

ResponsePapersCarnapandQuine

-- ShaughanLavine - 23 Aug 2006

Carnap

For Carnap there are two kinds of questions about existence: internal and external.

Internal

Insofar as we understand a framework, we understand how to answer internal questions of existence, because the framework tells us how.

• Constraint on an adequate discourse: it must give us a criterion of reference and existence.

Ian says that sometimes a framework gives us ways to find out whether something is evidence for the existence of something else (symptoms for a disease): we learn from our medical framework, not which symptoms are evidence for which diseases, but how to discover which symptoms are evidence for which diseases. Even that is too simple, since we have criteria for what constitutes a symptom and what constitutes a disease that come into play before the question arises.

That is all true, and it may not be compatible with Carnap's informal description of frameworks, but his official account is the formalized one, and there a framework is just an axiomatically presented theory that allows us to get evidence for the existence of things, and there he intends to be ok about such complex issues.

Orlin says that this suggests a bootstrapping account in which a framework suggests how that very framework should be extended. So far as I know, Carnap never considers such a thing. We will discuss principles of abstraction viewed in two ways: describing objects in a domain and extending the domain. Once a domain has been extended, that in itself may make further extensions possible. Kit Fine, who discusses that in the most formal detail generally favors immediately moving to a maximal closure. Orlin remarks that even if closing is a natural procedure for mathematics, it may not be in other cases.

Liz is discussing string theory. Since I know nothing about string theory, I want to discuss a simpler example: how we get to physical objects from a phenomenalist framework. Carnap tried to construct physical objects within a phenomenalist framework, so that physical ("material" might be better) objects would turn out to be nothing more than some precise version of "permanent possibilities of sensation" (Mill). Quine thinks that the project is doomed. He thinks that we just introduce a new framework of material objects, and—to the extent that the introduction is justified at all—it is justified only in the external sense of being a better (simpler) way of accounting for phenomenal experience or of just being a more useful theory.

We have now moved to

External

External "justification" is outside any framework. There is no standard, no criterion for existence, that stands outside any framework. Therefore the only possible kind of justification is pragmatic, and I just mentioned two versions: pragmatic reasons why a framework represents an improvement over an old framework, and autonomous virtues of a framework.

It is consequence of this attitude that the traditional metaphysical questions are incoherent, if posed as external questions, and trivial, if posed as internal ones. Carnap doesn't actually say anything like "incoherent," he says meaningless or noncognitive.

Carnap is a pluralist about frameworks: you are entitled to use any framework you like for any purpose for which you like it (provided it is successful). In particular, he countenances using different frameworks for different purposes. That does not eliminate the question whether various frameworks are compatible, whether one framework can be extended by another, and the like. I think Carnap has largely won on his view that external metaphysical questions can only be discussed on pragmatic grounds and grounds of the internal virtues and vices of particular frameworks: most metaphysics today is explicitly either about the internal coherence of proposed frameworks or the analysis of a framework envisioned as an addition to another, specified framework. (Arguments for trope theories take the form that they make it possible to solve problems with which object theories are inevitably stuck. Arguments against abstract objects often presuppose a physicalist or naturalist framework.)

Carnap is often read as thinking that there can be no final answers about what are appropriate frameworks. He does think that we have no such answers in our present state of knowledge, and he thinks the grounds for preference have to do with the (pragmatic) success of the frameworks, but he certainly thinks that the physicalist framework is superior to the Homeric gods framework, which is therefore no longer a live option, and that internal problems with framework could serve to rule it out (which he outright says about the most obvious case: consistency).

Liz says that the question whether there really are, for example, physical objects, seems meaningful and important even if perhaps unresolvable. Carnap uses the word "meaning" in an odd way, but what he is claiming is that there are no agreed-upon procedures or grounds on the basis of which one could arrive at evidence either for or against the claim that there are no physical objects. That is not to say that we can't assess the utility and internal coherence of the way in which we actually talk about physical objects.

David asks whether the view is that ontological disputes are just verbal. Carnap has been read that way, and there is some reason. That is clearly not what Quine thinks, and I don't think it is what Carnap thinks either, but that is less clear. The pragmatic utility of a framework certainly depends on extra-linguistic factors. Carnap, for example, thinks that physical object language is a better way of accounting for experience than purely phenomenal languages as they have been proposed, and so, since as an empiricist he is attracted to purely phenomenal languages, he took up the project of showing how to reduce talk of physical objects to phenomenal language. The Quine-Putnam indispensability argument also takes this form: our best physical theories are inelimably committed to mathematics (and that is a nonlinguistic fact) and so we are committed to the existence of mathematical objects.

Carnap is committed to the claim that if two frameworks are equivalent in their pragmatic virtues then the question of which to adopt is a matter of convention or decision, that is, that there is no further fact of the matter about which framework includes more truths.

Carnap says

• The mere use of names for some entities doesn't commit you to them: you need to quantify over variables for which the names may be substituted to get a commitment.
• He seems to think an empiricist can rest content with a framework with nonempirical commitments because such a framework doesn't really (presumably, in an external sense) commit you to the objects.
• He envisions nominalists and Platonists might both think number talk is meaningful, though they disagree about whether number terms refer, which comes close to raising the question when a framework is meaningful.

The question whether a discourse can sensibly introduce talk of abstract objects is about whether you can quantify, about general statements about numbers, not about, for example, just the number two.

Quine

Quine shows how we can avoid getting tangled in Plato's beard by revising the language we use to pose certain questions in a way suggested by Russell. Moves like that are what have resulted in the modern centrality of philosophy of language to philosophy.

When we say that something is, we are ipso facto committed to the existence of that thing. In so saying, Quine shifts his subject from what there is to what saying certain things commits one to, from ontology to ontological commitment. This idea of the ontological commitments of a theory, as distinct from the ontological commitments of a theorist is novel and important.

"There are white dogs" commits us to white dogs, not whiteness or dogkind: quantification, not naming, leads to ontological commitment.

Our conceptual scheme is what yields our theoretical and hence ontological commitments. It plays a role much like that of Carnap's frameworks. Unlike a framework, there is no place, not even linguistic, to stand to negotiate between conceptual schemes. That follows from the nonexistence of an analytic-synthetic distinction, which we have not discussed.