Yet another way to think about Quine's critique of Carnap

Several of the blog entries here have been about the Quine-Carnap debate over the status of analytic truth. Generally, I don't feel the force of Quine's arguments as they are usually presented, either because his interpretation of Carnap is unfair or inaccurate, or the arguments just aren't that persuasive. Multiple commentators on the Quine-Carnap debate have suggested that the two are 'talking past each other,' at least to some degree. So, I am constantly trying to find a way to make Quine's view make sense to me, AND simultaneously really disagree with Carnap. This seems like installment 19 or so in that endeavor.

Carnap and Quine agree that language can be studied at various levels of abstraction. Using Carnap's taxonomy, we start at the level of pragmatics, where we study how individual speakers use expressions under particular circumstances. This level contains the most detail: speakers, their circumstances, plus the meanings of the words for particular speakers under particular circumstances. At the next, more abstract level, we have semantics, which abstracts away from particular speakers and particular circumstances. And at the highest level, we have syntax, which abstracts away the meanings of words, leaving just the symbols, the way they are put together, and which strings follow from others.

In each transition from pragmatics to semantics to syntax, some information about language is omitted/ discarded. (Like the move from Euclidean geometry to neutral geometry, which drops the parallel postulate.) Now, we can conceive of Quine's indeterminacy of meaning thesis (the radical translation thought experiment) as critiquing Carnap in the following way: Carnap is importing or introducing new information at the semantic level, because the semantic facts Carnap includes in a semantically-characterized language [a "semantic system"] cannot be 'read off' even the information contained at the pragmatic level. The analogy in the geometry case shows why this is clearly an unacceptable maneuver. It would be: thinking that there exists some claim that could be proved in neutral geometry (= Euclid's first four postulates only) but couldn't be proved in Euclidean geometry.

This may not be Quine's actual worry; his concern may stem from the fact that applied semantics (or whatever branch of language study) underdetermines pure semantics (or whatever). However, Carnap is perfectly happy to accept that claim: Creath says this is why Carnap's copy of Word and Object Ch.2 (Indeterminacy of Translation) has no marginalia. [But how does the geometry analogy fare here? Would Carnap admit that applied geometry underdetermines pure geometry? My guess is yes; and that that's not so bad...

Help me write a book

This post falls under the category of "shameless begging." I am in the later stages of writing a book (based on my dissertation) called Carnap, Quine, and Tarski's Year Together. The plan is to send the manuscript to the publisher in about two months, i.e. very early July. Between then and now, the only thing I'm working on is the book.

I need your help with this. I would greatly appreciate any feedback, large or small, on the manuscript. It definitely needs to be looked at by a fresh set of eyes (preferably attached to a clever mind, but I'll take what I can get). The manuscript can be downloaded, as a 1MB Word file, here.

Thanks in advance to anyone who lends a hand. In my limited experience, the greater the number of people who tear apart something I've written, the better the final product is.

Carnap on what's 'really wrong with the Aufbau'

When I was going through some of the photocopied material I have from the Rudolf Carnap archive, I found the following interesting (to me) remark. Carnap is discussing with Nelson Goodman a draft of Goodman's dissertation (A Study of Qualities, which much later became The Structure of Appearance). Goodman points out various supposed technical defects with Carnap's method of quasi-analysis presented in the Aufbau, and Carnap says roughly the following:

The real problem with my Aufbau is not the various counter-examples that can be constructed against my particular version of quasi-analysis (which I already knew about), but rather the assumption of extensionality.

(I haven't quoted, because I don't think it's allowed.) I like this because (a) it's a really smart philosopher saying 'My book rests on a mistake,' and (b) it fits with my pet theory that the real break between Carnap and Quine stems in large part from Carnap moving away from the extensional languages he espoused in Logical Syntax, while Quine held fast to the extensional standard throughout his life (see the posthumous "Confessions of a Confirmed Extensionalist").

Is 'p is a priori' itself a priori?

My most recent posts have been far too long by blogging standards, so I am determined here to be brief. Most philosophers (though not all, e.g., anyone who holds the Quine-Putnam indispensibility thesis) believe the following sentences are true:
(1) "'2+3=5' is true a priori"
(2) "'The earth is round' is true a posteriori"
The question is: are (1) and (2) true a priori or a posteriori? Put metaphorically, does experience teach us that logical theorems are known independently of experience? I am not sure there are any non-question-begging arguments to be given one way or the other; if there is one, I would love to hear it.

That's all I wanted to say. If you are curious about what motivated me to think about this question, keep reading. There are two motivating sources:
First, my dissertation, which deals with the academic year Carnap, Tarski, and Quine spent together at Harvard in 1940-41. In a private conversation Carnap had with Quine, one way they formulate the difference between themselves is as follows.
Carnap: 'p is analytic in language L' is itself an analytic statement.
Quine: 'p is analytic in L' is a synthetic statement, to be settled by a behavioristic investigation into the linguistic habits of L-speakers.
Granted, 'analytic' is not identical to 'a priori' -- but for Carnap, they were extensionally equivalent, and the question above is very close to this issue.
Second, for the last 10 years, van Fraassen has been suggesting that we think of empiricism not as a theory or assertion but as a stance. (See "Against Naturalized Epistemology" (1995) in On Quine and 2002's The Empirical Stance.) The primary argument he offers is that empiricism, if conceived as an assertion, is self-defeating: "All knowledge about the world is a posteriori" (or any other slogan intended to capture the empiricist's thesis) will be difficult to construe as having experience as its source. And, van Fraassen says, if the empiricist thesis cannot be justified on the basis of experience alone, then it fails to live up to its own standards, and is therefore self-defeating. But van Fraassen is assuming that "such-and-such is a posteriori" must itself be an a posteriori claim. And that is taking for granted an answer to my question above.

This week's complaint about "Two Dogmas"

I am willing to admit that Quine's "Two Dogmas of Empiricism" has many virtues, and if I ever write anything one-tenth as intelligent or one-hundredth as widely read, I will count my philosophical career an unequivocal success. However, I find it a very frustrating piece when it is considered as a critique of Carnap's views on language circa 1950. A number of people who have considered "Two Dogmas" in this light have also felt that it does not directly rebut Carnap's views, but instead offers (to put it in Kuhnian terms) a different paradigm for understanding language (recently, see P. O'Grady's 1999 article in PPR). Interestingly, the first person to suggest this interpretation of "Two Dogmas" was apparently Carnap himself (see Howard Stein's "Was Carnap Entirely Wrong, After All?", Synthese 1992). This post is about my most recent round of frustrations in attempting to read "Two Dogmas" as a straightforward argument against Carnap.

My gripe here concerns section 4 of the paper, "Semantical Rules." Quine writes (page numbers from From a Logical Point of View):

Once we seek to explain 'S-is-analytic-for-L' for variable 'L'..., the explanation 'true according to the semantical rules for L' is unavailing; for the relative term 'semantical rule of' is as much in need of clarification, at least, as 'analytic for.' (34)

So Quine is telling us that there is no decent explanation or clarification of the term 'semantical rule of,' if we generalize away from particular languages.

It seems to me that the next question we should ask is: how does Carnap characterize these semantic rules? Is it as unclear as Quine alleges? To answer that, we have to look at an extremely underread book: Carnap's Introduction to Semantics (1942). There Carnap says:

a semantical system [is] ... a system of rules, formulated in a metalanguage and referring to an object language, of such a kind that the rules determine a truth-condition for every sentence of the object language, i.e., a sufficient and necessary condition for its truth. In this way the sentences [of the object language -GFA] are interpreted, i.e., made understandable. (22)

What are some of Carnap's examples of semantic rules? (Where English is both the object and metalanguage):
-'Chicago' designates Chicago etc.
-The truth-tables for the usual propositional connectives

Now I want to know what Quine finds unacceptable (unclear, unexplained, 'unintelligible' he says elsewhere in "Two Dogmas") about Carnap's characterization of semantic rules. This isn't just a matter of my pounding or kicking the table and saying 'I really do understand them' -- rather, it seems to me that if Quine is right that these things are unclear, then logic (at least logic that is not purely syntactic/ proof-theoretic) is unclear. In order to specify an artificial language in model-theoretic terms at all, we need to be able to say things like "Modus Ponens is an inference rule in such-and-such language" -- and if that is not clear, then logicans working in model theory are stumbling around in an unintelligible haze. If we cannot (intelligibly) lay down/ identify semantical rules such as this one and the others Carnap mentions, then the logician cannot (intelligibly) specify interpreted languages to study.

Thus concludes my main rant. A couple further things should be said before signing off, though. First: Quine, in "Notes on the Theory of Reference" and elsewhere, says that "'Chicago' designates Chicago," "'Snow is white' is true iff snow is white" and the like are basically comprehensible (p.138 in From a Logical PoV). That this is in tension with Quine's criticism in "Two Dogmas" has been pointed out (in different terms) already, by Marian David in Nous 1996. Second, there is one real difference between the semantic rules Carnap uses and the ones Quine favors and we usually use today, pertaining to hte interpretation of predicates and relation letters. Carnap sets up the following as a semantic rule: 'Blue' designates the property of being blue.
And here Quine would have a (more) direct disagreement with Carnap: Quine rejects intensional languages in general, and properties (for Carnap) are to be understood intensionally. So to generate a clean Quinean argument against semantic rules from the above quotation, we can fall back on Quine's rejection of non-extensional discourse as unclear and unintelligible. But this view of Quine's has won far fewer supporters than his rejection of the analytic-synthetic distinction.