[hist-analytic] Carnap on Philosophy

Roger Bishop Jones rbj at rbjones.com
Thu Aug 20 16:53:47 EDT 2009

This is a rather selective response to Steve's last
rather challenging message on Carnap.

First I must clarify a little the extent to which I
am defending Carnap's philosophy.

I have asserted in the first instance that Carnap "was clear"
that the propositions of philosophy are analytic, and in
response to Steve's scepticism on this I supplied some
I hope that I have satisfied Steve on that point.

In response, Steve has expressed doubts about the consistency
of that claim of Carnap's with Carnap's own philosophical writings.

On that point I have disagreed with Steve (and continue to do so).

I propose for the time being to stick at that, i.e.
to stick to the defence of Carnap against the claim that
his philosophy was inconsistent with his claim that the
propositions of philosophy are analytic.

I do not intend to offer any general defence of his philosophy.
For example, his methodological position in
"Philosophy and Logical Syntax" is more specific than that,
it describes how one should go about formally establishing
such propositions, and, though he does do a considerable
amount of technical work which is clearly intended to
facilitate that kind of demonstration, it is not the case
that he proves all his philosophical propositions in the
manner which he recommends.
Further to my accepting that Carnap falls short of his own
ideals in that respect, I would not myself recommend the
methods he proposes.

Now, on the point which I am still prepared to defend,
viz. that the philosophical propositions which Carnap
asserts either are or are believed by him to be
(usually the former) analytic, I have a couple of further
points of clarification to offer before responding in
detail to some of Steve's points.

First I must clarify the purpose of the analogy I drew
between mathematics and philosophy.
The point of this was simply that many things that
mathematicians and philosophers assert are not properly
construed as "propositions of" mathematics and philosophy
So I am not claiming that every assertion which Carnap
makes is analytic, and I am asserting that when Carnap
says that the propositions of philosophy are analytic
he is not saying that everything which he asserts is
Obvious examples of propositions which he asserts but
would not claim in the required sense to be propositions
of philosophy, are his claims about the views of
other philosophers.

Further to this, a great deal of what Carnap writes
are not "propositions" at all, they are definitions
or proposals, either of for languages or concepts.
This is a large part of what he conceives of as the
purpose of philosophy when he describes philosophy
as "the logical syntax of science".
The use of the term "logical" in this phrase is intended
to exclude the descriptive analysis of existing scientific
language (let alone "ordinary" language), which he does not
see as falling under his method, and does not see as
properly philosophical (he would classify this as empirical
science, possibly linguistics).
He sees philosophy as proposing languages for use in the
empirical science, in the same way as he sees Principia
Mathematica as offering and demonstrating the use of
a language for mathematics.

In this context, it would be consistent with Carnap's
philosophy if Carnap asserted no philosophical propositions
whatever, but exclusively concerned himself with the
construction of formal languages for various branches of
science and the discussion of the pragmatic considerations
which are relevant to the adoption of such languages.

I don't believe he does this, I think he does actually
assert propositions which he believes to be both philosophical
and analytic, though from my limited reading of Carnap
I am not actually aware of cases where he has demonstrated
such propositions in a manner consistent with his syntactic
method (I would however be rather surprised if there were
no examples in "Logical Syntax of Language").

It is worth noting here, that for Carnap to be consistent
on this point, his assertions only have to be analytic according
to his own conception of analyticity, they need not be
analytic in Kant's or in Kripke's usage of that term.
Furthermore, insofar as Carnap condemns metaphysics,
for consistency he has only to avoid asserting propositions
which comply with his own description of proscribed
metaphysics, and this is greatly different from Kripke's
conception of metaphysics (and I would say almost everyone
This is particularly relevant for those many people who
look at Carnap's writing and immediately see lots of
metaphysics, most especially if they see this in a volume
like "Meaning and Necessity" which benefits from the very
considerable widening in Carnap's conception of acceptable
(not metaphysical in the proscribed sense) language.
In particular, any apparently metaphysical language can
be rendered acceptable to Carnap once it is incorporated
into the definition of a language, for then the apparently
metaphysical claims are given meaning by the language,
the internal questions become respectable, and the external
questions can be passed over so long as the language passes
some pragmatic criteria.

On Tuesday 11 August 2009 17:21:04 Baynesr at comcast.net wrote:
>The very first thing I want to ask Roger is this: give me an example of the
> successful employment of the methodology he advocates in solving a
> philosophical problem. 
> If you are going to defend a method
> give me an example of its success, NOT in general terms, but input/output.
> Other minds? External world? The Self? Ethics? State the philosophical
> problem and then in a few words tell me how it’s solved. Russell could do
> it; Tarski did it. Ok? So now give us a philosophical problem that has been
> solved.

The kind of problem you are looking for here seems to me not to be
characteristic of Carnap's philosophy, for two reasons, one positive
and one negative.

On the positive side, the kind of problem which interested Carnap
was typically one of devising the best language for some scientific
purpose, or of devising the best method for the conduct of science
or philosophy.  The result of solving such problems are not

On the negative side, very many of the "traditional problems"
of philosophy are in Carnap's scheme of things trivialised by
Carnap's method.

Take for example, the question, "Are there numbers?" which is
discussed by Carnap who make this point about it.
Carnap's recommended method for doing arithmetic proceeds as follows.
First you devise a language for talking about numbers and define
its semantics, for this purpose simply assuming the existence
of numbers (though taking reasonable steps to ensure
that the assumptions on which the semantics is based
are logically consistent).
Questions about the existence of numbers are then expressible
in this new language for arithmetic, and will often be settled
be demonstration using the inference rules of the language.
However, the general question "are there numbers" is in this
context a trivial problem.
There remains a question about the legitimacy of the language,
and this might be thought to involve the "external question"
of the existence of numbers.
This external question Carnap regards as meaningless and therefore
as irrelevant to the acceptability of the language, which is
to be determined by pragmatic considerations.

Many of the traditional questions are philosophy are resolved
by Carnap's philosophical methods by proposals about the use
of language.
Thus, the question "What is a proposition" will be resolved by
Carnap (if at all), by a proposed usage for the term proposition,
together with any theoretical claims about propositions which
follow from the proposed usage (which will be analytic).

I appreciate that I have not answered your question.

>I think we have to distinguish between philosophy’s being
>analytic, philosophy’s being a set of analytical propositions,
>and Carnap's philosophy as distinguished from other forms
>of analytical philosophy.

Carnap's position is specifically about the kind of philosophy
which he recommends, not a claim about the propositions affirmed
by other philosophers.
He does not say that philosophy is a set of analytic propositions.

>There are other forms of analytical philosophy that meets
>this test. For example I think what Austin, Ryle, Grice and others were
> doing were not much like what Carnap recommends, but they are all fine
> analytical philosophers. So the metaphilosophical question is not resolved.
> How do we resolve it? Roger suggests that we do it by analogy with how we
> would answer the same question in mathematics. He says,

I don't understand what you mean here by "the metaphilosophical question",
and I doubt that my analogy between maths and philosophy is intended
to address it.

>“The propositions of mathematics are those whose
>subject matter is properly mathematical and which have
>been proven by accepted methods. “
>Well, I think there are mathematical propositions which have yet to be
> proven, and the question of “methods” is not beyond controversy, as
> historians of Cantor will recall from the reception of the diagonal
> argument. Nor is it very encouraging to be told that mathematical
> propositions are those whose subject matter is mathematical. So I can’t
> really buy in to this as clarifying matters. Roger brings up Ayer.

It was not my aim to give in a sentence a watertight definition
of the propositions of mathematics.
Merely to point out that very many propositions truthfully uttered
by mathematicians in the course of doing mathematics are not properly
considered among "the propositions of mathematics".
This is not just for certain philosophers, there is a distinction here
which is a part of the culture of mathematics.

>I can’t help but recall Sellars’s criticism of Ayer’s use of Carnap in
> speaking of “sense-data languages.” Ayer, following Carnap, defends a form
> of logical reconstruction wherein one translates from one language to
> another. I’ve mentioned the pitfalls etc. of this, and I’m a little
> disappointed that Roger hasn’t addressed them because they are at the heart
> of the issue.

I am not defending Carnap's use of translations.

>So I have a challenge for Roger! Give me an example of a philosophical
> problem that can be solved using Carnap’s methods.

Carnap is primarily concerned with the problems of science.
So he is aiming to provide languages for use in science.
The methods he describes in "Meaning and Necessity" are for defining
the semantics of such languages.
The philosophical problem then is "how to define a language" and
the result is a description of methods, not one or more propositions.

> I mean a philosophical
> problem such as the problem of the external world, the reality of the Self,
> the nature of choice, free-will, causation. Now clearly Carnap addresses
> some of these problems, but when he does his method of intension/extension
> or reconstruction into a logical “perspicuous” language simply plays no
> role.

Does he?
If you have among them a counterexample to his assertion that
philosophical propositions are analytic then I would be happy
to discuss it.

>Well now, “propositions of philosophy,” (in quotes). All philosophical
> problems appear to be pseudo-problems! Can you give me an example of a
> philosophical problem, distinctively philosophical (to mimic Bruce) that is
> not a pseudo problem by Carnap’s lights? His conception of philosophy is an
> exercise in explicating the language of science. I don’t see any legitimate
> philosophical problems, really. Cite me an example of what YOU consider to
> be a philosophical problem! <G>

It is entirely possible that there might be a complete disjunction
between what you count as a proposition of philosophy and what
Carnap counted as one.
But that in itself would not render Carnap inconsistent in the sense
we are considering.

Let me give you 

>Some of my reasoning here is ba sed on Bergmann’s _Metaphysics of Logical
> Positivism_ There are selections on Hist-Analytic that may interest you.
> For example, why prefer realism over phenomenalism? Say, in Carnap’s case.
> He looks like a phenomenalist to me in some places and not so much in
> others. How can logic cure us of this “pseudo” issue. I don’t see it.

Well of course Carnap's position did change quite a lot over his
But he is immediately misrepresented if he is held to be denying realism,
for I don't think he ever did that, and once we get through his
"liberalisation of empiricism" he embraces "theoretical languages"
which are just the same as a realistic physicist might use.

Does Bergman argue that the points which he considers metaphysical
would be so regarded by Carnap?

>I see very little similarity between Russell and Carnap after 1927. Where do
> you see it?

I don't see much similarity between Carnap's philosophy and Russell's.
Carnap explicitly describes the way in which his own conception of
philosophy is based on Russell's and his program consists (in my words)
in doing for science what Russell had done for mathematics.
This kind of high level influence is in my opinion enormously
important, and my support for Carnap's position is at a similar
level. I applaud aspects of the general thrust of Carnap's philosophy
but am inclined to retain very little of the detail.

I have not found a nice example of a "proposition of philosophy".
perhaps one will come to me.
If you have counterexamples I would be happy to consider them.

I will mention two little points.

Firstly, it does seem to me that the claim that Hume's fork
identifies a fundamental and objective partition of some kind
is one which I would count as metaphysical.
I don't think Carnap did, and in his scheme of things, this
is just an aspect of his conceptual scheme which he can
reasonably consider to be justified by pragmatic considerations.
When stated in an appropriate language, Hume's fork is analytic.

Secondly the question whether there are necessary synthetic
truths is often considered to be a metaphysical question closely
related to metaphysics.
For Carnap its denial is analytic, and is trivially so.
In some places Carnap defines both "analytic" and "necessary" 
as "L-true", in others he defines necessity in terms of analyticity.
Probably there are other variations with similar effects,
viz the analyticity of there being no necessary synthetic truths.

The supposed refutation of Carnap by Kripke is pure equivocation.
Perhaps Kripke's conceptions of analyticity and necessity are
better than Carnap's, but they cannot serve to refute Carnap
in this matter.

I appreciate that I probably have not met your challenge.
However, technically, it is not necessary for me to do so
to sustain my support of Carnap in the matter on which I
am supporting him.
For you to refute this position, you must offer a counterexample,
for Carnap's position would be consistent if he had never
tabled a proposition of philosophy.

You talk as if this were easy, lets just take one example.


More information about the hist-analytic mailing list