[hist-analytic] Steve's and Roger's recent interchange

Roger Bishop Jones rbj at rbjones.com
Wed Jul 29 18:15:01 EDT 2009

On Wednesday 29 July 2009 21:00:06 steve bayne wrote:
>My dispute was over whether "Carnap was surely quite clear..." Now an
> argument may be made for this, but Roger was pretty confident about this so
> I'm hoping no argument is needed; just give me a page number, that's all.
> Bruce says,

No, I can't find one.
Mainly I think because Carnap uses "logical truth" normally
rather than analytic, though I think he was quite clear that
these two are the same thing!

More on this later.

Some points of clarification.

Firstly I said "Carnap was quite clear".
I did not say that Carnap said it was quite clear, or
that he believed that it was easy to establish whether
a sentence is analytic.
Just that it seems to me central and plain in Carnap's
philosophy that he regarded philsophical propositions
as analytic.

I do not claim that Carnap believed that every proposition
asserted by a philosopher is analytic.
He was stipulating what should properly be regarded as

Thus, if G.E.Moore makes a true claim about natural English,
Carnap does not consider him to be making an analytic claim.
He would regard it as an empirical claim, and consider it
not to be analytic philosophy (in his syntactic phase, he says
"philosophy is the application of the syntactic method",
and the application that method yields propositions
in the formal mode, which are about logical syntax,
and are if true, logically true).

When Wittgenstein makes an apparently metaphysical claim
in the Tractatus, then this will be regarded not as philosophy
but as sheer nonsense, unless it proves possible to translate
it from the material mode into the formal mode, in which
case it may be philosophical, and will be analytic if true.

>"The legitimate objects of the human understanding are exhausted by
> relations of ideas and matters of fact and existence."
>I think this gets right to the main point. This is the statement of an
> analytical philosopher, indeed a very good one, but part of the problem
> though is that the sentence begs the question of what a "legitimate object"
> is. If we presuppose as true that all there is in this world are ideas or
> matters of facts then our understanding of 'legitimate object' may be
> question begging. So let's be clear on legitimacy first, then whether this
> class is coextensive with ideas, facts or some combination. John's going to
> the store is not a fact because he didn't go to the store; his going to the
> store is not an idea, nor is John so John's going to the store is neither a
> relation of ideas nor a matter of fact; and, of course, the store may not
> exist. My point here is that while there may be a good Humean answer to the
> issues raised, there is nothing "quite clear" about how we ought to address
> them.

This is Hume not Carnap.
Carnap is defining a language in which we get a classification
of propositions, and if he gets the details right it will be possible
to prove that there is the stated exhaustive dichotomy.
This can be done formally, though we can question whether Carnap
got it right.

His syntactic phase was inspired by the techniques for arithmetisation
of metatheory in Godel's papers, which made him think he could provide
a method for translating philosophical propositions into arithmetic
(i.e. logic, for a logicist).
Carnap's position is coherent.
He is putting forward a conceptual scheme and a method,
and in these activities he does not offer philosophical claims.
But when he does make claims then he believes these to be
logically true and provable in the appropriate linguistic

>When Wittgenstein remarks 'The world divides into facts not things' what is
> he denying? Certainly not a particular relation among facts, nor ideas,
> unless things are ideas. Again, we have a problem circumnavigating the
> conceptual terrain. In addition, there are other problems. E.g. I am
> obliged to keep my promises. Right? Suppose this is right. Is this a fact;

I don't think it is "factual" for either Hume or Carnap.
Though to say it is a fact might mean simply that it is true,
and I don't know that they would deny that.
(we use this term in relation to mathematics, "2+2=4 is a fact",
but in our context, that of Hume and Carnap, we are using "fact"
specifically to mean "empirical fact", we have a narrow usage
relative to which neither mathematical nor ethical propositions
are factual.) 

> well what then of the is/ought distinction? Again, it may be depend on what
> we take to be a "fact" but isn't that really part of our dilemma, and one
> for all Humeans as a well?

As far as the present issue is concerned, evaluative propositions
are not considered by Carnap to be part of philosophy.
There might be some analytic propositions in ethics, and these
might belong to analytic philosophy but then they would not
be genuinely evaluative.

>Recall that Carnap maintains that all topological properties of space and
> time can be dealt with using purely non-metrical methods. Suppose this is
> true. Is it "analytic"? Suppose it is not. Is it then a fact. Well yes, but
> it is about theories and theories are relations of ideas. So are the
> classes of fact and ideas (and their relations) such as to have a null
> overlap? This is another problem. If so is this a matter of fact or a
> relation of ideas. There are two big problems.
>First, there is no argument for believing the class of relations of ideas
> and matters of fact are exhaustive; if not, then some analyses may not
> involve relations of ideas etc. Second, if the classes are so broad then
> there is always, available, the useless "shuffle" of saying "Oh that's a
> relation of ideas," or "Oh that is a matter of fact.

Yes there are such arguments.
Actually they are very simple arguments.

I don't understand your shuffle.

>There is another question: are all truths acquired by analysis analytic.

The ones obtained by certain specific methods of analysis are.
Among these methods are the method of logical syntax.

> Mightn't their be theories that follow upon some analysis that might have
> been arrived at some other way? Well, what would exclude this.

The fact that Carnap is speaking of specific kinds of analysis,
and makes no claim about analysis in general.

> Until
> Carnap's notion of analysis is made clear and I get at least a couple of
> page references, I'm afraid I'm as skeptical as ever that all statments of
> analytical philosophy are analytic.

The central purpose of Carnap's "Philosophy and Logical Syntax"
(which I have read) and I guess also LSL (which I have not)
is to present "the method of logical syntax".
This is probably the most precise and complete a description of
a method of philosophical analysis as has ever been given.
In this method, the truth of all philosophical propositions
is to be achieved either by logical demonstration (if they
are already in formal mode) or by translation into formal
mode so that they can be so demonstrated.
Hence, any philosophical proposition which can be shown
to be true by Carnap's method must be a logical truth
(and hence, in Carnap's terminology, analytic).

Carnap does not consider a couple of pages enough to properly
articulate his method.
Will you settle for a couple of books?


More information about the hist-analytic mailing list