[hist-analytic] A Scientist's Landscape

Roger Bishop Jones rbj at rbjones.com
Sun Jul 26 15:40:29 EDT 2009

On Friday 24 July 2009 14:30:11 Baynesr at comcast.net wrote:
>I don't recall Carnap ever saying that he was unable to
>comprehend a pseudo-problem.

Probably not.

Let me explain how I got there.

Carnap's own precis in his "intellectual autobiography"
describes his position on pseudo-problems (by which he
seems to mean metaphysics) as having evolved through the
following stages:
	  1. They are useless
	  2. They are meaningless
	  3. They lack cognitive content

Now for a first cut stage 2 corresponds to a
confession that he does not understand pseudo-problems.
If they are meaningless, how could he?
Of course, its more in-your-face than that,
his words imply that neither he nor anyone else
understands them because they are nonesensical.

In time Carnap becomes more conciliatory, but
this is not apparent in the story about pseudo

These three stages all appear very early,
he has already arrived at stage 3 in 1931, with
nearly another 40 years of further development to

Perhaps the shortfall in the account arises
because he stopped using the phrase.

However his position in relation to metaphysics 
does continue to evolve, and this is apparent in
things like "the liberalisation of empiricism"
and his "principle of tolerance".

In his "liberalisation of empiricism" Carnap
admits first physicalistic and then theoretical
languages, even though these are in radical
positivism going beyond the given into progressively
remote "metaphysics".

His principle of tolerance elucidated through the
internal/external distinction effectively commit
Carnap to accepting the most speculative metaphysics
provided only that it is presented in the right way.
(or at least reduce his opposition fall back to
a claim of uselessness).

This is how it works, in relation to any metaphysical
theory which is held to consist of necessary truths.
Carnap defines necessity in terms of analyticity, so
to persuade Carnap that your theory is necessary
you just have to define the semantics of your
metaphysical language appropriately.
If your theory is logically consistent then this
will always be possible (actually, you don't even
need to know that its consistent).
Once you have done this, the claims of the theory
become internal to the language you have defined,
and they will in that language be necessarily true.

There remains the external question of whether
the language in which you have formalised your
metaphysics is acceptable.  For Carnap, the external
question is to be resolved pragmatically, and
therefore the worst he can say about it is that it
is useless, which if you think of metaphysics in the
way in which Hardy thought of pure mathematics, won't
greatly worry you.

This may seem like a trick to you, but it seems
to me that Carnap is wrongly thought of as dogmatist,
though he undoubtedly did pass through a pretty
dogmatic phase (which he partly attributes to
Wittgenstein's influence), and that there is a
pretty good chance that he could have been persuaded
by arguments on these lines to take one step further
of liberalisation.  The step further would be not
simply to accept that his principle of tolerance
(etc,) demand a return to a pragmatic consideration
of metaphysics, but, on the basis of arguments I
have not yet produced, to accept that even speculative
metaphysics can sometimes be useful.

>If pseudo-problems were
>incomprehensible they wouldn't present much of a
>problem. The "problem" with pseudo-problems is that they
>look like real problems but in fact arise from confusions
>in ordinary language. They look this way because the
>sentences in which they are formulated are "pseudo-object
>statements" in the "material mode" of speech.Generally,
>Carnap speaks more about "pseudo-sentences" rather than
>pseudo-problems. But not always. My reason for bringing this
>up is that pseudo-sentences ARE incomprensible within
>an "syntactical language." In the material mode they
>look fine. Here I address the Carnapian view.

This does not correspond to my understanding of Carnap's
position in his syntactic phase.
The role of "semantic ascent" is not in dealing with
pseudo-problems, but in dealing with analytic or
necessary propositions, for example the propositions
of mathematics.
These propositions can be understood by translating
them from the misleading material mode into the
formal mode. Then you can see that they are necessary.
A pseudo-question is one in the material mode which
talks about pseudo-objects but cannot
be translated into the syntactic mode.

>For Carnap "pseudo-questions" (and here he does use this
>expression) arise from metaphysical statements purporting
>to be about objects that go beyond the domain of science.
>_Logical Syntax of Language_ p. 331.Compare Kant in this
>regard; they are close to this point, at least with respect to
>Kant proclamations against metaphysics. Consider the
>following example: "The world is the totality of facts not
>things." Of course, this is the sentence Wittgenstein uses
>to begin the Tractatus. This sentence occurs in the "material
>mode" of speech. It raises pseudo-problems about the
>nature or ontology of facts, etc.But the point Carnap is
>making is that by translating the pseudo-sentences that
>give rise to these pseudo-problems we state the scientific
>content without the problmatic dross. So instead of this
>sentence we have in the syntactical language: "Science is a
>system of sentences not names." (op cit. p. 303).

I think if you can do the translation then Carnap would
not regard the original as a pseudo-problem, even though
he would describe it as a statement in pseudo-object mode.

>Let's take an examplle. You remark:
>"It seems to me that your nominalistic tendencies are
>much the same kind of thing. You decline to consider
>explanations which invoke entities about whose existence
>you are sceptical. How does this differ from the
>rejection of pseudo-problems?"
>I am a nominalist; in fact a radical nominalist! The only
>entities are particulars (of the lowest type). Much of what
>passes for semantics on my view is purely a matter of
>psychology. Philosophical psychology is central. Philosophical
>problems associated with explanation etc. are also central.
>But ontologically I'm a nominalist. Carnap would consider
>my ontology, or any ontology outside the domain of
>science a breeding ground for pseudo problems; but my
>view is that the "ideal language" (syntactical language)
>approach is barren. When I ask: "What is a number?" for
>Carnap this is a pseudo-question. Not for me. For me this
>is a legit question.

I would like to understand your position better than
I can from the above description.

Presumably in your belief that your particulars are
unacceptable to Carnap you are thinking of his
phenomenalistic phase.  But I would guess that the
first stage in his liberalisation (embracing
physicalistic language) is pretty much the admission
of an ontology of particulars.

Carnap doesn't actually have a problem with numbers,
even though in his early phase he would have considered
them pseudo-objects.  Even the question "What is a number?"
would be OK provided you settle the context.
In the syntactic phase it surely is held to
translate into a question about numerals?
I don't care for this, but I think he could give
a much more satisfactory answer later on.
(e.g. in the language of set theory we have standard
definitions of numbers).

Most importantly, to have a conversation with you
about many of the topics which are common on this
list, I need to know how as a radical nominalist
you can talk about the meaning of sentences.
Is this possible?
If so, how can you have discussions about necessity?
Or indeed, how can you do any philosophy at all!

I point out to you, that you effectively duck
out of conversations as soon as I mentions meanings
or propositions, without any enquiry into what I
consider their ontological status.

How do you know that I am not a radical nominalist?
(I'm pretty close to being a fictionalist in respect
of all non-particulars).

>I think Grice is right about meaning. Meaning is the thorn
>in the side of the Carnapians, in my opinion. The wind up
>taling about worlds, and these in my opinion are pseudo
>objects if ever there were one! Grice is a pragmatist of a
>SORT. He views language neither as a mirror of nature
>nor as a medium of exchange; it is an instrument. I think
>this is right, but it leaves my metaphysics untouched.

"Possible worlds" are, in my book, just abstract objects,
which for present purposes we can treat as fictions.
So what's the problem?
(I think Carnap's principle of tolerance is pretty
much like fictionalism, it rules out the idea that
questions about the existence of abstract objects
have objective answers, but allows that we use them

>You remark:

From here on in you are commenting not on my remarks
but on your own remarks as quoted in my last response!


More information about the hist-analytic mailing list