[hist-analytic] RBJ's Proposal and and Hume's Fork

Roger Bishop Jones rbj at rbjones.com
Sat Feb 28 15:52:44 EST 2009

On Wednesday 25 February 2009 18:46:45 Bruce Aune wrote:

> RBJ proposed that a sentence appropriately disambiguated should be
> said to be analytic iff it expresses a necessary proposition, the
> latter being a proposition that is true is every possible world.  I
> think this is unpromising for the following reasons:

The principle issue at stake is whether or not analyticity and
necessity co-extensional concepts.

I don't believe that the Logical Positivists argued the case
for this, it was just obvious.
If you think the relationship is obvious then to suggest,
as my definition does, that analyticity is the one to use
for sentences, and necessity for propositions is reasonable,
and my definition is then reasonable.

This is not so good in a context in which most people think
that they are obviously not the same.
It is still a tenable position, provided that this definition
is offered in the context of some argument to the effect
that the identity holds when some less question begging
definition is used, such as "true in virtue of meaning".

It is the case that I believe my definition is consistent
with the view that "analytic" means "true in virtue of meaning".
This is because I believe that:

   true in virtue of meaning

means the same as:

   expresses a proposition which is true in every possible world

This seems to me sufficiently obvious that to find a convincing
demonstration is hard.
It was this which motivate me to come up with the mathematical
model which I referred to earlier.
In that exercise I provided a lightweight model of the semantics
of descriptive languages, and then defined necessity using that
However, when I came to define analyticity, the result was too
obviously the same concept that no proof was necessary.

The only relevant part of the meaning of a sentence for
determination of either analyticity or necessity is the
truth conditions.
We know that a statement is analytic when those conditions
tell us that the under all conditions the sentence is true.
But for a language which talks about the world (as opposed
to some abstract domain for example) then the "conditions"
in question are "possible worlds" and being true under
all conditions just means being true in all possible worlds.

I can see, that this "obvious fact" will have to consume
a much larger space than this in my monograph, for though
once obvious to many, it is now obvious to few.

If for a second we assume that this argument is good
and that analyticity is the same as necessity, there
still remains a problem in identifying the fallacy in
Kripke's argument to the contrary.

I am no scholar of Kripke, so my suggestion on this score
must be tentative.
From my recollection, Kripke, perhaps thinking it obvious,
does not really offer an argument in favour of the denial.
He openly assumes the existence of rigid designators,
which are defined as phrases which designate the same thing
in every possible world, and observed that an
identity between rigid designators must be necessary.
Then somehow we get the denial that these identities
are analytic, and I don't recall exactly how we get this.
Is this part of the definition of a rigid designator,
or is it a second assumption, or just a bald uncorroborated
claim, or something he offers an argument for??

If my argument above is sound, and the non-analyticity
is somehow incorporated into the definition of rigid
designator, then the definition is incoherent and fails
to define anything.
If it is a second assumption, then Kripke is making
contradictory assumptions.
If it is a bald assertion then Kripke's argument is

Does he offer an argument?

A counter argument is that if an identity of any
kind is necessary, then it is true in every possible
world and this information is part of the truth
conditions and hence part of the meaning of the
language and the identity will also be analytic.

It looks to me like Kripke is begging the question.

My remaining comments therefore largely point out
the obvious consquences of my position above for
your points.

> 1.     1)  As Kripke pointed out, ...

I think this is now covered.

> 2.   2)  Many propositions claimed to be synthetic a priori truths by
> epistemological rationalists are generally acknowledged to be
> necessary, but anyone who thinks they are really analytic would
> generally be taken to have serious work to do.  One such proposition
> is expressed by “Nothing determinately blue on some region also has
> some other color there.”  I argue in my recent book that this should
> be considered analytic, but there is nothing trivial about the case I
> make for this claim.  I am convinced that I am right, but most
> rationalists would not share my conviction.

The "serious work" is as above, perhaps expanded somewhat.

It seems to me that propositions like the one you mention
are problematic because of doubt about what they mean,
not because of doubt about the concepts of analyticity
or necessity.

I would say that the proposition you quote is false
since something determinately blue on some region might
also be azure there.
We probably don't understand the sentence in the same way!

> 3.    3)   Useful conceptions of analytic truth purport to explain
> why analytic truths that are necessary have this further property.
> The statements (or “judgment”) covered by Kant’s conception give some
> indication of this.  If a predicate concept is contained in a subject
> concept in an affirmative way, anything in any world falling under
> the subject concept would be guaranteed to fall under the predicate
> concept because the latter is just one of the concepts it falls under
> if the subject is applicable to it.  This is why the statement is
> true in (or at) any possible world.

This is nugatory in the context of a convincing demonstration
that the concepts are coextensive.

> 4.   4)  Hume’s epistemic fork was the doctrine that all truths
> concern either mere relations of ideas or matters of fact and
> existence.  The former are considered analytic by empiricists: their
> truth can be ascertained by “mere analysis” and does not, as Hume
> said, depend on anything that is anywhere existent in the universe
> (except the relevant ideas).  Matters of fact and existence are,
> empiricists emphasize, synthetic truths that can be known only by
> observation, memory, and “experimental” inference.  A conception of
> analytic truth can be considered plausible only if makes clear the
> kind of analysis that can plausibly show that a given analytic
> statement is indeed true and, if necessary as well, why it has this
> additional property.  I cannot see that the conception RBJ intends to
> develop has this plausibility.

I don't think my proposal makes any difference in this area,
though this is something I intend to say something about in
the final section of my proposed monograph.

Can you poke some holes in my arguments above, or fill in
the hole in my memory of Kripke's argument?

Roger Jones

More information about the hist-analytic mailing list