[hist-analytic] Davidson's Hume

Roger Bishop Jones rbj at rbjones.com
Thu Jun 11 15:41:29 EDT 2009

On Wednesday 03 June 2009 14:23:45 steve bayne wrote:

>We have to distinguish at least three things.
>1. Attributive uses of definite descriptions and referential uses.
> (Donnellan)
>2. Rigid and non-rigid designation (Kripke)
>3. Purely referential vs non purely referential designators.
>Each has a history and some would argue, incorrectly in my opinion, that
> there was overlap.

Thanks for pointing that out.
I have to confess that I don't really understand any of these
distinctions. (groping in the dark, as usual)

A brief search for enlightenment on Donellan left me doubting
that the distinction in use which he is discussing reflects
any underlying difference in meaning or sense.

On Kripke, so far as I understand him, his separation of
necessity from analyticity depends on it being coherent
to suppose that something can be a rigid designator without
that being a consequence of its meaning.

So that brings me to wonder about what is now meant by
speaking of something as purely referential.
Presumably it means that the designator has no sense
apart from its reference, but I don't know whether the
case that the meaning *is* the denotation counts as
purely referential or whether that term is reserved for
the case that the reference is contingent and the designator

>In "Reference and Modality" Quine notes that while '9>7' is a necessary
> truth 'The number of planets > 7' is not a necessary truth because the
> number of planets is a contingent fact. The source of the problem is that
> 'The number of planets' is not purely referential. We find something
> similar in Russell's "logically proper names." Donnellan's distinction may
> show itself within a single world and says nothing about other possible
> worlds, while (as you know) rigidity is very much about "worlds." These
> distinctions do not require much of anything with respect to what we know.
> It is a semantical not an epistemological conception.

If meaning is used in the same way as Carnap and I use it,
then, if the meaning of a designator is the thing
which it designates, that designator will be rigid.
Furthermore in relation to this kind of semantics, (which
we might call a full truth conditional semantics), this is
the only way in which you can get a rigid designator, and
results in the disappearance of Kripke's counterexamples
to the analytic/necessary identification.

Whether we are talking about different uses in one world
or uses in different worlds, the touchstone is the semantics,
and this ought to result in there being connections between
these issues (if indeed they are anything to do with meaning,
as one might doubt for the Donellan distinction).

>In 'The cause of e caused e' IF we take ' the cause of e' attributively'
> then my claim is that the sentence 'The cause of e caused e' is contingent.

Perhaps you could expand for me on the "attributive" use, and how it
yields this result?

Having just read Strawson I think he would say, and I would agree,
that e having a unique cause is presupposed here and that the
sentence has the status of being true whenever it has a truth
value, but in many possible worlds lacking one.
Though one might very reasonably insist that in those cases it
is false.

> On the other hand if we think of the sentence as completely devoid of
> pragmatic elements, and here I have in mind Jerry Katz's notion of
> 'linguistic meaning' then the sentence is trivial, like 'I married my
> wife'.

Can't say I like the idea that its trivial.

> So it's not so much which is the right reading but what you get on
> different readings.

But we have been here before and this looks to me like a non-sequitur.
How have you come to the conclusion that there is no basis for
considering one reading correct and the other mistaken?

I did offer some reasoning for not taking descriptions
as "purely referential" but you haven't responded to them.

Roger Jones

More information about the hist-analytic mailing list