speranza and grice on analyticity

RBJ

Here are some responses to JLS followed by some more from on
the same topic.
.
> --- Lovely. We cherish REG's pun there:
> dog-dogma-underdogma.
.
Who's pun?
I thought it was yours.
.
> (Incidentally, perhaps Hume's terminology is, more than
> from Aristotle, who Grice worshipped -- in the figure of
> "Kantotle" -- from Locke direct, whom Jones has edited.
> For Locke, analytic propositions are 'trifle'.
.
I think myself that Hume's account is in some ways more
suggestive of Plato, especially when we see his scepticism
about matters of fact.
.
Philosophy in those days was still breaking away from Plato
and Aristotle, and the departure perhaps began with
translations of Sextus Empiricus commissioned by Savonarola.
This gave ammunition for sceptical arguments (initially
against papal (over?)dogma) which unfortunately had to be
mitigated to make room for the new science.
.
The first major philosophical (rather than ecclesiastical)
mitigation (if we discount less famous figures such as
Gassendi and Mesennes) was from Descartes, who used
sceptical arguments to sweep all aside before erecting his
own dogmatic rationalistic philosophy.
.
Locke may perhaps be seen as reacting against the
rationalistic side of Descartes, re-asserting the empiricist
elements of Aristotle.
But in doing so he did not come up with a credible account
of the fork, he leaned to far in the direction opposite to
Descartes, allowing to small a role to the a priori
(trivia).
.
Hume takes matter to their logical conclusion, and has a
much sharper conception of the divide.
.
First of all, he broadens the realm of the a priori,
relative to Locke, while, like Locke, denying it to any
matter of fact.
Secondly, in the sphere of the a posteriori, the matters of
fact of which our knowledge depends upon sensory experience,
he goes into just exactly one can infer from this sensory
evidence, which turns out (if implicitly we are asking the
question : what is logically entailed by the evidence) to be
almost nothing.
.
Locke does not look to me (though I am substantially
ignorant) as a man influenced by scepticism to break with
Aristotle, but as one impressed by Newton's science to
endorse empirical methods rather than Descartes rationalism.
.
Hume however, is taking seriously the influence of sceptical
thought, even though he too is under the spell of Newton,
and models his philosophy on the methods of empirical
science.
Hume's moderation of scepticism consists in cutting of the
scepticism at the line between relations between ideas
(which he considers to consist of certain truths) and
matters of fact (which is a very thin and uncertain soup).
His conclusions about matters of fact are very
controversial, but his drawing of the line which is his fork
(the first one) is much better than his predecessors, and the
importance he attaches to it is significant.
.
> -- For what about a trifle proposition like
>
> 2 + 2 = 4
.
This is Speranza talking about Locke.
But I had the impression that trifiling propositions for Lock
were even more trifling than that.
He seems to require that a proposition is trifling only if
the the subject appears explicitly as a part of the
definition of the predicate.
.
By contrast with Hume, whose "relations between ideas" does
encompass the whole of mathematics (he says).
.
Did you get this example from Locke?
.
> ... to use an example that fascinated Grice:
>
> "Nothing can be green and red all over". -- Synthetic a
> priori.
.
Interesting...
Is that what Grice says it is?
I have a colour problem to dscuss later.
.
> Jones:
>
> "It seems to me that some philosophers have, and others
> do not, a good intuitive grasp of when a proposition
> says something about the world and when it does not.
> Hume had, but others do not."
>
Speranza
> I'm always fascinated by Mill. To me, Mill just followed
> the empiricist -- and why not, positivist? -- position
> to the extreme. For he thought that
>
> 2 + 2 = 4
>
> rather than a formalism alla Hilbert that merely reflects
> a play on words, is a statistical empirical
> generalisation. Analytic, mabbe, but surely not a
> gratuitous statement as most philosophers think it is.
.
Though I have not read him, I have the impression that Mill
takes the a priori to be just as small as Locke, possibly
even smaller.
So he is a philosopher who, in the terms I stated above,
doesn't get it, or doesn't get it "right", by a substantial
margin.
.
This is an illustration that a positivist is not simply an
extreme empiricist.
Hume and Carnap (not sure about the positivists in-between)
had a more generous conception of the a priori than
empiricists like Lock and Mill, and positivists such as
Hume, possibly Carnap take a less generous view of what we
can know a posteriori.
(a positivistic science is expected not to go beyond the
experimental data, except to the very limited extent of
deductive inference, but an empiricist is more likely to
accept wide ranging "inductive" inferences).
.
> Excellent. "Positivism" has a very long pedigree to it.
> "Rationalism" has undergone so many changes that in
> "Logic and Conversation" (WoW:ii) Grice feels the need
> to grant that he is
>
> "enough of a rationalist"
>
> to want to say that ... people are rational.
.
I'm not sure that I would regard that as rationalist in the
usual philosophical sense of the term.
.
> Indeed. He was talking about too many things! I am always
> fascinated by the fact that in his big oeuvre, he found
> the time to write a "History of England". (I'm more
> fascinated by the fact that he was a Scot, and that his
> surname was spelled 'Home').
.
Apparently both the British Library and the Cambridge
University Library still list Hume as a "Historian" as which
he was best known in his day.
.
Speranza on Grice:
> His idea of 'demonstration' is perhaps best made evident
> in his third book, "Aspects of Reason". Here he
> considers 'trivial', which may relate to Locke 'trifle'.
> Grice wants to say that some reasoning is 'trivial' when
> it is not even (or almost) reasoning.
>
> "I have "1 + 1" hands; therefore, I have '2' hands".
>
> Say. He wants to say that in some 'demonstrations', which
> are, granted, analytic, nothing _is_ demonstrated. He
> makes the point that he will STILL consider such
> 'demonstrations' demonstrations, and he appeals to
> 'implicature'. It is a mere implicature that has us
> think that trivial demonstrations are no demonstrations.
.
I don't understand the source of this problem.
Who is it that has cast doubt on whether trivial
demonstrations are indeed demonstrations?
.
> In unpublications, he considers
>
> "Nothing can be green and red all over"
>
> -- and a few other instances of 'synthetic a priori'.
.
What does he say about why he thinks this synthetic?
.
As far as the demonstration is concerned, Carnap would have
something very close to this as a meaning postulate
(perhaps, "red is not green"), and so it would in effect not
be a demonstration in Aristotle's terms.
Aristotle's usage is reflected in Hume when he talks of
relations between ideas as "intuitively of demonstratively
certain" since for something to be demonstrative it has be
shown by derivation (syllogistic) from premises which are
themselves intuitively certain (necessary, essential truths)
but those premises are not counted as demonstrative by
Aristotle.
By contrast in moden logic the axioms are counted as
theorems, and have a one-line proof which is just the axiom.
.
> This is indeed a special concern of his metaphysics,
> because even for an ordinary-language philosopher like
> him, it is not easy to grasp where such 'demonstrations'
> or premises derive their validity. He was interested,
> informally, in the ontogenesis, as it were, of such
> claims. When do children start to think that "nothing
> can be green or red all over" states an a priori
> analytic truth, as it doesn't?
.
Or indeed, how could they?
.
> And if it expresses an
> a-posteriori synthetic truth, what does it mean to say
> that it IS 'a posteriori'.
.
That's tricky, but I have an example to come.
.
Returning now to the matter of the Grice defence of
analyticity, I have, in the course of thinking about the
history of the analytic/synthetic dichtomy give
consideration to how far back deductive reasoning goes.
.
Note first the semantic characterisation of deductive reason.
Reasoning is deductive if its conclusions are entailed by
its premises, and A entails B iff "A implies B" is analytic.
So there would be a way here of arguing that someone grasps
the notion of analyticity even if he doesn't have the word.
If he can tell sound from unsound inferences, then he
understands analyticity in all but the word.
.
It is generally thought that the Greeks were the first to do
deduction, but what this really means is that the first
mathematical texts in which mathematical results (e.g.
algebraic equations) are accompanied by proofs are
from 6th century BC Ionian philosophers.
.
However, it is easy to argue that the ability to grasp
certain elementary inferences is an essential part of
understanding a language.
e.g. one does not understand the English language if he does
not know that "azure" is a shade of "blue" and if one cannot
therefore infer from "the wall tiles are azure" to the
cruder "the wall tiles are blue".
.
Does this amount to a competence in the analytic/synthetic
distinction?
.
I think not.
I am inclined to doubt that common sense has any clear
distinction between what follows of logical necessity from
what follows of any other kind of necessity, i.e. what
follows on the implicit assumption of any amount of factual
background.
.
So I suspect, that though you could obtain agreement from
the philosophically naive on the truth of a large number of
analytic propositions, but not actually on their analyticity
.
Is there a better way of testing comprehension of analyticty
than expecting a grasp of whether an argument is deductively
sound?
.
Probably there is.
Talk about meanings.
Does "azure" mean a shade of blue?
This takes us right back to Plato and Aristotle who
recognised that the starting point in demonstrative proof is
definitions.
Shame that so many words (usually including colour words)
don't actually have definitions.
Here Carnap's meaning postulates might come to the rescue.
.
So I am vacillating here between believing the
Grice/Strawson defence of analyticity from ordinary language
and common sense, and doubting it.
.
Carnap did not attempt to make the case for analyticity or
semantics natural languages being determinate or definable,
and this is probably why in two dogmas Quine is forceful in
asserting that he believes the problem of giving a definite
semantics is just as intractable for formal as for natural
languages.
.
This leads me into a final observation in relation to those
colour words and the question of the analyticity of colour
exclusion statements.
Whether they are analytic or synthetic depends upon their
meaning, and these are cases where its hard to disentagle
sense from reference.
Does a colour word mean that particular colour or does it
just happen to refer to it?
In the former case the exclusion will probably be analytic,
in the latter, synthetic.
.
It may not be the same in all cases.
That red and green really do mean different colours is
plausible, but possibly red and aquamarine don't.
Does "aquamarine" mean a specific colour which is a shade of
blue, or does it mean the colour of sea water, which
contingently is a shade of blue?
.
Though I hate Quine's indeterminacy arguments, because I find
his story about the way field linguists work highly
unbelievable, without argument i accept indeterminacy of
translation, because it just is hard to translate between
language, and even between particular persons idioms (e.g.
Frege's sinn and bedeutung).
And also indeterminacy of meaning in natural languages, it
seems to me that there is no fact of the matter in many
cases about exactly what words mean (which is one reason why
science, or any advance in knowledge, evolves new language).
.
Having lightly debated the matter, I conclude of the Grice
Strawson defence, that it does support a not very precise
notion of analyticity, and is a worthwhile response to
Quine, but that it falls short of carrying the case
sufficiently well for the defence of analyticity as a
philosophical concept of great importance and utility (and
of course Quine later fell back from outright repudiation to
asserting philosophical inutility of the dichotomy)
.
For this I think one also does need a more direct response
to the detailed sceptical arguments of Quine.
.
Roger Bishop Jones