Sharpless quotes from Morris:
"Another central issue in contemporary philosophy is whether a sharp
distinction can be made between analytic and synthetic sentences (or
"propositions"). In the present context this is the question of whether a
sharp distinction can be made between formative and lexical discourse."
One may view Morris as considering Carnap's attitude towards "meaning
postulates", i.e. in a formal discourse such postulates can be introduced
freely. On second reading, perhaps Morris has another thing in mind when he
mentions "formative" vs "lexical". From what follows it seems he means
"grammatical" vs "lexical". Thus a syntactic sequence would be "formative",
but any replacement of the formative sequence by a given set of "lexemes"
would be "lexical". The question is, is there a purely "formative"
discourse? I guess it should look as very abstract...
"I have suggested elsewhere that the distinction can be made only in terms =
of
pragmatic considerations - and not in terms o£ semantics or syntactics al=
one.
This seems to be implicitly involved in Carnap's introduction o£ "meaning
postulates" in his defense o£ the distinction o£ the analytic and the
synthetic. To decide whether the sentence
1. All crows are black.
is analytic or synthetic involves reference to the sign structure
(and hence to the dispositions to respond) of a specific
interpreter (or a group of interpreters)."
To a truthful utterance of
2. That's a white crow, I tell you, as I seen in the wood,
and she was eating YOUR seeds...
"If the interpreter is disposed at a certain time to respond
to all denotata of the term 'crow' by the term 'black' (i.e., if he would
not call anything a crow unless it were black), then the sentence is
analytic at that time; otherwise it is not. The criterion is thus pragmatic
and involves the use of signs (i.e., the acceptance of a sign framework) by
a specific producer of the signs. 'Acceptance' is a basic term, in
pragmatics."
What Grice would say about this is (still) a different issue -- worth examining.
Subscribe to:
Post Comments (Atom)
Posts
This seems very strange to me.
ReplyDeleteFor Carnap the analytic/synthetic distinction is a semantic distinction, and meanings postulate play a role specifically in the definition of the semantics of a language.
If "meanings postulates" belonged to pragmatics (and hence were not properly called "meanings" postulates) then they could not feature in the definition of a semantic concept, and if used to define analyticity would define analyticity as a pragmatic rather than a a semantic concept.
This would not address any supposed difficulty with the concept of analyticity (the one which Quine attacked).
RBJ
Good point.
ReplyDeleteI recall when I was having this section on the analytic/synthetic distinction for a memorial to Grice elsewhere (my "The Way of Conversation" referred in pdf in "Carnap & Grice", pdf. -- references --) I had to actually quote from
D. E. Cooper,
who distinguished between
"pragmaticism" and 'pragmatics'.
For I saw that Grice's later approach to the analytic-syntehetic distinction (and thus to the 'dogma') had attained a 'pragmatist' (NOT 'pragmatic') bent. In Grice's view, as with Carnap, the 'choice' to HAVE 'analyticity' belongs to the realm of practical decision -- reached by a consideration of what 'use' that distinction would play.
Sharpless however indeed quotes from Morris -- in his rather obscure paper -- as suggesting that the 'analytic-synthetic' distinction wouuld be 'pragmatic' per se, which seems to avoid, as R. B. Jones says, the very problem that Quine signalled.
----
There IS possibly a roundabout way in which the analytic-synthetic distinction COULD be said to be 'pragmatic', but only if first we define the 'semantic' in term of the 'semantic'.
Indeed, the 'defense of the dogma' by Grice is said to be a way out of the vicious circle because it DOES examine 'analyticity' as being guided by a criterion which is 'behavioural': if an utterer cannot assent (in any situation) to a proposition, -- e.g. p & -p, -- THAT is what makes 'p & -p' contradictory. Its contradictory a tautology, and thus analytic. But there are some problems there as well, as there should.
For Carnap the decision to adopt a language (whose definition would normally include the semantics) is pragmatic, but once adopted the content of the semantics and any consequences of the semantics, do not belong to pragmatics.
ReplyDeleteThe same consideration apply to the definition of analyticity. The adoption of a definition of analyticity may be done on a pragmatic basis but this does not mean that the resulting notion of analyticity is pragmatic. If the definition of the concept of analyticity is in terms of semantics (which it normally would be) the resulting concept is semantic, and claims about analyticity (at least in pure rather than descriptive semantics) would then themselves be analytic.
Morris fails to understand the distinction and seems to think that Carnap's pragmatic criteria for language adoption result in the assimilation of semantics into pragmatics.
RBJ
Right, and I'd expand on Grice's way out of the circle.
ReplyDeleteAs Bennett notes in his "Linguistic Behaviour" (Cambridge UP, 1973) -- Bennett has some dates wrong. But he hypothesises:
---- Quine published "Two dogmas of empiricism"
---- Grice and Strawson published "In defense of a dogma" in 1956.
---- The following year, 1957, Grice published "Meaning".
Bennett suggests, wrongly, that Grice's "Meaning" paper is a sequel to "Defense of a dogma". Pity Grice had written that in 1948!
In any case, if we use Grice/Strawson's example of analytically false sentence (and thus of 'analytic sentence' for I dislike the idea of equating 'the analytic' with the 'analytically true'):
we have:
i. This child is an adult
("My neighbour's three-year-old child is an adult").
Nobody would ASSERT that. In Grice's and Strawson's hypothetical situation, one DOES assert it, and the interlocutor is logically puzzled. "Different linguistic frameworks?". "Went round the bend?", "What the heck?"
But we don't need to get so drastic. For Grice and Strawson are suggesting:
---- for sentences which are NOT analytic ("My child is a genius") you can assert them -- and you can assent to them, or not.
---- for sentences which are NEVER 'analytically true' but are always 'analytically false': "This circle is square") nobody can assert them.
----- for sentences which are ALWAYS 'analytically' true ("Women are women", "War is war") everybody asserts them, on request.
IN FACT, the behavioural criteria are SIMPLER, as expressed by Grice and Strawson.
"If your interlocutor says, "I can't get what you mean", then what you meant was analytically false ("Caesar is a prime number", to use Carnap's example -- a controversial one).
"If your interlocutor, RATHER, says, "I can't believe you, even if I DO understand you, the sentence is NOT analytic".
This possibly begs as to "Women are women" in that people do understand it, and do believe it.
But in any case, we at least have a criterion for 'analytically false' sentences, and so once we get one, we just add "not" and we get an analytically true sentence.
"My neighbour's three-year-old child is NOT an adult".