Donnellan was a genius, as everybody knows.
His referential/attributive distinction dwells on 'the'. Grice, just to be original, would rather be seen dead than using Donnellan's terminology, so he preferred 'identificatory/non-identificatory'.
It all has to do with the logical form of 'the' sentences.
R. B. Jones notes:
"I see that there was an error in my remark about iota in Church's Simple Type Theory. Without choice iota is definite description, when choice is added you get indefinite description."
I see. For the record, we should also explore what Donnellan (never mention Grice -- that's FOR GRANTED) did say about "There is one King of France who is bald", or "Some king of France is bald", or "A king of France is bald".
Granted, 'king' INVITES the definite article, 'the'.
But perhaps it shouldn't!
We do speak of the seven kings of ancient Rome.
And it is not impossible to think of a dual monarchy?
Kripke says it would, because 'monarchy' necessarily means "one", i.e. 'the'.
But as Russell says, he is going to dedicate one chapter to 'the' singular, and the other chapter to 'the' plural. So surely 'the kings' makes sense, even if
"The kings of France are bald"
has an odd ring to it, if that's the expression.
In general, pragmaticists seem to consider indefinite descriptions as misleading. Never misleading to Grice!
A: Where does Smith live?
B: Somewhere in the South of France.
He seems to be seeing 'somewhere in the South of France' as some sort of 'indefinite description'. Or as he'd prefer, under-informative. The implicature, Grice says -- the example is in "Logic and Conversation" is that to be more specific (or definite -- as in "he lives in the village that Brigitte Bardot made famous for sun tanning in the nude") would hardly be cooperative. He is presenting a context where the only point of A's question is whether Smith lives far enough TO AVOID PAYING HIM A VISIT -- or something.
Note that 'the' in "The South of France" is definite enough -- although perhaps not for the French!
Similarly, in America, "The Hamptons" has caused a lot of problems. Southampton and Bridgehampton and East Hampton qualifies as the Hamptons, but Hampton Bays doesn't!
A: Where does Martha Steward spends her summers?
B: Somewhere in THE Hamptons.
(Not exactly in East Hampton, but the outskirts of East Hampton, which still should qualify as "the" Hamptons).
I suppose "the Midlands" in England triggers the same implicatures, but admittedly, "some Midlands" sounds rude.
"A midland" would be the equivalent indefinite description to the definite description "the midland".
But one problem is that "the midland" is hardly used in the singular!
Jones goes on:
"So if you do tweak to have separate symbols, this is good because iota remains definite description instead of being commuted by choice into indefinite, and of course they are both useful so you do want both."
Yes. Unfortunately, Grice left Oxford because he thought he would find expertise in logic AND LINGUISTICS (he adds in "Prejudices and Predilections") in the USA, and he did. But linguists seldom use the vocabulary (keyword) "definite description" -- never mind "indefinite description". I have seen loads written on DEFINITENESS, though, which is of course, the same idea that 'the' and 'a' are important features in English -- I wouldn't say 'language', because I can imagine a language where such distinction does not exist.
In fact, in field work by anthropologist Elinor Ochs, she thought she had found a population that did not follow Grice's maxims of definiteness.
"A sister did it."
"MY sister did it"
was found to be too informative and intrusive. I'm not surprised if they uttered
"Some female did it"
even when the utterer KNEW it has his own sister. Ochs found this as evidence against the universality of Grice's conversational implicature, but Grice never responded, because he thought that conversational reason, while universal, can display a cunning or two!
(A 'cunning' of reason is Hegel's idea that Kant's universal reason may display an exception here or there, because history is contingent and Kant (and Grice) are talking in an abstract way, rather).