Sunday, November 16, 2008

Is King's account too fined grained?

Is King's account too fined grained?

King considers whether his view individuates propositions too finely across languages. He wants us to think about the following three possibilities:

(A) At least some proposition(s) can be expressed in different natural languages.
(B) At least some proposition(s) expressed in one natural language can be expressed in any natural language.
(C) All propositions that can be expressed in one natural language can be expressed in any other.

Which do we accept?

1. If King's view is correct, then a proposition Q is expressible in different natural languages L and L' iff

(i) they contain sentences SL and SL1, whose syntactic structures at the relevant level of syntax are identical.

(ii) the semantic significance of these syntactic relations are the same.

(iii) the semantic values of the lexical items occurring in the same places in the syntactic structures associated with SL and SL' are identical.

2. King's view is correct.
3. So (A) entails a substantial empirical claim about the syntactic of the languages in question.
4. So (B) entails all languages to have sentences that are syntactically identical.
5. So (C) entails all languages to be structurally identical at the level of LF.

It seems we should conclude that (B) and (C) entail successively stronger and implausible claims about the LFs of natural language.

6. Thus, the entailments from (A) and (B) and (C) are too strong.
7. Because King's view makes them too strong, King's view is false.

This would, of course, make King unhappy. To see our way past this objection, we need to see that (A) differs from both (B) and (C) in its pretheoretical plausibility.

(A) is a sort of constraint on any theory of propositions. It is desirable for any theory of propositions to yield the result that 'Scnhee ist weiss' and 'Snow is white' express the same proposition.

1. If (A) is true then 'Scnhee ist weiss' and 'Snow is white' express the same proposition.
2. They do.
3. So (A) is true.

(B) and (C) contrast strikingly with (A). They are lacking pretheoretical plausibility.

Perhaps a more natural pretheoretical reading would be

(B') Some sentences of some language can be translated into any other language
(C') Any sentence of any language can be translated into any other language

Sadly, these rewrites only suppose the originals given a further assumption. We need to add that translation is pairing sentences which express the same proposition. So we need to ask: What is the evidence for (B') and (C')? How plausible is the additional premise?

Sometimes we do make strict translations, like in the case of 'Scnhee ist weiss' and 'Snow is white'. But usually in practice we make loose translation. We make what are closer to paraphrases.

It seems that the looser sense of translation is used to give a push for the pretheoretical plausibility of (B') and (C'). But this doesn't seem to help us bolster the case for (B) and (C), since they seem to rest upon the strict sense of translation.

To our questions we must answer: The evidence for (B') and (C') is the loose sense of translations. The additional premise is plausible given the strict sense of translation.

King thinks that it is a virtue of his theory that if we want to figure out whether (B) or (C) is true, we only need to look at the empirical evidence about language and do some theorizing about propositions. This lets King block the objection (6) above. (A) seems to be common-sense, and we should accept it. Whether (B) or (C) is true or not is an open question, perhaps.


Dan said...

I happen to think that King is attacking a straw man here. We don't need (B) or (C) to make the point that his theory is too fine grained. All we need is:
(D) There are two sentences (maybe the same language, maybe not) that violate (i), (ii), or (iii) but nevertheless express the same proposition
It's worthwhile to note that this objection (in any form I've seen) is not a formal one, but one that draws upon our intuitions about how to individuate propositions. This is why he trades on the plausibility of (A), (B) and (C). So if we can bolster our intuitions about two sentences that violate (i) (ii) or (iii)to the effect that they express the same proposition the objection goes through. So consider:
(P) Mary swims (in english)
(NP) Mary doesn't swim (in Nenglish)
These differ in syntactic structure, and so violate (i). In English the syntactic structure gives different instructions than in Nenglish, thus (ii) is also violated.
Intuitively (P) and (NP) express the same proposition. But to bolster the point a bit, they have in common two properties that are usually used to individuate propositions:
(1) truth conditions
(2) aboutness facts
That is, they are true in the same situations, and they are about the same thing. Unless King offers a good independant reason to include (i) and (ii) as individuation criteria, it's hard to see why we should accept it. His answer can't be that it fits the data, because clearly it doesn't.
I think this deteur through (A) (B) and (C) is sort of a red Herring. His real opponent is the individuation criteria (1) and (2).

Adam said...

I think it's important to note that King seems happy to hold that two distinct propositions may have exactly the same truth conditions, and also that two distinct propositions can be about the very same fact.
This seems intuitive to me, anyways. The proposition that 2+2=4 and the proposition that Dan is identical to Dan seem likely to be true in exactly the same (possible) contexts. But we should think that they are distinct. And the proposition that Wes is tall and the proposition that Wes is 6+ feet in height are arguably distinct, yet they plausibly represent (are about) the same non-propositional fact- the fact consisting of Wes instantiating the property of being some determinate height.

Wes McPherson said...

I understand Adam, but am not so sure I understand Dan.

It seems to me that (P) and (NP) express different propositions. One has some negation going on, for example.

It seems that cases like

(1) Wes is a bachelor


(2) Wes is an unmarried man

also express different propositions, even if

(3) bachelor


(4) unmarried man

mean the same thing, have the same truth conditions, are about the same thing, etc. One of these expressions is basic, one is complex.

Dan said...

Well, as I said, it's an intuition pull sort of argument. However, I do think there's some merit in saying that certain English statements can be translated into the corresponding Nenglish statements. When going from (P) to (NP), really, what's lost in translation? If the answer is nothing (or something that speakers of the language don't usually pay any attention to) then that's reason to think that they express the same proposition.

Adam, I'm not quite sure what you're getting at. Doesn't any structured proposition dude hold those two things?
Plausibly not the latter. The proposition < Adam,Tall > and the proposition < Adam,6+ft. > arguably don't correspond to the same fact. Tallness and being 6+ft. are two different properties (even if having one necessitates having the other). If a fact is just a property being had by an individual, this individuates those two facts.

Wes McPherson said...


If (P) and (NP) are translations, it seems like we have to take translation in the loose sense. This is like:

Adam says: Hamburgers are delicious.
Wes says: Adam says he loves to eat burgers.
Dan says: Adam says he enjoys to consume cooked, ground up beer shaped into patties and put between buns.

If we have the loose sense of translation in mind, then plausibly what Adam said is what Wes said Adam said, and what Dan said Adam said.

But in the strict sense of translation radically different things are being said. What gets lost in translation is something fine grained. The expressions may mean the same thing, but their structures are not identical.

Chris Tillman said...

Mmm. Cooked, ground up beer-shaped patties.

Wes McPherson said...

I promise I meant beef...

Adam said...

I think Dan's right; initially I don't think I understood the comment fully.
It's surprisingly difficult to think of two distinct propositions that are the same with respect to both aboutness facts/representation and truth conditions, IF* you don't already buy the view that differences at the level of syntax contribute to the individuation of propositions. I think the examples given so far suffer from this problem.

What about the propositions expressed by "First order logic is undecidable" and "It is not the case that first order logic is decidable"? These seem like distinct propositions, independently of any commitment to the idea that syntax contributes to individuation. But they seem to agree with respect to aboutness facts- they are both about the fact consisting of Church's theorem and truth. They also agree with respect to TC's.