Kaplan and Lewis

Let's see if we can make sense of the differences between Kaplan and Lewis. 

Here is one way to see what Lewis is up to:

Kaplan-to-Lewis Translation Method

S is true relative to context C iff the character of S determines in C some content CON and CON is true relative to C

It is important to see that what the above says is that Kaplan’s two levels of meaning are reducible to one level of meaning (remember my remark that Lewis is the arch-reductionist). The left side (the stuff before the ‘iff’) is equivalent to the right side (the stuff after the ‘iff’). So, if the above is correct, Kaplan’s two levels of meaning can be reduced to one level of meaning. It is not as though character and content are gone. They are just combined in one stage. The context C gets all of the stuff relevant to the character stage AND all of the stuff relevant to the content stage. So, C gets bloated in order to handle everything that is relevant for determining a truth-value of S. We just need ONE context to determine the truth of S, according to Lewis. 

However, if at the character level the context C does something different from what the context C does at the content level, then mashing character and content levels together and having them both take contexts as inputs to yield a truth value as output will distort those different operations. 

In other words, if the character of S is a function from the context of an utterance of S to the content of S (as Kaplan claims) and the content of S is function from the context of S to a truth-value of S, AND what the context does at the character level is different from what context does at the content level, then Lewis’s move will fail to capture that difference. 

Analogy: addition and multiplication are similar functions, but they function differently depending on the order. 3 + 4 x 5 = ? We have to separate the functions clearly to know what the value is—(3 + 4) x 5 is very different from 3 + (4x 5). The order of operations determines a different answer or value. 

Kaplan assumes that we lose valuable information if we collapse the two levels into one. But this is true only if the semantic value (the meaning) at the character level + the context is different from the semantic value at the content level + the context, such that their interaction – i.e. (character level + context) + (content level + context) is different from (character level + content level + context). 

So, in order to see who is correct, we have to come up with a case where Lewis’s reduction fails to account for some semantic fact that Kaplan’s two-level theory accounts for. 

Recall that one of the things that a theory of semantics must do is account for the fact that different types of utterances (different types of sentences) can say the exact same thing and that the same type of utterance can say different things. So, if Bob says on Monday “I will take out the trash tomorrow” and Mark says on Monday “Bob will take out the trash the next day” both Bob and Mark have said the same thing. 

Bob says on Monday, Nov 18th “I will take out the trash tomorrow” 
Mark says on Monday, Nov 18th “Bob will take out the trash the next day” 

Bob and Mark have said the same thing. They have both said

            Bob will take out the trash on Nov 19th.

Kaplan’s analysis works all this out in two stages. At the character stage, the sentences are different, whereas at the content stage, the sentences are the same. More precisely, the character of 

            S = I will take out the trash tomorrow

Is a function from the context of utterance to a content. Since the context is Bob’s saying S on Nov 18th, we plug that into the function and get the content

            Bob will take out the trash on Nov 19th

(Note that if we did not know the context of utterance we could still grasp the meaning at the character level. What we would grasp is this: the speaker of S will take out the trash the day after the utterance. In other words, we would grasp that the context needs to be filled in to get the exact content, but we would grasp what that context in the abstract is)

The character of

            S* = Bob will take out the trash the next day

Is also a function from the context of utterance to a content. Since the context of this utterance is Mark’s saying S* on Nov 18th, we plug that into the function and the content

            Bob will take out the trash on Nov 19th.

(Note that if we did not know the context of utterance we could still grasp the meaning at the character level. What we would grasp is this: Bob—referent determined by context—will take out the trash the day after the utterance. In other words, we would grasp that the context needs to be filled in to get the exact content, but we would grasp what the context in the abstract is)

So, we have different characters that deliver the exact same content because of the contexts of their utterances. 

The character/content distinction has allowed us to make sense of how different types of utterances can say the exact same thing.

But Lewis’s account appears to obscure this. All we have on his account, you will remember (see above) is that S is true relative to a context C. Since S and S* are different types of utterances, their truth conditions are different. But they say the exact same thing. So it is possible to say the same thing even though the truth conditions are different. So, having the same truth conditions is not necessary for saying the same thing. So, Lewis needs to tell us what is necessary for saying the same thing. With what we have thus far, Lewis’s account is incomplete. He needs to add something, while Kaplan does not. If we scored these two theories at this point in the game, Kaplan’s theory is ahead. 

Comments

Post a Comment

Popular posts from this blog

Grice's First Try

1 st Try : “x meantNN that p” iff x was intended by S to induce a belief in H First, let’s get clear on what this 1 st Try is saying: ‘x’ refers to an utterance of some sort, where even gestures and various other kinds of conventional signals count as utterances (e.g. three rings of the bell…). ‘MeantNN’ stands for ‘conventionally means’ or ‘has a meaning that is not determined by some natural law or whatever.’ ‘p’ refers to the proposition that the utterance expresses. For example, the proposition that my utterance of “Snow is white” and your utterance of “La nieve es blanca” both express the same exact proposition.  So, on the left side we have: an utterance of x conventionally means that p The right side should be fairly clear. It is saying that the utterance is intended to bring about a belief in someone. Putting this all together we get: an utterance of x conventionally means that p iff the utterance of x is intended by S (the utterer) to bring about a belief in someo...
Read more

Donnellan on Russell

Attributive “A speaker who uses a definite description attributively in an assertion states something about whoever or whatever is the so-and-so”  Referential “A speaker who uses a definite description referentially in an assertion ... uses the description to enable his audience to pick out whom or what he is talking about and states something about that person or thing” “Smith’s murderer is insane.”  Russell : definite descriptions only apply to attributive uses (if it applies to anything) Donnellan : it is possible for someone to say something true even though nothing satisfies the description ‘the F’ in ‘the F is G.’ Semantics or Pragmatics              “Smith’s murderer is insane.”              Does the meaning of this sentence change on different occasions of use?          ...
Read more