...which is? Your "something else" isn't part of the definition of "meaning" per Merriam-Webster. Again, you sidestep a point I made. — Harry Hindu

Whatever it means in the context of "the meaning of a word". According to Wittgenstein, and probably other philosophers, it would be wrong to interpret this as "the intention of a word".

Also whatever it means in the context of "my girlfriend means a lot to me". It would be wrong to interpret this as "my girlfriend intends a lot to me".

Or whatever it means in the context of "if the grass is wet then it means that it rained earlier". It would be wrong to interpret this as "if the grass is wet then it intends that it rained earlier"

And any other contexts in which the word "mean" isn't synonymous with "intend".

It's a nonsensical scenario.

It's not a nonsensical scenario. It happens all the time. People often say the wrong thing, either because they weren't thinking clearly or because they don't know their left from their right. The fact is that even though I intended for you to turn left, my command "turn right" doesn't mean "turn left". Therefore the meaning of an expression isn't the same thing as the speaker's intention.

If that is what you are saying, then why does the phrase, "you didn't say what you meant" apply where "meant" refers to intent?

Because you intended to say one thing but instead said something else? I don't understand the relevance of this question.

There's a neat argument by Charles Travis that I think illustrates quite well what was meant by Wittgenstein when he said "meaning is use", and it also shows that the traditional view of meaning is mistaken (like that a word means that which it stands for or refers to etc.).

Travis gives an example of a sentence that can be used in one context to say something true and in another context to say something false about the very same object. And so if Travis' example is convincing, then it shows that the meaning we associate with each word in a sentence (whatever it is) is not sufficient to determine the meaning or content of the whole sentence on the occasion of utterance (and by 'content' Travis means truth-evaluable content, i.e. that which determines the truth conditions for the sentence).

Suppose we utter the sentence "the leaves are green", and point to a bunch of dead brown leaves that have been painted green. Is the sentence true or false? It depends according to Travis on the purpose for which we use the sentence: if we are interested in the superficial color of the leaves, then we would be saying something true when we use the sentence, whereas if we are, say, interested in botany then we would say that the same sentence is false (or imagine a cease of brown leaves that are lit by intense green light, or leaves that glow green in the dark and so on).

So the moral is this: whatever 'meaning' we associate with the sentence or any of the words of which it is composed, it doesn't determine in advance what the sentence means on a particular occasion of use. Knowing what 'leaves' and 'green' mean doesn't by itself tell you how to use the sentence when you talk about some particular leaves, because you have countless options to choose from. The sentence can have a determined meaning (i.e. to say something concrete about the leaves) only if we have a clear purpose in mind for which we want to use the sentence on a given occasion.

The example you've given would be a good way to explain what a proposition is. It doesn't really prove semantic externalism, though (not a hard version of it anyway.) But maybe you didn't intend it that way.

I don't think the traditional view is that we never have to look to context to understand what's meant.

First I think the argument is neutral on the issue of externalism, semantic or otherwise.

And you are correct that the 'traditional' view doesn't completely ignore the role of context in determining meaning, but unlike Travis, most philosophers believe that whenever the meaning of a sentence seem to be context dependent, then that sentence must contain some semantic 'parameter' that fixes in advance how any context of utterance can and cannot contribute to its meaning.

So a paradigmatic example for this is indexicals (like "me" "now" "here" etc.). All philosophers agree that the reference of 'I' in "I'm hungry" for example, changes on most occasions of utterance, depending on the person who utters it. However, many philosophers argue that the pronoun 'I' changes reference because of its meaning, so they regard the the word 'I' as synonymous with something like "the person who uttered the word", and this they claim explains how we can know to whom the word refers on different occasions.

And what is nice about Travis' example is that it appears very implausible that a sentence such as 'the leaves are green' should contain some sort of semantic 'parameter' or indexical that already anticipates what meaning should it have for every conceivable context (after all, neither the meaning of 'green' or 'leaves' contains a disjunction which specifies what you ought to say about any conceivable case, i.e., painted leaves, gloving in the dark leaves, brightly lit leaves; or what about leaves that are 80% green? or leaves that are 53% green? Is it semantically determined for each case when exactly should we stop calling the leaves green?).

most philosophers believe that whenever the meaning of a sentence seem to be context dependent, then that sentence must contain some semantic 'parameter' that fixes in advance how any context of utterance can and cannot contribute to its meaning. — Fafner

Does Travis deny that rule-following is an aspect of communication?

No he doesn't, he just says that rules are not enough for meaning, or better - that applying a rule is always a matter of excising a special sort of capacity which is not itself determined by antecedent rules (which I think is also Wittgenstein's point).

No he doesn't, he just says that rules are not enough for meaning, or better - that applying a rule is always a matter of exercising a special sort of capacity which is not itself determined by antecedent rules (which I think is also Wittgenstein's point). — Fafner

Then Travis is like most philosophers who

believe that whenever the meaning of a sentence seems to be context dependent, then that sentence must contain some semantic 'parameter' that fixes in advance how any context of utterance can and cannot contribute to its meaning. — Fafner

The parameter in question is simply that the sentence follows the rules (leeway given for creativity).

So the question is: why should we think of Wittgenstein's views as extraordinary?

Let me put it this way. From the Travis/Wittgenstein point of view the relevant rule is that the sentence "the leaves are green" apply only whenever the leaves are green (of course you'll have to tell a much more complicated story here, and furthermore there's really no universal rule for 'green' or being a leaf on this account, but this blunt formulation will suffice for my purpose); whereas most philosophers would say that you need an additional rule, for example a rule which specifies for each case (when the leaves are painted, the leaves are glowing in the dark and so on) whether the sentence is true or false in advance. Travis and Wittgenstein would say that there's no such rule, and furthermore that you don't need it in order to understand or apply the sentence "the leaves are green" in new cases.

So the dispute is about whether you can describe language algorithmically, such that the meaning of every expression would completely determine in advance what you should say on each occasion. Travis' and Wittgenstein's view is that the idea of such a "super rule" or algorithem is a confusion, and that rules can never replace human judgment and sensitivity, but essentially depend on them for their application (so it's a pretty deep disagreement about the nature of rules and understanding).

whereas most philosophers would say that you need an additional rule, for example a rule which specifies for each case (when the leaves are painted, the leaves are glowing in the dark and so on) whether the sentence is true or false in advance. — Fafner

Putting aside for a moment which philosophers believe that, here you are explicitly using sentences as truth-bearers instead of propositions. Why are you making that choice?

So the dispute is about whether you can describe language algorithmically, such that the meaning of every expression would completely determine in advance what you should say on each occasion. — Fafner

That's contrary to common sense, so the philosophers who adhere to that view have some explaining to do. You wouldn't say the majority of philosophers have made this blunder would you?

Putting aside for a moment which philosophers believe that, here you are explicitly using sentences as truth-bearers instead of propositions. Why are you making that choice? — Mongrel

The first reason is that Travis' original argument is concerned with sentences, not propositions, so I couldn't use something else instead when presenting his argument. Secondly and more importantly, it is because we are talking about the semantics of natural language, aren't we? So we are concerned with the meaning of sentence as physical signs, since propositions (on the standard view at least of propositions as abstract entities) don't have any semantics because they are not composed of signs.

That's contrary to common sense, so the philosophers who adhere to that view have some explaining to do. You wouldn't say the majority of philosophers have made this blunder would you? — Mongrel

I can give plenty of examples. Kent Bach is one such philosopher (see for example - userwww.sfsu.edu/kbach/Bach.ContextDependence.pdf). Another example is Jerry Fodor. Here's a random example from one of his paper that I read recently: (from "Connectionism and Cognitive Architecture")

the difference between ‘good book’, ‘good rest’ and ‘good fight’ is probably not meaning shift but syncategorematicity. ‘Good NP’ means something like NP that answers to the relevant interest in NPs: a good book is one that answers to our interest in books (viz. it’s good to read); a good rest is one that answers to our interest in rests (viz. it leaves one refreshed); a good fight is one that answers to our interest in fights (viz. it’s fun to watch or to be in, or it clears the air); and so on. It’s because the meaning of ‘good’ is syncategorematic and has a variable in it for relevant interests, that you can know that a good flurg is a flurg that answers to the relevant interest in flurgs without knowing what flurgs are or what the relevant interest in flurgs is.

(don't ask me what "syncategorematicity" is because I don't know, but from the context it is pretty clear that he means by that something like a variable)

So we are concerned with the meaning of sentence as physical signs, — Fafner

A sentence isn't a physical sign.

since propositions (on the standard view at least of propositions as abstract entities) don't have any semantics because they are not composed of signs. — Fafner

A proposition is semantic content.

I can give plenty of examples. — Fafner

Isn't Fodor talking about rules?

And by the way, there are even more extreme views than what I've described. There's something called "minimalist semantics" (advocated by Herman Cappelen for example) and on this view, save for a very small number of cases (like indexicals), context has nothing to do with semantics whatsoever. Their idea is that whenever you put together any words to form a sentence (which is 'well formed' by some criteria of grammaticity) then you automatically get a meaningful sentence with a determinate 'literal meaning' that is not open to interpretation.

So they would argue that even Chomsky's famous nonsense sentence "colorless green ideas sleep furiously" says something meaningful, though false. Incidentally, I took a few philosophy classes with someone who subscribes to such view. He argued in class that whenever we say a sentence such as "everybody came to the party" what the sentence itself literally means is that "every entity in the universe came to the party", and therefore it is always false. So yeah...

A sentence isn't a physical sign. — Mongrel

Then what else could it be...?

A proposition is semantic content. — Mongrel

Fine, but this is not the same as saying that they have a semantics like sentences.

Isn't Fodor talking about rules? — Mongrel

Obviously yes, his whole Language of Thought theory is about syntax.

Then what else could it be...? — Fafner

Whatever it is, it's not physical sounds or marks. You should be able to get that intuitionally by noticing that you and I can utter the same sentence. Two utterances (physical sounds or marks), one sentence.

There's an argument that buttons it up tight if your intuition fails you :) A sentence is a particular pattern of words.

Do you agree that they are abstract entities? — Fafner

Definitely. So?

Obviously yes, his whole Language of Thought theory is about syntax. — Fafner

If he's laying out rules, then he's not saying something Witt would disagree with, is he?

I just made almost exactly the same suggestion (as Fodor does) elsewhere, but about assertion & truth rather than goodness.

I'm not sure such an approach requires every possible use to have been determined in advance -- some can kick out as not (yet) defined, and we probably actually want that to happen, I should think.

Whatever it is, it's not physical sounds or marks. You should be able to get that intuitionally by noticing that you and I can utter the same sentence. Two utterances (physical sounds or marks), one sentence.

There's an argument that buttons it up tight if your intuition fails you :) A sentence is a particular pattern of words. — Mongrel

You are talking here about the distinction between sentence token and sentence type, and I don't see why this should contradict what I said (that they are physical entities). We can after all talk about physical entities as particulars (a cloud) or as being a type/kind to which particulars belong ("being a cloud"). If your argument were right then it would prove that physical objects don't exist...

Definitely. So? — Mongrel

What I wanted to say is that this is not the same as saying that propositions have a semantics like sentences.

If he's laying out rules, then he's not saying something Witt would disagree with, is he? — Mongrel

Witt' would definitely disagree with Fodor more or less on everything... And I don't agree at all that Witt' was "laying out rules" (whatever that means).

I'm not quite sure what you had in mind, but what Fodor meant (and many other philosophers) is that you need such rules, that tell you in advance what the truth conditions of each sentence in the language are, because they assume some strong form compositionality in order to account for the productivity of language (they argue that unless you have an algorithm that tells you what "___ is a good X" means independently of what '___' and 'X' mean and the context of their use, then we could never understand new sentences of this form, which is exactly what Travis denies).

ou are talking here about the distinction between sentence token and sentence type, and I don't see why this should contradict what I said (that they are physical entities). We can after all talk about physical entities as particulars (a cloud) or as being a type/kind to which particulars belong ("being a cloud"). If your argument were right then it would prove that physical objects don't exist... — Fafner

You utter sentence A. I utter sentence A. We uttered the same sentence.

If this was a type/token situation, then you uttered your sentence and I uttered mine. We didn't utter the same sentence. Is that what you're arguing?

What I wanted to say is that this is not the same as saying that propositions have a semantics like sentences. — Fafner

A proposition is semantic content. What do you mean they don't have semantics?

And I don't agree at all that Witt' was "laying out rules" (whatever that means). — Fafner

Witt agreed that rule-following is a prominent part of communication, didn't he?

You utter sentence A. I utter sentence A. We uttered the same sentence. — Mongrel

Yes, that's what I meant.

If this was a type/token situation, then you uttered your sentence and I uttered mine. We didn't utter the same sentence. Is that what you're arguing? — Mongrel

No, I was argueing the opposite. On the type/token distinction if you say 'cats fly' and I say 'cats fly' then we have uttered the same sentence (type), which is perfectly consistent with saying that sentences are physical entities.

A proposition is semantic content. What do you mean they don't have semantics? — Mongrel

Semantics is concerned with signs and symbolism. Since nobody claims that abstract propositions have symbolic meaning then they don't have a semantics by definition (it is fine to say that they are identical with semantic content, but it is not the same as saying that they have semantic content - it's a rather pedantic point, but this is what philosophers mean by "semantics", so it is better to follow their use in order to avoid confusion and misunderstanding).

Witt agreed that rule-following is a prominent part of communication, didn't he? — Mongrel

I don't know, maybe, but this wasn't his main philosophical concern when he discussed rules.

unless you have an algorithm that tells you what "___ is a good X" means independently of what '___' and 'X' mean and the context of their use, then we could never understand new sentences of this form, which is exactly what Travis denies — Fafner

Now I'm confused because Fodor explicitly says in the passage you quoted that the meaning of "good" is fixed and its varying applicability is explained by something else besides its meaning, and he called that its "syncategorematcity."

Another point about propositions: propositions are usually postulated in order to explain the semantic content of sentences. However, if you treat proposition as themselves having semantic content, then the question would arise, what is their semantic content? Another proposition?

I don't see how what you say contradicts what I said about Fodor. Could you explain a bit more?

However, if you treat proposition as themselves having semantic content, then the question would arise, what is their semantic content? Another proposition? — Fafner

Yeah that has to be right. Sometimes "proposition" gets used to mean something like: the sentence under consideration, disambiguated, indexicals eliminated, ellipses eliminated, whatever is needed from context explicitly added in, and so on. A sentence "normalized" in whatever way is needed. That's a useful thing but I don't know a standard term for it.

Yeah that has to be right. Sometimes "proposition" gets used to mean something like: the sentence under consideration, disambiguated, indexicals eliminated, ellipses eliminated, whatever is needed from context explicitly added in, and so on. A sentence "normalized" in whatever way is needed. That's a useful thing but I don't know a standard term for it. — Srap Tasmaner

Hmm I don't remember ever seeing such a use of 'proposition', can you give an example?

Maybe you mean when a philosopher just explains the content of a sentence by paraphrasing it with a longer or more detailed sentence?

Another use of 'proposition' can be found in Wittgenstein's Tractatus, where it simply means a sentence which is put to a meaningful use (or in the Tractarian terminology, a sign which is used as a symbol).

No, I was argueing the opposite. On the type/token distinction if you say 'cats fly' and I say 'cats fly' then we have uttered the same sentence (type), which is perfectly consistent with saying that sentences are physical entities. — Fafner

I don't think so. You have a car. I have a car. We do not have the same car (type.) We have the same type of transportation.

You utter A. I utter A.
A=A
Your utterance is not equivalent to my utterance.

A sentence is not identical to the sounds or marks that are used to express that sentence.

Since nobody claims that abstract propositions have symbolic meaning then they don't have a semantics by definition (it is fine to say that they are identical with semantic content, but it is not the same as saying that they have semantic content - it's a rather pedantic point, but this is what philosophers mean by "semantics", so it is better to follow their use in order to avoid confusion and misunderstanding). — Fafner

Semantics is the study of meaning. Propositions are all about meaning.

I don't know, maybe, but this wasn't his main philosophical concern when he discussed rules. — Fafner

What was his main concern when he discussed rule following?

Another point about propositions: propositions are usually postulated in order to explain the semantic content of sentences. However, if you treat proposition as themselves having semantic content, then the question would arise, what is their semantic content? Another proposition? — Fafner

I'm unfamiliar with propositions being spoken of as having semantic content. They are semantic content.

You could just say "content" instead of "proposition" and get along just fine.

I don't think so. You have a car. I have a car. We do not have the same car (type.) We have the same type of transportation. — Mongrel

You have to distinguish between two senses of "the same". On one sense, indeed we don't have the same car (our cars are not numerically or token identical), but on another sense we do - say if our cars are of the same brand and model (they belong to the same type). This distinction should not be very controversial unless you are some sort of extreme nominalist (are you?)

Semantics is the study of meaning. Propositions are all about meaning. — Mongrel

As I said, it's just a matter of terminology. If you don't want to confuse people, then you ought not to talk about the "semantics" or propositions.

What was his main concern when he discussed rule following? — Mongrel

The false things that people say about rules and meaning.

I'm unfamiliar with propositions being spoken of as having semantic content. They are semantic content.

You could just say "content" instead of "proposition" and get along just fine. — Mongrel

Yeah, that's exactly my point. If propositions don't have semantic content then they don't have a semantics period (because this is what "semantics" means - a theory which assigns contents to signs).

You have to distinguish between two senses of "the same". On one sense, indeed we don't have the same car (our cars are not numerically or token identical), but on another sense we do - say if our cars are of the same brand and model (they belong to the same type). This distinction should not be very controversial unless you are some sort of extreme nominalist (are you?) — Fafner

What is a brand such that millions can own the same one? Does it have physical wheels?

Yeah, that's exactly my point. If propositions don't have semantic content then they don't have a semantics period (because this is what "semantics" means - a theory which assigns contents to signs). — Fafner

OK. I don't know how this fits in to the conversation. Lost track, I guess.