How could you put together shapes to compose "I'll stop at the store on the way home tonight and get milk."?

srsly?

What do you expect when no question is asked in your OP :/

I was mainly expecting indifference, misinterpretation, maybe a little ridicule. ;)

Compositionality and the context principle are two of the absolutely central concepts of philosophy of language, but I for one have been having difficulty seeing how they fit together. Dummett has trouble explaining how they fit together. If it already made sense to you, that's cool.

in language there is no word that means, "I'll stop at the store on the way home tonight and get milk." — Srap Tasmaner

Oddly enough, though, between two people, there often are such words. I go walking with a mate every Tuesday and he gets a text at some point during the day. 'Milk' means, 'Please stop at the store on the way home and get milk.' 'No milk,' means...well, guess :)

I don't get it either, Srap. But we have to rewrite the philosophy of language from pragmatics upwards to get it all straight. I fear I may not have the time or energy.

Oddly enough, though, between two people, there often are such words. I go walking with a mate every Tuesday and he gets a text at some point during the day. 'Milk' means, 'Please stop at the store on the way home and get milk.' ' — mcdoodle

No, that's just ellipsis, and the rest of the sentence is understood from the wider context of the relationship between these people, their housekeeping habits, the rigors of communicating on mobile devices, etc. "Milk" still just means milk. (That whole sentence will never appear in a dictionary as one of the meanings of the word "milk," and for good reason.)

The word "milk" is also ambiguous, but it's pretty clear that the parties to this exchange have a usual agreed-upon meaning, so the ambiguity is not an issue here for them. You could also think of this as part of the context of their exchange.

None of this is really the sort of context at stake in Frege's context principle.

You cannot perform a complete linguistic act with just a word or any bunch of words. (Except, as noted, elliptically.) Sentences are special. So you want to say something like, "the meaning of a word is its use in sentences." (We can get much fancier about formulating the context principle if needed.)

The trick (besides dealing with the potential circularity) is to avoid denying compositionality. You can end up thinking the sentence, or the wider context of its utterance, "gives" the words in it meaning, which they lacked until they appeared in that sentence (or until its utterance in a particular context, etc.). This is patently false. If it were true, language would be impossible.

Here's how I think the two principles could be reconciled.

The primary goal of the principle of compositionality (the idea that the meaning of sentences depends systematically on the meaning of their parts, viz. words) is to explain how can we understand a potentially infinite number of new sentences which we never encountered without much trouble. And the natural idea seems to be that we can understand them because they are composed of words that we are already familiar with, combined according to rules which we have learned in the past. And so the idea is that we deduce the meaning of the sentence (which is unknown) based on the meaning of its parts (which are already known).

However, according to the context principle, words don't have any meaning on their own, but only sentences do, and this seems to be at odds with compositionality. Notice however that the principle of compositionality has two components: 1. the basic thought that to understand something new, it must contain old elements which we already familiar with 2. what the old elements contribute to the new sentence is self-standing meanings that we can grasp independently of their function in sentences. And my proposal is that we can accept (1) (which is not at odds with the context principle) and reject (2) (which is), thus preserving the general explanatory framework of compositionality without accepting its specific explanation of how the meaning of sentences is determined by its parts. And once we make this distinction, it becomes clear that there could be other explanations of how words contribute to the meaning of a sentence, which do not assume that words have meanings on their own.

Here's my alternative explanation of how this can work. When we encounter a new sentence, what we do is not to first analyze the meanings of its individual words (and then deduce the meaning of the complete sentence) but we compare the new sentence with the vocabulary and syntactical structure of other sentences which we are already familiar with. So to take a somewhat simplistic example: suppose that we encounter a sentence like "the dog is on the mat", and suppose also that as children we learned the two sentences "the cat is on the mat" and "the dog is on the sofa", but have never encountered "the dog is on the mat". We can see that there are analogous components and structures in common between the sentences: we know what it is for a dog to sit on a sofa, and we know what it means for a cat so seat on a mat, and from that we can understand that a dog is something that can also sit on the mat analogously to a cat. So what I'm saying here is that what we know is not the "meanings" of individual words like 'dog' and 'cat', but rather we are familiar with sentences that make claims about cats and dogs, and by analogy we employ this knowledge to grasp new sentences about dogs and cats (so there's nothing in this process which we grasp that is less than the meaning of a complete sentence). There's no such thing as abstract knowledge of the meaning of words like 'cat' or 'dog', rather we have a stock of propositions about cats and dos which we understand, on the basis of which we can understand new propositions about cats and dogs. We do however need the new sentences to be composed from old familiar words (this is the part which is correct in compositionality), however they don't have a meaning on their own which they carry into new sentences, but they serve as syntactical cues which help us to connect the new sentences with the old familiar ones.

However, according to the context principle, words don't have any meaning on their own, — Fafner

Why would anyone claim that?

The short answer is that the philosophical notion of "meaning" is pretty obscure (as it was argued for example by Quine in "Two Dogmas of Empiricism"), and the notion of the meaning of a single word is even more obscure. As I see it, the real question here is not whether words have "meaning" on their own, but what is "meaning" in the first place? What kind of philosophical work is the notion supposed to do?

So if we go back to Frege, who first formulated the context principle, what interested him was truth and not meaning, and therefore he proposed that we should ask about meaning in language only so far as it helps to illuminate the concept of truth. However the concept of truth clearly applies to whole sentences (propositions) and not single words, therefore Frege proposed (as methodological maxim for investigating truth) that we shouldn't ask about the meaning of words outside the context of a sentence, but ask what kind of logical contribution does a given word makes to the truth conditions of the sentence in which it appears. So there is no such thing for Frege as taking a single word like 'cat' and asking about it's "meaning", rather we should ask what is in common between all sentences that employ the word 'cat'.

It is also connected to Frege's rejection of psychologism, that is treating logical questions as psychological, since he thought that thinking about words an isolation from sentences encourages a picture where we take "meaning" as something psychological or subjective, which would undermine the objectivity of logic and language that Frege wanted to defend.

I should also add that I'm not objecting (on the behalf of the contextualist) that words can be legitimately said to have "meaning" in some non-philosophical sense of the term, if this means that some strings of letters or sounds are recognized to belong to this or that language (e.g. the word 'cat' is part of/has a role in the English language, as opposed to 'ajklorlsd' or 'חתול'). The objection is rather to a special 'philosophical' sense of meaning, where it is usually associated with notions such as concepts, ideas, universals etc.

The short answer is that the philosophical notion of "meaning" is pretty obscure (as it was argued for example by Quine in "Two Dogmas of Empiricism"), and the notion of the meaning of a single word is even more obscure. As I see it, the real question here is not whether words have "meaning" on their own, but what is meaning in the first place? What kind of philosophical work is the notion supposed to do? — Fafner

Well, meaning is simply a mental association one makes with a term (for example). One can easily do that with a single term.

I should also add that I'm not objecting (on the behalf of the contextualist) that words can be said to have "meaning" in some non-philosophical sense of the term, — Fafner

I don't buy that we're talking about some "special sense of the term 'meaning'" in philosophy. If Frege only cared about truth, then that apparently led him astray in his analysis of meaning.

Well, meaning is simply a mental association one makes with a term (for example). One can easily do that with a single term.

That's fine, but in itself it doesn't have any philosophical significance. As I said, it really depends on your philosophical goals. Usually in philosophy of language "meaning" is supposed to do some specific explanatory work of this or that phenomena, as in the argument for compositinality which I presented (i.e., if words had no meaning on their own we couldn't understand unfamiliar sentences). "Meaning" as a psychological association between words and mental images has no philosophical interest in itself (maybe it's true, maybe it isn't, but philosophy obviously can't decide that).

I don't buy that we're talking about some "special sense of the term 'meaning'" in philosophy. If Frege only cared about truth, then that apparently led him astray in his analysis of meaning.

But it's a fact, philosophers did propose all sorts of definitions and theories for the notion of "meaning", so it does become pretty technical in many discussions.

That's fine, but in itself it doesn't have any philosophical significance. — Fafner

The philosophical significance is that it gets right what meaning is.

The point of philosophy isn't to make shit up that has no resemblance to what the world is really like.

But it's a fact

On my view, but it's not. Philosophers propose all sort of definitions and theories for the notion of meaning, sure, but they're not doing so by way of making shit up. They're trying to analyze what meaning is. Lots of philosophers are getting that wrong.

The philosophical significance is that it gets right what meaning is. — Terrapin Station

Well, and biology gets things right about beetles and flowers - would you also say that it is philosophically significant?

The point of philosophy isn't to make shit up that has no resemblance to what the world is really like. — Terrapin Station

So far you are the one here who makes shit up. Do you have any empirical evidence that people always have mental associations with every word they know? That sounds to me like a totally far fetched claim. For example I have no idea what kind of things I associate with most of the words I know, except perhaps some faint images with familiar nouns. What mental associations do you have with a words like 'and' or 'because'?

Philosophers propose all sort of definitions and theories for the notion of meaning, sure, but they're not doing so by way of making shit up. They're trying to analyze what meaning is. Lots of philosophers are getting that wrong. — Terrapin Station

There's no "correct" analysis of 'meaning', because there are many different senses in which this word is employed, inside and outside philosophy.

Well, and biology gets things right about beetles and flowers - would you also say that it is philosophically significant? — Fafner

Analysis of what meaning is isn't a field other than philosophy. What biology gets right about beetles and flowers is certainly important for philosophy of biology. Philosophers of biology aren't just going to start making shit up that has no connection to facts per biology.

Do you have any evidence that people always have mental associations with every word they know? — Fafner

You're attempting to lecture me about what constitutes philosophy, and then all of sudden you start acting as if it's an empirical science? Seriously?

There's no analysis of "meaning", — Fafner

Wow, the bullcrap is deep in this one. You'll fit in well here.

Thanks for your thoughts. I'm tied up at work right now but will get to this soon as I can.

Absolutely we'll deny (2) as formulated, but I don't think the "only sentences have meaning" view can be made to work.

More later. Have fun with Terrapin.

Analysis of what meaning is isn't a field other than philosophy. What biology gets right about beetles and flowers is certainly important for philosophy of biology. Philosophers of biology aren't just going to start making shit up that has no connection to facts per biology. — Terrapin Station

What does it mean then to make a philosophical analysis of meaning? How should we decide who is right?

You're attempting to lecture me about what constitutes philosophy, and then all of sudden you start acting as if it's an empirical science? Seriously? — Terrapin Station

I'm just saying that someone who's complaining about people making things up, shouldn't do the same thing himself...

Wow, the bullcrap is deep in this one. You'll fit in well here. — Terrapin Station

That's not an argument.

but I don't think the "only sentences have meaning" view can be made to work. — Srap Tasmaner

What's the alternative?

And by the way, don't change my words when you quote me. I didn't say "There's no analysis of "meaning""

And by the way, don't change my words when you quote me. I didn't say "There's no analysis of "meaning"" — Fafner

As I'm doing here, I highlighted your post and hit the "quote" button. It's not as if I retyped anything. You apparently edited your post after I quoted you.

What does it mean then to make a philosophical analysis of meaning? How should we decide who is right? — Fafner

Tell me something about your background with philosophy. I'm curious how this question can fit with the content and tenor of your other comments in the thread so far.

I'm just saying that someone who's complaining about people making things up, shouldn't do the same thing himself... — Fafner

And indeed I wasn't. I was talking about what meaning is, in general. I wasn't making up something that bears no resemblance to that and saying that it's meaning--"Oh, just in a special philosophical sense."

That's not an argument. — Fafner

If only that could imply that it's not the case.

Tell me something about your background with philosophy. I'm curious how this question can fit with the content and tenor of your other comments in the thread so far. — Terrapin Station

I don't see how this is relevant.

And indeed I wasn't. I was talking about what meaning is, in general. I wasn't making up something that bears no resemblance to that and saying that it's meaning--"Oh, just in a special philosophical sense." — Terrapin Station

So on your account, do you want to say that all meaningful words must have mental associations? (because that's an empirical claim whether they do), and if they don't does it follow that they are not meaningful?

I also don't understand what you mean by something that bears "resemblance" to meaning (I suppose you meant something like a definition). How do you identify something as "meaning" in the first place, and know whether your definition "resembles" it or not? (and it's connected to my first question about what it is to analyze something in philosophy).

As I'm doing here, I highlighted your post and hit the "quote" button. It's not as if I retyped anything. You apparently edited your post after I quoted you. — Terrapin Station

You are probably right, my fault.

I don't see how this is relevant. — Fafner

I don't see why it matters if you see how it's relevant.

HI Fafner, nice to see you here.

Hi mcdoodle, nice to see you too. Are you from the old forum?

The alternative is to say that words do have sense, and that the sense of a word is the contribution it makes to the sense of a sentence in which it is used.

You avoid circularity by beginning with sentences simple enough for you to have a way of grasping their sense independently. These atomic sentences are just what Frege forms concepts/predicates from.

More later.

The alternative is to say that words do have sense, and that the sense of a word is the contribution it makes to the sense of a sentence in which it is used. — Srap Tasmaner

Not that anyone is necessarily doing this, but I think it's important to remember that it doesn't work in a "robotic," black & white way. Words in isolation often have meaning (and they can express fairly complex thoughts or actions), but it's also the case that phrases and complete sentences can be "units" of meaning on their own. In other words, it's not that people always assign meaning to individual words, and then concatenations of them are always concatenations of the meanings of the individual words in the phrase or sentence. It varies by person, by situation, by context, etc.