Hmmm

[philena]

Remember how I said in my last entry that I like semantics? I don't know how much of that is still true. So far I'm keeping up, but I can tell that in the future serious trouble will come from the symbolic representations of x being in this set and not in that set when we apply set applications of 1- or 2-place characters to various different collections of elements of type tau and so on and so on and so on, and all that just to say that Kim doesn't smoke. I felt that diagramming my minimalist syntax was fun once I knew how to move heads and phrases and merge and make various features agree, because it was like this new computer game that I was figuring out how to make work, and once I got it right, I would end up with a good sentence. But I never really believed that minimalist syntax is a good representation of how language works. My distaste for semantics, however, is different. I don't feel that all my set representation might be fun once I figure out how to get a handle on all the mechanics of double brackets and and braces and semi-colons. I mean, it might be, but I have my doubts that Ill ever get a handle on all the logical notation. I sometimes look at problems and feel exactly the same sense of bafflement that I felt in honors calculus my first year (the lowest mark on my GPA, a factI joke about when giving tours, but which I don't actually think is all that funny). I might be able to crank out some sort of proof that "Kim runs and smokes" is true if and only if Kim is a member both of the set of things that run and of the set of things that smoke, but even at best I'll only be mindlessly applying stuff I was taught in class, but not really understanding what's going on.

The problem with this different attitude towards semantics from the one that I had towards syntax, though, is that I get much more of a feeling that these representations of truth values, truth conditions, membership towards sets, unionizing and intersecting sets, and so on, is much closer to the way we understand sentences than syntax trees are towards how we form sentences. I understood the syntax theory very, very well towards the end of the quarter (or at least, as much of it as we had been taught), but I felt as if all I was doing was learning the rules of a game that had no real-world applicability. But I can see the applicability of logical representations of truth conditions for sentences. Knowing exactly how we can see that one sentence is true while another is false seems incredibly useful for understanding how our perceptions of language relate to our perceptions of the real world. And if I'm right, if the current semantic theory I'm learning is in fact more related to the real state of things, then it's not good that I can tell right from the beginning that it's going to leave me in the dust exactly the way math (whose applicability no one disputes (as far as I know)) did my first year. What good is being a linguist if the parts of linguistics that seem the most real are also the parts I can't understand?

Comments

[(Anonymous)]

The sentence "Kim runs and smokes" could be referring to Kim smoking while she ran, or running so fast that she looks like she, metaphorically speaking, "smokes" from air friction, or even - I admit this is a little bit of a stretch -- she has a bad cold with a runny nose while she smokes. I think the meaning of a sentence is based first on its context-- that is the foundation -- and set theory is not the right mathematical tool for handling this basic fact. Set theory tries to assume that language is basically like mathematics or a computer-programming language.

[philena.livejournal.com]

I'm certainly aware of the various interpretations of the sentence. However, all the different ones you give are built the same way: they assume that while Kim is engaged in one activity, she is engaged in another activity at the same time. The meaning of those activities is embedded in the lexical entries, which everyone just has to memorize. What the semantics we are dealing with right now addresses is how those sentences are built: lexical entries interact with each other to build a different meaning. The reason we know that they interact in a regular way--like mathematics or computer-programming languages--is that we can take different lexical entries and use a series of rules to build them into propositions of infinite variety and almost infinite complexity, which are comprehensible to anyone who has access to those rules. For that to work, the rules must be regular, like mathematics.

Although you're right that set theory is not the right way to approach it. We've just started dealing with a new notation, which assumes that verbs act as functions and take nouns as their arguments.

[(Anonymous)]

We can't assume that she engages in another activity at the same time.
Suppose "Kim runs and smokes" is used to mean that she is a jogger and a smoker. There is then no assumption that she ever is smoking while she is running. Again, context rules.
I'll tip my hand about how I think semantics should be approached. Suppose any listener (or reader) before hearing (or reading) the next sentence has already formed a list of possible meanings for that sentence, and these meanings are based on context -- previous sentences, what is going on around them, etc. The person constructing the sentence has much the same list (after all, both speaker and listener share their context). Then all the speaker has to do with the next sentence is to separate out the meaning that is to be communicated from the shared list -- and that is all any sentence is designed to do. Strictly speaking, then, a sentence taken in isolation has no particular meaning without a specified context. We are so used to forming contexts that people reading example sentences in discussions of semantics automatically tend to "back create" appropriate contexts for the example sentences based on the discussion that follows them.

[philena.livejournal.com]

But how is the listener going to have a set of meanings for that sentence? He has no way of knowing what the sentence will be, or how it will be constructed. The only way he arrives at the meaning is through a knowledge of all the possible words that can be used (the lexicon) and a knowledge of all the possible words that can be applied to combine those words into various meanings. All that he already has going into a speech utterance is a previous knowledge of context.

But even if we accept that he has a list of possible meanings in his head already, how is he going to construct those various meanings? He will be using the words that make up the sentence, and then applying certain rules to them, which will result in the various meanings. And one of those meanings will be "Kim is a habitual smoker and Kim is a habitual runner." Our assignment was to take that specific meaning and analyze how it was arrived at, although we were all aware that there were other meanings. We were just instucted to ignore them.