Okay.

Okay, so.

I think I have everything.

So today is a theory section again.

Tomorrow will be live again. I've seen quite a few have uploaded homework, which is excellent.

But I think it's reasonable.

And since we're not only going to do theory today that you have until tomorrow.

And tomorrow we'll just basically talk about what you did.

And then we can press it into a shape or one thing that will also happen. And I'm kind of setting this up this way is that part of the problem was too big for you.

You didn't have the prerequisites, but it's important that you think about it before so that you're ready for the technical solutions.

And that's the way we set up the last homework as well. And then we'll just see how to deal with things.

So, last time we've looked at essentially the problem or what could the topic of this course in these natural language semantics even be.

How can we get near that? How can we understand what the problems are? And I tried to give you a couple of just basically phenomena that have something to deal to do with it so that you have kind of a

shape of the problem.

They will also kind of

appreciate some of the solutions we will look at later. So the first thing to for me to say was that semantics of natural language is difficult.

Everybody has their own version of what the problem might be. Psychologists basically say oh it's all we have up here.

And mathematicians just basically say oh yeah there's math in the background somewhere and philosophy has good ideas but no constructive solutions that you can compute with.

And those are the things we're going to try to get at in this course.

We've looked at translations.

We've looked at words.

And systematic phenomena like here we had in a various, various instances of the so called type token.

And this is a instance of this kind of understanding newspaper is something like a type.

It's the type of today's newspaper.

This here is a token.

And this is something completely different.

We've looked at various phenomena, things that can go non trivial when we pass from language to some kind of semantic information, or the other way around.

We have to work around all of that. And we saw a couple of examples in the abstraction, we have the problem of synonyms.

So we can express the same concept in various ways. The other thing is a single utterance can have multiple distinct meanings.

We have the bank.

We have all kinds of composition issues where it sometimes becomes difficult to see the contribution of words, of single words in the new representation we have later, as evident kind of the contribution of the word every kind of smeared all over this first of its own.

And the last thing we really talked about was that we kind of have

If we look at this, at the pipeline between meaning and meaning and utterance in a little bit more depth, we've kind of identified two phases one is kind of a syntactic phase that gives us something which we could call meaning.

And then we still need a process that kind of fits that into the current context.

And that actually takes stuff from the context very often into the meat.

This is the kind of raw or generalized leading representation that instantiates to what the utterance tries to convey.

One thing we see time and again, is that an utterance, the actual speech or piece of text conveys more information, or wants to convey more information than it actually linguistically realizes.

And we're going to see more, more instances of that as well.

And this phase is kind of what we call semantics construction.

Taking an utterance, a string or a waveform or something, and make meaning out of it.

And here we have semantic pragmatic, or just pragmatic analysis.

Any questions so far? That's all we got.

But of course it's even more difficult.

Think of the same story I told you last week. We need somewhere out there, when does the course start?

Somebody said begins at 11.

It would just as well if you say it begins at 11.

There are a lot of things that are not linguistically realized here.

What is it? The lecture.

Well, that's kind of a movement from the context into the meaning.

We're copying stuff down, if you want.