Now recording. So, okay. Are there any questions before we start?

By the way, the reason this quiz was funnily timed was I got scared that the questions

would be too long and that I think gave you two more minutes and I didn't want to kind

of start earlier after having already announced it. So, but then I think the result is that

I did, that that was correct, my assumption that some of the problems actually needed

to be done by the other members. Okay, good. So, where are we? We've completed the extremely

boring time of one. The only thing that might be surprising to you is kind of all the syntactic

categories. What's a noun phrase? What's an adjective? Some of those I'm assuming you

know from school, but many may not actually remember the difference between a common name

and a noun phrase and all of those kind of things. They are kind of what we need to deal

with here because syntax is important. We'll add others to that. Okay, so fragment one

was essentially just for getting to know the framework and the probably most important

thing is this, where we used the translated sentences or at least the sentences that we

three readings, the claimed readings, they gave us different semantises and then we've

tested those against the truth conditions, the truth conditions were kind of cast into

situations which can be expressed in the same logic and then we get the entailment. And

of course, importantly, the disentailment, namely that the situations we dreamt up make

the one reading they're intended for true and the other readings that they are supposed

to distinguish against make them false. That's what truth conditions do. And we kind of come

full circle in test driving fragment one by seeing that the claimed truth conditions,

which is kind of our observation, namely native speakers are convinced by the situation where

there's a big red sports car and the gangster and John, Peter, John, chasing each other on the back

seat. The observation is that native speakers are convinced by that, that this is a reading

our prediction, which all of this gobbledygook leads to

is born out of the right prediction. So we're doing science, natural science actually.

And now of course, that's about as far as we can go with fragment one because it's so boring.

But the kind of thing we've established is

we have this kind of Christmas tree picture where we will see that quite a lot where we

start out from the natural language utterance, say the draw and gangster

utterance, we can start by parsing, we can start with a syntax tree by semantic construction.

We create a logic expression. Sometimes we call those quasi-logical form because we still need

pragmatic analysis afterwards. Only that in fragment one, which is so boring, we don't have that.

Okay. Professor, would you mind sharing the screen with us? No, thanks for reminding me.

Thank you. You should do that earlier because I forget almost everything and the people here in

class don't know that I forgot. So you have to tell me earlier. Okay. Now you see the Christmas

tree picture, right? Yes. Excellent. So we haven't experienced the last bit here, pragmatic analysis,

but we've kind of done something with the logic expression, which is verify readings in this case.

Okay. And we might understand that a pragmatic analysis as well because if we hear that sentence

in a particular situation we are in, for instance, one where we can see the sports

car, the small sports car and John is sitting in it, then we can discard by just exactly the

reasoning we have done so far, we can discard two of the readings and instead of having the sentence

be ambiguous between three readings, it's now after pragmatic analysis no longer ambiguous

because we know that two of the readings are inconsistent with our observations of the situation.

Okay. So that's kind of what we do when we do this whole waterfall, namely we analyze,

where we analyze the meaning and one of the things is that by the intrinsic processes

we have like ambiguity, we have this multiplication of readings, right? We have syntactic ambiguity

here. There might be semantic ambiguity that we have to take into account during semantics

and so that makes more and more and more and more readings, but we can use the pragmatic analysis

to condense them, to prune again and the hope is that will be more or less non-ambiguous.

Of course that hope is false until we kind of get much better in pragmatic analysis.