Good afternoon again. Thank you all for joining. I'm very pleased to introduce the speaker

of our Moll lecture today, Professor Dr. Stefanie Ebert. Stefanie is a professor here at the

Moll. She is running the chair of computational linguistics, computational and composed linguistics.

Actually, she has a background as a bachelor in mathematics.

Well, okay. So, she will tell us more about the trajectory. But this is a particular case of

successful path in which from math you can get, you know, impactfully into different areas, like

in this case computational and composed linguistics. So, Stefanie is the vice spokesperson of Moll. So,

she has been with us from the very early of the conception of the center, that as you know was

as a merging point or a point of intersection of different groups and disciplines here at the Moll.

Certainly, this is all related to mathematics, mathematics and data science,

artificial intelligence, computational mathematics. So, this link us very naturally to other groups,

like the groups in computational mechanics here in the faculty of engineering, computer science,

and so on, but also through machine learning to computational and composed linguistics.

Maybe probably this is a very superficial understanding, but if somebody had any doubt

that there was a link between linguistics and mathematics, now I will check, check it,

and you will see that indeed, you know, there is a connected path of a steady configuration that

leads you from one configuration to the other. So, Stefanie Ebert, before she's a transgender woman,

so you want to find her full academic achievements, you will have also to pay attention to

the fact that earlier she was Estefan Ebert. So, you can see this name change, well, which nowadays

that, you know, counting in Google Scholar is so important makes, you know, is not totally irrelevant.

I'm still not sure how well Google Scholar has dealt with the name change. Are they counting well or not?

I'm not sure whether it gets, it seems to get, it seems as if you have difficulty getting a few of

the new papers, although I've specified both names there. Definitely. So, the best thing is to use

APA-like citations, or you just have abbreviated first names. We have this problem also because we use two family names

normally, officially, and then in the standard international system, they only take the last name,

so many, many people in Spain have to merge both names to make one full name so that, you know,

these systems recognize your work properly. Anyhow, so we are very delighted that today,

in principle, we are learning about this field in a way that we, you know, in particular,

us as very simple-minded mathematicians can understand. Thank you, Estefan.

Well, thanks very much. So, today I'm going to talk about a topic from my research field of

computational corpus linguistics, which means there will be a lot less mathematics in this

talk than we're used to from the lecture series. Instead, we're going to look at linguistic research

questions and certain fairly simple quantitative techniques we can use to find evidence to answer

these questions. So, thanks very much for coming anyway. I hope it'll still be interesting,

and there's a little bit of math there too. Okay, so the research question I'm addressing

actually is situated in a division, in an artificial division that has a long tradition

in linguistics. It's sort of something that we've overcome in the field, but for a long time,

people used to think of linguistics as a division between grammar and lexis. So, lexis is the

vocabulary, the words. Lexis is dealt with in dictionaries. Grammar is dealt with in

grammars of the language, often prescriptive grammars that tell you how, what the right way

is to form sentences from the words of a language. Grammar and lexis were seen as two entirely

different fields or different aspects of language because grammatical rules are fully schematic and

productive. That means they are regular. You can apply them in any case. You can form many

new sentences using these rules, and then applying a completely regular and transparent, or as we say,

compositional manner. Compositional means that the meaning of a sentence, when you use a grammatical

rule to form a sentence, the meaning of the sentence can be derived from the meaning of the

parts. So, there's this, it's meant to be unambiguous that you can get to the meaning of the sentence

from the meaning of the parts. On the other hand, the vocabulary, the lexis, was always seen as

completely fixed and arbitrary. So, the relationship between words and their meanings is arbitrary

convention. That's why we need to list all the words of the language in the dictionary.