"Good afternoon everyone. Tea is now served

in the kitchen. Have a great weekend."

Every day at 3 in the afternoon,

the scientists and staff at SFI meet for

tea time! This is a time to get together

for a break, have some delicious snacks,

and chat informally with whoever happens

to be around. Tea time has been an essential

part of SFI's culture since its founding.

And the promise of food and conversation

manages to get everyone out of their

offices and into the institute's large

kitchen for a needed afternoon break.

"Are you a St. John's person?"

"Graduate, yes."

"Oh. Now you're working here?"

"Yeah."

"Are you working with Simon?"

"Sometimes, yes."

Simon: "On occasion!"

"I think it would fit in beautifully

you know, next time around."

"Mhm. Okay. Well we should talk."

"I would really encourage. I'd be glad to

give you any number of examples, ways in

and so forth." "Okay."

This is Sarah, who I've just met here at

tea time today. "Hello." Sarah, can you

tell us how you are and what you're doing

here at SFI?

Sarah: Sure, my name is Sarah Klingenstein

and I am a graduate fellow. I graduated

last fall with a Masters in Eastern Classics

from St. John's College in Santa Fe and

previous to that, I studied linguistics,

also at the graduate level, at the University

of Colorado, Boulder.

"What brought you to Santa Fe?"

Sarah: Well it's kind of funny actually.

When I was in college, my first job was

as a research associate and it was this

project for the Cave, which is this

immersive environment and so I was doing

a lot of the background research. And

all the interesting...I was 18... all the

interesting stuff that I was coming across

came out of the Santa Fe Institute and so

the first time I came to visit Santa Fe,

I wanted to come to the Santa Fe Institute

and it just wasn't exactly a tourist site

so it didn't work, but uhm, so then

a couple years ago, when I started at St. John's,

there was a posting for a graduate intern

at the Santa Fe Institute and I can

honestly say that it was a childhood dream

of mine to work at the Santa Fe Institute.

"So what are you doing here?"

Sarah: Uhm a lot of my work is computational

linguistics. It's funny, it's at the uhm

infrequently explored intersection between

computational linguistics, British legal

history, and information theory.

"Oh wow!" <i>laughter</i>

Sarah: And in a broad sense, what we are

trying to do is come up with a sophisticated

way to...uh...kind of quantify changes in

uhm...in semantics over time. And linking

that to institutional change.

"In this course, we covered some of

Shannon information theory and uhm, basically

found that it doesn't really have a lot

to say about semantics. "Right."

So I'm wondering how you're using it in

your work on semantics.

Sarah: "Well, what we've had to do is

come up with a way to course-grain the

English language, and so we've actually

borrowed the structure that Roget used

when he first made his thesaurus.

What he initially did wasn't just making

synonym sets, he also built this whole nested

semantic structure of the English language.

At the very top, there are 6 categories

and all of English fits into these six

categories. What that has done is kind of

provide us with a ready-made...uhm...

nested, semantic classification system.

And so what that allows us to do is

course-grain the semantics of English

into these categories. And so

we've been treating them like probability

distributions, and so in that way...

"You can do the kind of...Shannon entropy."

Sarah: Exactly, yeah.

"Okay, well thanks a lot Sarah, very interesting."

Sarah: It was nice to meet you.

"And nice to meet you."

<i>Inaudible conversation</i>