Carrion Luggage

Curiosity Killed the Categories

So, you were told there would be no math? Lies. There is always math. From Quanta:

    Where Does Meaning Live in a Sentence? Math Might Tell Us.  
The mathematician Tai-Danae Bradley is using category theory to try to understand both human and AI-generated language.

And no, I don't understand it, either. If everyone understood this stuff, it wouldn't need a whole article, would it?

Growing up, Tai-Danae Bradley had no love for math. In 2008, she entered the City College of New York, where she played for the basketball team and hoped to start a career in sports nutrition.

Which would have required math.

But in her sophomore year, her calculus professor changed her mind. Mathematics, she learned, was the language that all the sciences are written in.

Well, all the real sciences, anyway.

Now, as a researcher at the artificial intelligence company SandboxAQ, and a visiting professor at the Master’s University in California, Bradley is using the language of math to try to better understand language itself.

I've known for a long time that language is mathematical. Not that I had the means or ability to analyze it myself, of course, but, like with a turbulent river or a thunderstorm, there's math that describes it, even if we haven't quite figured it all out yet.

Her lens is category theory, a way of stepping back from the specifics of any individual field in favor of a broader underlying framework that bridges all of them. By thinking of language as a mathematical category, she’s been able to apply established tools to study it and glean new insights.

Which is one reason this article caught my eye; I like it when things from different fields of study can be connected.

The bulk of the article is in interview format. (Don't worry; there's nothing we'd call "math," unless you count some equations in illustrative photographs.) I'm not going to quote much of that; it's there and I think it's pretty much accessible. There's really only a couple of quotes by Bradley that I wanted to highlight, like this one:

I’m very interested in this phenomenon where things that feel different turn out to be fundamentally related.

And I am, as well, though I don't have the formal tools to analyze such things. This is why I like to learn a little bit about a lot of things rather than a lot about one or two things. Ideally, I'd learn a lot about a lot, but I'm entirely too lazy for that.

So I pick up what I can, when I can. Like this article, for instance.

One final quote, then:

I think in five years, we might have new mathematical ideas that were inspired by language.

And this, I want to see.