Ling 499R: Understanding Language Understanding, Spring 2026

Course Description

The verb "understand" gets thrown around a lot these days. We use it to refer to the process of human language comprehension. We talk about, and debate, the idea of AI understanding what we're saying to it. We also use the word to refer to our goals in scientific research: we seek to understand how language works. In this class we'll ask: What does it mean to understand language? How do we decide if someone (or something) is doing it? What are the problems that a language-understander has to solve, and what do we know about how it's accomplished, by humans and otherwise?

In the past, this was an advanced seminar limited to students majoring in computer science, linguistics, or a combination. I'm now re-gearing it as a general interest course. It's going to be calibrated for advanced students who don't necessarily have either of those as their focus. What does that mean, exactly? You don't need to program, or solve math problems, or do linguistic analysis (though if you do know those things, that's great, too, and people studying CS and/or Linguistics are very welcome). The focus of the class is on understanding ideas, not on building practical skills -- except in the sense that grappling with ideas, understanding arguments, and assessing evidence are among the most important skills you could possibly have.

You do need to be open to learning about what it means to understand language from a computational perspective. If you decided "I'm never going to look at page that has an equation written down on it, ever again, never, never, never!" when you finished high school, this probably isn't the right course for you. If you didn't mind math so much, and you like rigorous, careful thinking, and you think language understanding is a cool topic -- something that includes anyone -- come try it out!

In terms of previous knowledge, I'd like for students coming to the class to already have seen either the basics of linguistics as taught in LING 200, or have foundations in other fields that are adjacent to the scientific study of language like computer science, hearing and speech, information science, neuroscience, philosophy, psychology. More advanced students (e.g. those who have taken CMSC 470) are also welcome. If you're not sure about the fit, or if you're an interested student from a less-adjacent field, please feel free to mail me at resnik+ling499sp2026@umd.edu or come to the first class.

I consider every course I teach a work in progress. Although some formal requirements are necessarily spelled out in advance in the syllabus, as well as the overall plan for the semester, the exact content of classroom time, readings, and homework assignments may evolveas we become better informed about the capabilities and interests of the people taking the class.

If you might be interested in the class but aren't sure if it's a fit, please feel free to email me at resnik+ling499sp2026@umd.edu or come to the first class.

Course Goals

After successfully completing this course you should be able to:

Discuss key components of research on NLP and human language understanding
Identify correspondences (and lack thereof) between artificial and human language understanding
Wow people at parties with your erudite assessment of the latest developments in AI

Logistical info

Course Times: TuTh 11:00am-12:15pm

Classroom: Jimenez (JMZ) 1224 [map]

Syllabus: TBD

Schedule of Topics:TBD

"Flavor" of the class

I know it can be hard to get a sense of what a course will be like if you don't know what the materials look like, and you don't know the professor. I'm still working on the materials, but if you want to get a sense of me as a professor, see my web page. And if you want to get a sense of what I'm like talking in front of a room full of people, see various videos of me doing just that. (For an example of me talking about AI to people who don't have technical background, see "What You Should Know About ChatGPT" for a talk I gave to the university's Alumni Board of Governors meeting a few years back.)

What you'll be expected to do

The most important work for the class will involve you doing readings, and then me talking in front of the room lecture-style but with a lot of opportunity for discussion. Reading and responding to those readings will be responsible for a big part of the grade. I'm still on the fence about whether the class will have zero, one, or two exams, or if you'll do a course project instead. There will be weekly assignments, e.g. reactions to the readings, but there definitely will not be problem sets, programming, etc. -- my focus is on computational thinking, not computational doing.