Month: June 2021

Chris Collins interviews Noam Chomsky about formal semantics

Over on his blog, Chis Collins has posted the text of a conversation he had over email with Noam Chomsky on the topic of formal semantics. While Chomsky has been very open about his views on semantics for a long time, this interview is worth reading for working linguists because Collins frames the conversation around work by linguists—Heim & Kratzer, and Larson & Segal—rather than philosophers—Quine, Austin, Wittgenstein, Frege, et al.

You should read it for yourself, but I’d like to highlight one passage that jumped out at me. Of the current state of the field, Chomsky says:

Work in formal semantics has been some of the most exciting parts of the field in recent years, but it hasn’t been treated with the kind of critical analysis that other parts of syntax (including generative phonology) have been within generative grammar since its origins. Questions about explanatory power, simplicity, learnability, generality, evolvability, and so. More as a descriptive technology. That raises questions.

p 5. (emphasis mine)

It’s true that formal semantics today is a vibrant field. There’s always new analyses, The methods of formal semantics are being applied to new sets of data, and, indeed, it’s virtually impossible to even write a paper on syntax without a bit of formal semantics. Yet it is also true that almost no one has been thinking about the theory underpinning the analytical technology. As a result, I don’t think many working semanticists are even aware that there is such a theory, or if they are aware, they tend to wave their hands, saying “that’s philosophy”. Formal semanticists, it seems, have effectively gaslit themselves.

Chomsky’s framing here is interesting, too. He could be understood as suggesting that formal semantics could engage in theoretical inquiry while maintaining its vibrancy. It’s not clear that this is the case though. Currently, formal semantics bears a striking similarity to the machine-learning/neural-nets style of AI, in that both are being applied to a very wide array of “problems” but a closer look at the respective technologies very likely would cause us to question whether they should be. Obviously, the stakes are different—no one’s ever been injured in a car crash because they used lambdas to analyze a speech act—but the principle is the same.

But I digress. Collins and Chomsky’s conversation is interesting and very accessible to anyone who familiar with Heim & Kratzer-style semantics. It’s well worth a read.

Break up Big University; Create Jobs

This argument in this article (tweeted out by Shit Academics Say) is just designed to pit one group of workers (sessional lecturers) against another (tenured faculty). This is because it ignores the fact that the number of faculty positions at least in Canada is kept artificially low.

Notice that there is no mention whatsoever of class-sizes. Coming from UofT, I can tell you, class sizes have been out of control for years. My intro bio class so big that no lecture halls could house it. Lectures were in the 1730-seat Convocation Hall.

Convocation hall is a beautiful building but it is designed for ceremony, not pedagogy. There is no chalkboard or whiteboard, and if there were, the students in the upper balcony wouldn't be able to read them. What's more the seats have no writing surface for note-taking.

More recently, I taught a "general interest" linguistics course (a "bird course") so big that it also couldn't be housed in a proper lecture hall. Instead we had what was basically a movie theatre. The lights were perpetually dimmed, and again, no chalkboard.

These sorts of non-classrooms really only allow for one type of teaching style, possibly the worst type: A lecturer droning on about a slide deck.

Beyond just the lectures, it's quite impossible for all 1000+ students of such a class to have direct access to their professors in office hours. There aren't enough hours in the day.

(Of course, most students don't go to office hours. It might make a good action, though, for student unions to organize students to go to office hours en masse. Not to shout slogans at professors, but just to ask for help)

Clearly, UofT, the largest university in Canada, has reached its capacity of students.

Imagine, though, if we kept tenure and the researcher/teacher model of academia and put hard limits on class sizes. Say, 200 for 1st yr classes, 100 for 2nd yr and so on. How would that affect things?

The neoliberal response would probably be "well, you'd have to have fewer students, probably only well-off white students."
But there's another possibility: Expand the faculty size by creating new universities.

This could mean founding a brand new university, or it could mean splitting up oversized universities. UofT, for instance, has three campuses: Downtown, Scarborough, and Mississauga. Why not spin them off from each other?

There are definitely ways to do this that I haven't thought about, and none of them are perfect, and all of them require public funding. But that's true of any societal problem.

But we can't really expect to solve the problem without an adequate diagnosis of the problem's source.
There's no shortage of qualified educators, nor is there a shortage of people who want/need an education.The problem is infrastructure.

So, whenever someone makes an argument pitting workers against workers, it can only really serve to obscure the fact that the problem is elsewhere—with management, with bureaucracy, with politicians.

Originally tweeted by Dan Milway (@thrilway) on June 26, 2021.