Form, function, fieldwork: An interview with Michael Yoshitaka Erlewine

Kevin Martens Wong speaks to Michael Yoshitaka Erlewine (“Mitcho”), an assistant professor and linguist at the National University of Singapore. Mitcho specialises in doing fieldwork on under-studied languages and understanding how they contribute to our current understanding of formal linguistic theory and the nature of language.

---

You started with teaching, and then you went to Mozilla, and then you went to graduate school in linguistics at MIT. Why this path?

I started college as an undergraduate who knew that linguistics existed and it could be something that I would like to do. Then, in college I was lucky, I was in a very good place, and I discovered that, yes, it was what I wanted to do. So when I finished my undergraduate studies, I knew that I was probably going to go back for graduate study, but I decided to take some time to do some different things. I’d enjoyed language teaching and I’d enjoyed programming just as a sort of hobby, so I decided to try those out a little bit. I was fortunate in that I could connect those things with my interests in linguistics along the way. At Mozilla, for example, I worked on this experimental project, Ubiquity, that actually took advantage of some of my linguistics expertise. But my goal was always that I would be going back to grad school.

I generally think that taking a little time off and having the experience of not being a student, and being an employee, and having a chance to actually make things, is a good experience. For me, these experiences confirmed that, yes, this is what I want to be doing, and yes, I was very happy to be going to graduate school. Of course, it also gave me an opportunity to be in Taiwan, and I was very interested in living in some place I hadn’t lived before—particularly some place I could improve my Mandarin (laughs).

Why was MIT good for you?

It’s an incredibly stimulating place, and there are lots of students and faculty working on all sorts of different things. When I was starting graduate school, I thought my focus would probably be on Chinese syntax, and one concern I had was that there wasn’t anyone at MIT who worked specifically on Chinese languages. And it turned out that that wasn’t a problem at all. It’s the kind of place where you have to be driven yourself and have to have your own questions and problems you want to work on, but then there are people around you who can engage with anything. I really benefitted from that, and my interests broadened as well. I still work on Mandarin, but I’ve since also worked on a variety of other languages.

Yeah, you ended up working on a number of other languages like the Formosan language Atayal, and the Mayan languages Kaqchikel and Chuj. Why did you decide to do that, instead of just continuing with Mandarin?

There are lots of fascinating things in many languages that we in the field generally don’t know much about. When I lived in Taiwan for a year, I lived in the middle of nowhere, in this town that was predominantly ethnically Atayal. Most people there did not speak the Atayal language anymore, but while I was there, I found an Atayal teacher, this retired guy who used to work for the train service, and an Atayal textbook in Mandarin. We worked through it together, but while we were doing that, at some point, I switched from just learning basic conversational Atayal to starting to ask my teacher some questions about the language—basically doing some elementary linguistic elicitation and gathering more data on how the language was used. And at some point I asked him why he was willing to spend time tutoring someone—a foreigner, no less. He said that not that many people speak Atayal anymore, and maybe in a few generations really very few or no one will, but if there are ways that through talking with people, outsiders like me, that other people can learn about the language, that would be great and something he could be proud of as an Atayal.

That experience was really powerful and really important for me. I had not done fieldwork per se before then—I did elicitation work with Mandarin speakers but it’s different than being in the field—and that made me start to think more about the process of doing fieldwork and working on understudied languages, and how to do that in a way that respects the community and which simultaneously can give us very different perspectives on this thing called ‘language’. That kind of perspective, working outside of the very well-studied Indo-European type languages, is really valuable for the theoretically informed kind of work that I’m interested in.

Kaqchikel and Chuj were happy accidents (laughs). At MIT, everybody has to take a field methods class, and when I took it, the class was on Kaqchikel, a Mayan language. We worked with this wonderful speaker who lived in Boston, and we continued to work with her after the class for some time. In doing that I learned more about Mayan languages and that literature, and ran into some really interesting things in Kaqchikel. I also happened to have a great opportunity to later work with a speaker of Chuj, another Mayan language.

But more broadly, the questions I’m interested in through all this fieldwork on different languages are about how language works in general. What is this thing that somehow we have that other mammals don’t have, that allows us to communicate in this way, allows us to form thoughts in this way, and allows us to acquire this kind of skill in this way? Every language has the potential to teach us something unique about this whole system called language itself. Furthermore—and this has been the core discovery of our young field over the past sixty years—when you actually look under the surface, across totally genetically unrelated languages, you often run into the same patterns. Obviously that’s something we want to explain, but in order to do that you have to start looking at languages that are genetically unrelated.

And so something that I enjoy—and it’s the way I happen to work—is to go out and explicitly look at languages that are really somewhat unrelated. And that takes time, because you have to invest in knowledge of the language and each relevant literature. But I think the potential rewards, the kinds of connections I might be able to make, definitely make it worth doing.

You’re currently working on Batak Toba. Why Batak Toba?

One attraction for me of moving to Singapore was the ability to do fieldwork in this part of the world. There are hundreds of languages that you can get to within a few hours from here and are really fascinating options for fieldwork. Batak Toba is the first language that I’ve started working on in Singapore. There’s a colleague of mine in Theatre Studies who is herself Batak, and she took me to the Batak church that meets here in Singapore and introduced me to a couple of friends who grew up in the Lake Toba area near Medan, in Sumatra, and now live in Singapore. I’ve been meeting with some of these speakers and slowly learning about the language, and that’s been incredibly fun. It was a nice coincidence but also speaks to the opportunities that are available in a city like a Singapore.

Singapore is exciting for me, not just because I can fly to interesting places where there are understudied languages, but there are also understudied languages that are right here in the city. It can be hard to find these people, but it’s possible, and there are tremendous opportunities there.

Formal theory traditionally studies the internal grammar of individual speakers. How do you deal with variation in your data?

That’s a great question. One thing that’s important is to be clear about what we are studying. Something that does come from this generative tradition is that even the internal grammar of an individual speaker, their I-language, is still worth understanding, and I truly believe that. At the same time, there’s a question of how you present that and how you work with it. If I work with one speaker of Batak Toba, I don’t want to say, this is Batak Toba—because this is actually the grammar of this particular speaker of Batak Toba. And I think we need to be clear about that, because we know that variation exists, and that is very relevant for linguistic theory.

In terms of dealing with variation in my own work, there are lots of practical questions when doing elicitation-based work. So one thing I try to make sure of is, I want the data I am building a theory based on to accurately represent at least the speaker’s variety of the language—in other words, I want to make sure that important contrasts are reproducible with that speaker across sessions, and I want the data to be internally consistent.

Once I have that and then I have multiple speakers that truly diverge, then I think it’s actually an opportunity to understand the way that language can vary. At first, when you run into this, it can feel like a headache, but if you spend more time with the data and try to sort it out, and try to understand, what exactly is the difference between these speakers, then there’s an opportunity to understand what aspects of this behavior go together. If there’s one thing that separates these speakers who are otherwise very similar, can we use this to our advantage and understand what is that one thing that differs? If we deal with it carefully, we can definitely make use of observed variation, and it can tell us a lot more than what we previously thought we might find.

Your formal work kind of crosses into functional linguistics. Do you think the boundary between formal and functional linguistics is becoming less rigid?

Yeah that’s a very good question, and I actually think about this question a lot. I think our field has an unfortunate history where these terms, capital-F Formal and capital-F Functional, have corresponded to certain groups of people and traditions. These are not inherently incompatible, but have in practice been separated for a period of time.

I think speaking from the Formal side—because that’s my training and that’s generally the academic community I work in—and working in syntax/semantics, there’s actually a lot of work that seeks explanation through function. Even though we might not call this kind of work “functionalist,” at the end of the day it is, in the sense that when I look at some puzzle and try to explain why this language does what it does, my explanation can take the form of, “this is how this expression gives us a usable meaning, to accomplish this particular communicative function”. And having to actually be able to compute a meaning for a certain structure limits the kinds of syntactic analysis I might consider. And so, working at the interface of syntax and semantics, I am in a sense very much interested in functionalist explanations with a lowercase ‘f’.

I absolutely don’t think that everything in language could or should be explained in terms of purely formal properties of grammatical structure, independent of their contribution to meaning. That sort of extreme caricature of formalism, I certainly don’t agree with, and I don’t think I am an outlier in the capital-F Formal community. Although the shift might have happened quietly, I think there are a lot of people who use formal approaches to pursue very explicit, explanatory theories where the explanations have to do with what the function of these expressions are. If you really care about the connection between syntax and semantics I think you’re forced to pursue what is in some sense lowercase ‘f’ functionalist explanations.

In general, I’m often attracted to puzzles where it’s not obvious whether the answer is syntax or semantics. I poke at the puzzle, and sometimes the answer is syntax, and sometimes it’s semantics—but you have to know what is possible cross-linguistically, what kinds of things happen. You also need to understand how meaning is conveyed, and how we want to model meaning, and what is possible there.

What are some challenges you’ve faced doing fieldwork in this part of the world?

One is that this is a part of the world where there aren’t clearly monolingual native speakers. People are just immersed in cultures where they are exposed from a young age to many different languages. All the speakers I work with are at least bilingual, but very often actually speak more than that, and to try to understand and recognise the possible influences of the broader linguistic background that this individual speaker has, is something that is important to do and can be difficult to do. Working with students here, I’ve also recently become interested in Singlish, and this is also a challenge there—because Singlish is on this diglossic spectrum where Singapore Standard English and Singapore Colloquial English coexist, and so it’s difficult sometimes empirically to pinpoint exactly where a particular speaker lies in a particular instance.

Another, more general challenge is that a lot of the work I’m interested in theoretically has to do with how we respond to different discourse situations. So that has to do with what the preceding discourse is, what the speaker’s knowledge is, what the speaker thinks the addressee’s knowledge is, etc. Practically, in order to do that kind of work, it means really having to control for various factors, setting up stories and elaborate contexts and things like that. It also means having to try to get inside the speaker’s head and share some of their relevant cultural expectations. And that’s tricky.

Do you have any advice for people who want to start work on understudied languages?

My first reaction is, what’s stopping you? You should get out there!

My next reaction is that you want to be cognisant of how you get out there. You have to understand the desires of the community or the individual speakers you work with and to respect their time and interests. However, in general, I think this is tremendously important work and work that all of us should be doing. There are languages that are understudied or completely unstudied all around, many of which will not be spoken in 100 years. If that community allows it and is interested, it takes a lot of work, and it takes a lot of time, but I think there is almost a moral imperative for linguists living in this age to work with these communities. Both to support those communities’ interests and also as a scientific endeavor, to document what the range of this human ability called ‘language’ is.