Tuesday, December 24, 2019

Two Poles in North American Linguistics Departments


In this post, I will lay out the contours of two evolving poles in North American linguistics departments: psycholinguistics/computational versus fieldwork/documentation. I am not claiming that any particular department exactly matches the descriptions I have given below. Many departments are a mix of the two basic types, but my feeling is that things are changing very rapidly and departments are tending to gravitate toward one of the two poles (more frequently toward the psycholinguistics/computational pole).

 I place these two poles with respect to what I call the classical model, which is basically MIT during the period when I was a graduate student there (1988-1993) and other departments from around that time. One can see the two poles as two very different ways that the classical model is being transformed empirically. The following diagram illustrates the metaphor of two poles around a center:

Psycholinguistics/Computational     ß     Classical     à     Fieldwork/Documentation

Ultimately, the questions are: (a) What do these types of linguistics departments have in common in terms of fundamental goals (e.g., understanding the human language faculty)? (b) How they can co-exist? (c) Are there drawbacks for departments being located at the extremes of the scale? (d) What shape should an American linguistics department take in the future? (e) What is the future of linguistics as a distinct scientific discipline? I will not attempt to answer these big questions in this post. Rather, I hope that by writing this post I lay the groundwork for further discussion.

Classical

In this department, the focus is on the core areas of linguistics, meaning syntax, phonology and semantics. The star faculty members are the ones who have made scientifically significant contributions towards understanding the human language faculty in one of these areas. Usually, there are some other minor specializations represented in the department (varying from department to department), such as sociolinguistics, phonetics, historical linguistics or language acquisition. These specializations are minor in the sense that there may be only one faculty member working in them, and a small minority of the students specializing in them.

In each of the core areas there is a dominant theoretical framework that the students are working in. For example, in syntax the framework could be Principles and Parameters syntax. In phonology, the framework could be Optimality Theory. And in semantics the framework could be Montague Grammar. Students take it as their primary goal to test and develop such theoretical frameworks.

Students typically write a thesis focusing on one of the core subfields. Methodologically, these theses are defined by standard argumentation in the different subfields. For example, a thesis on syntax would be expected to apply standard tests for constituent structure, islandhood/locality, binding theory, etc. The data for theses usually comes from secondary sources (such as descriptive dictionaries and grammars, linguistics papers), the native speaker judgments of the student in their own language or some basic/rudimentary fieldwork (asking your Italian friends about the interpretation of sentences with impersonal si).

In the classical department, there may be some psycholinguistic or computational research being done. Or there may be theses based on fieldwork. But this kind of work does not dominate the department in anyway. Rather, the department is defined by the goal of understanding the human language faculty (Universal Grammar) by developing theoretical frameworks (formal models) for syntax, phonology and semantics. Psycholinguistics, computational and fieldwork are seen as tools to achieve this goal.

Psycholinguistics/Computational

In this department, the focus is on psycholinguistic research. Note that I do not use the term “experimental linguistics” since I think that the way the term is applied nowadays is misleading. For example, eliciting grammaticality judgements or getting sentence translations in fieldwork definitely constitute carefully designed experiments. But when people use the term “experimental linguistics” they are implicitly excluding such data. This topic is the subject for a different blog post.

The star attraction in this enterprise is neurolinguistics (fMRI, MEG), and the acquisition of costly machines. But in addition to neurolinguistics, other methods include eye tracking, self-paced reading, child language acquisition studies, second language acquisition studies, Mechanical Turk studies, amongst many others. Students specializing in psycholinguistics need to have a working knowledge of statistics in order to interpret the results of their experiments. And so such students will typically take statistics classes during their graduate education.

In addition to psycholinguistics, such a department has a strong computational component, including computational phonology, etc. Students specializing in computational methods need to learn programming languages such as Python, and to create computational models to model various aspects of language acquisition and use.

The role of classical core linguistics courses such as Syntax/Phonology/Semantics is to provide students topics to study in their psycholinguistic and computational research.

An important goal of work in this kind of department is to obtain psycholinguistic and computational results that bear on the theory of the human language faculty (UG). People are trying to answer questions that come up in trying to formulate a theory of the human language faculty, using psycholinguistic and computational methods. But another related goal is simply to understand the processes involved in using language.

On the psycholinguistics/computational model for a department, there may be some fieldwork activity going on. For example, there may be Field Methods class, and a couple of the students might be involved in fieldwork. But most of the activity (e.g., 90%) in the department (talks, workshops, QPs, dissertations, RA positions, visitors, grants, etc.) revolves around psycholinguistics and computational linguistics, and not around fieldwork.

Fieldwork/Documentation

In this department, the focus is on linguistic fieldwork on less studied and endangered languages. The faculty and students in such a department will often specialize in a particular language or language family. Students will spend a significant amount of time (maybe a year or more) during their graduate studies in the field doing fieldwork. Funding for this fieldwork will come from a faculty grant, a grant that the student applies for (e.g., DDRIG, DEL, ELDP, etc.) or summer funding from the department. Dissertations usually have the name of a particular language or language family in their title. Students need to master fieldwork methodology and software programs such as Praat, ELAN, FLEx, etc. They also need to master the use of audio and video recording equipment. Of special interest to such students is CoLang, a summer school on language documentation, where many of the necessary skills are taught.

The role of classical core linguistics courses such as Syntax/Phonology/Semantics is to provide students topics to look into while doing their fieldwork.

An important goal of work in this kind of department is to obtain results from fieldwork that bear on the theory of the human language faculty. But another goal is to help communities document their languages. So students in such department will often be involved in creating grammars, dictionaries and pedagogical materials.

Of course, there may be people in such a department who apply psycholinguistic methods in doing fieldwork. They may run controlled psycholinguistic experiments in the field with large groups of consultants and then do statistics on the results. But still, the main focus of research in the department is fieldwork and documentation, not psycholinguistics and computational linguistics.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.