Informal biography

It has taken me a long time to find myself professionally. This had a lot to do with my unfortunate tendency to avoid settling down, but also with my combination of interests: a fascination with language going back as long as I can remember, the idea that I should Do Science, and the naive view that the world really can be a Better Place and that people who believe this should not just sit around and wait for it to happen. As far as bringing the first two together, it never occurred to me that there might be anything like a science of the mind — in high school, we weren't exposed to social science (it's nice to see that my high school has changed in this regard) — and, encouraged by my father, I headed off to college to major in physics (later math), with the idea that I'd eventually go into astronomy, satisfying at least the second interest.

At San Diego State College, now San Diego State University, what I remember the most were freshman English (the novels that had been too shocking for my high school), the sociology class where I first saw what it meant to investigate human behavior systematically, the ecology class where I was finally exposed to the wonders of evolution, studying Chinese, but most of all my Junior Year Abroad in Heidelberg, Germany, where, in addition to a lot of plain old fun, I had my first opportunity to learn a second language by interacting with native speakers and my first exposure to people from a variety of other places. At the end of four years, I found myself not really inspired but with a vague sense (based on the advice of a math professor) that what I should be going into was oceanography, of all things.

At this point I conveniently entered the Peace Corps. As with many in my generation, this was a way to avoid deciding what to do with your life, as well as a chance for some adventure and — who knows — even doing Some Good or something. During my four years in Ethiopia, I went from teaching math and science to teaching English to supervising English teachers, never really feeling that what I was doing was at all what Ethiopia needed. So much for Doing Good (or Changing the World). In the process, I got hooked on Ethiopian languages and never opened the oceanography book I'd taken along. From then on, there was no turning back from language.

But I left science behind for awhile. When I returned to the States in 1973, I entered the Master's program in Teaching English as a Second Language (TESL) at UCLA. This started a weird kind of limbo period in my life, in which I lived among Ethiopians, supported myself at odd TESL jobs, and took forever to finish my Master's degree. I found I could not get excited about school and was considering going back to Africa because my do-gooder side was nagging at me.

But I decided to give academia one more shot. I finished my Master's thesis and started the PhD program in Applied Linguistics at UCLA. I hoped that the study of second language acquisition would excite me. The more I knew, however, the more I realized how rudimentary our knowledge was of what was really going on. Without a clear idea of how first language acquisition worked, I wondered what we could say about second. I was inspired for a period with research on memory, especially with John Anderson's ACT model, and did my first-ever experiment, on ways to present second-language vocabulary items so that they would be memorable.

But the real inspiration came in discovering artificial intelligence and computational modeling, initially just the idea of it as presented in an education class. This was in 1983, around the time I should have been well into my dissertation. But it was just the beginning for me. I had to learn enough computer science to be able to do what I wanted to do. Well, at this point I still didn't know quite what it was I wanted to do, but I was going to make a machine smart and it was going to have something to do with learning language. It took me five more years, and it was a very odd PhD I hacked together. All of my research was conducted in the AI Laboratory under Michael Dyer, but I was still officially in the Applied Linguistics PhD program as far as UCLA was concerned. My dissertation was about a sort of spreading-activation model of second-language vocabulary learning, an unwieldy monstrosity which took almost as much time to explain to people as it did to program (which was a very long time indeed). And like most other network models in which the concepts are represented locally, it had very little to say about learning itself.

But it got me out of there (in 1988). And it got me a job. In the Computer Science Department at Indiana University, where I still am. The most exciting thing about IU has been the Cognitive Science Program, which is more than just a label applied to a group of faculty and students; it's a hotbed of interdisciplinary activism. The combination of a a community of people committed to the new dynamical view of cognition, the kind of in-your-face radicalism that I love on the part of a few key colleagues, and an unusually talented group of students was just what I needed. But it wasn't easy. I had a lot more to learn — about psychology and neural networks and what modeling is good for — but I finally found a niche of sorts, using neural-network models to try to figure out how language is learned.

I wrote the above (minus a few edits that I just added) in 2000, I think, and at that time, I didn't expect to have to go back and change it. After all, I was 52, about time to have figured out what you want to be when you grow up. But that was still only two out of three. I'd found a way to do both Science and Language, but what about Changing the World? As I got involved in extra-curricular activism, developed increasingly radical views, and became concerned about the urgency of issues that were very far from theories of how language is learned, I wondered if the expertise I'd built up during my strange career could be brought to bear on these issues.

With a little help from a few friends with similar concerns, I realized that it could. We can only expect the world to change for the better if people everywhere are well informed, if information and knowledge are democratized. New technology has opened up all sorts of new possibilities for how to achieve a revolutionary new democratization of information and knowledge, but this will require tools. These tools can build on some of the achievements of artificial intelligence and cognitive science, which are in large part the sciences of information and knowledge. In particular, because most information is passed on in the form of language, research on how people use language to inform and misinform, on how meaning can be extracted from large amounts of text, and on how translation can be performed automatically are all relevant. In other words, I find that I'm able to apply what I have learned about language, computation, and the mind to the development of these sorts of tools. It has meant something of a shift in direction, and at this writing I am far from achieving anything of actual use, but I have never felt so committed to what I am doing. And I have never gotten so much pleasure out of it.