What I think Natural Language Technology, AI and Cognitive Science are all about: A Non-Technical Explanation

NLT.

Work on natural language technology deals with designing computer systems that use natural (aka human) language. It's interdisciplinary by nature, drawing mainly on work in theoretical linguistics and techniques from research in artificial intelligence, machine learning, and computer science more generally. It encompasses work in computational linguistuics and natural language processing. Multimodal interfaces are systems that allow you to supply a computer with input via multiple natural mode at a time, usually speech combined with, for example, gestures, handwritten symbols or facial expressions.

AI.

AI doesn't mean building machines that are "conscious" or that "think." This is science fiction. Science does not yet understand what consciousness or thinking are. (Do you?) AI also doesn't necessarily mean building robots, although that is one application. That said, artificial intelligence is a bit of a misleading term, maybe even somewhat antiquated these days, especially since no one seems to be able to agree upon a satisfactory definition of "intelligence," human or otherwise. People who do work in AI are usually trying to encode some kind of human inference in such a way that computers can do it. This covers a lot of ground, and you'll find people in AI labs trying to get computers to do things like learning from data (recognizing patterns and doing intelligent approximation), deduction, logic, playing games, communicating naturally with language, gesture, or other natural modes of expression, and seeing (computer vision). So basically, AI is about building computer systems that are flexible (or dare I say, smart) enough to work in the real world.

And yes, this does sometimes include wacky robot hijinx, like soccer-playing robots, museum tour-guide robots, and even... space robots!

So if you ever meet someone who works on artificial intelligence, please don't express your moral opposition to their attempting to build a conscious computer. Consciousness is a fascinating subject, but unforunately it's not science yet, and thus remains the domain of philosophers (some neuroscientists may disagree).

CogSci.

A lot of people like to explain cognitive science as a structured, scientific approach to the study of mental phenomena. Isn't that psychology you say? Well, I think that that's what psychology started out as, e.g. in the work of William James, but at some point clinical treatment of people with mental ailments and diseases also became known as psychology. I have to admit that my knowledge of the history of psychology isn't the most detailed, but the work of people like Freud, Jung, etc. is often all put together under the heading psychology, so it's clearly time for some new labels. So, some more history:

In the 1950's, the behaviorist school, championed by every pigeon's favorite scientist B.F. Skinner, became the dominant approach to experimental psychology, which is the psychology that the people who were doing psychology at the time, that is the people who were still being paid by psychology departments at academic and scientific institutions, but who weren't doing clinical pyschology, were doing. (Now that's a long-distance dependency).

Towards the end of the 1950's, some other ideas starting taking hold in the field. The catalyst for this change is typically regarded as the work of Noam Chomsky (linguist, philospher, political dissident, prolific writer, supposedly the most quoted scholar alive, etc.). These ideas were so influencial, and had affected not just psychology, but also linguistics, philosophy, computer science, etc., that they decided they should give the new interdisciplinary field a new name :) enter cognitive science. There is a famous adage that says that any field which has the word "science" in it's title probably isn't one - but every rule has its exception, and I think cognitive science is just such an exception. According to John Searle, in a lecture I heard him give at Edinburgh in 2002, this was great for philosophers, because now they were doing science, so they could find a way to get funded. Alas, I don't think this trend lasted very long, else I would probably be working on a PhD in philosophy at the moment. But I digress (yet again).

So a popular way of explaining what it is that cognitive scientists do is to talk about the computational theory of mind, or the idea of treating the mind as a computational system (like that thing connected to the monitor that's sitting in front of you, allowing you to read this text right now). People used to make the analogy that the brain was like the hardware and the mind, the software. However, this analogy is a confusion and (I hope) no one really believes it anymore. It stems mainly from a misinterpretation of a really nice metaphor by David Marr. Work in cognitive science often assumes that we can formulate and test hyphotheses about how the mind works, without actually understanding how the brain works. This has produced a wealth of theories and insight into human behavior and cognitive mecahnisms -- but work in cognitive science rests on assumptions that are still very contraversial. People often like to enumerate the various disciplines that get involved in the cognitive science enterprise, so in keeping with the tradition, cog sci draws upon cognitive and evolutionary psychology, linguistics, computer science, neuroscience, anthropology, and the list continues to grow. People are also increasingly realizing the importance of studying social systems to our understanding of the mind. Today, cognitive neuroscientists are trying to understand cognitive mechanisms, or "how the mind works" as Stephen Pinker says, through research on "how brain works."

Please forgive my drastic oversimplification of these issues! I'd like to give these topics the proper treatment and explanation they deserve, but since I'm doing this, I'm procrastinating, and my "real work" beckons :).

So why do I care about this stuff?

Well, not only do I think it's intrinsically interesting, but also, in addition to helping us answer some deep, philosophical questions about language, meaning and thought, as I mentioned above, the discoveries made here have practical and immediate relevance for applications in the real world. Systems incorporating HCI and implementing language technology are still inchoate, but their presence is already widespread, and their potential is incalculable.