This is an interview with Martin Kay. He's one of the grand old men of computational linguistics. He was there at the beginning, in the mid-1950s; he's made major contributions (discussed in this video) and he still has an active research program. He worked with my teacher and colleague, the late David Hays, whom he discusses starting at 4:25 or so. The interview runs a bit over an hour and gets a bit technical here and there. But it gives you a good flavor of the discipline.
Starting at roughly 48 minutes he discusses the statistical methods that came over the field starting in the early 1990s or so. He speculates that the rule-based thinkers like Kay (and his generation and the much of the next) will be making a rapprochement with the statistical folks. Or vice versa. "As a scientist you want to explain what gives rise to those statistics."
This part of the discussion is very important. Much recommended. You might also want to nose around at the site, for the Oxford Internet Institute, which has many interesting videos, such as this one by Ted Nelson musing his way through his career.