I'm a big fan of Karl Popper's work. I learned about him when reading the book Empirical Linguistics by Geoffrey Sampson. At the time, it was a pretty iconoclastic publication, since it directly struck against the assumption of nativism by framing the study of language as something that could be evidence-based in a way where hypotheses were truly falsifiable. The ability to collect and process large amounts of data pertinent to language make it a lot easier to strike down some of the more inscrutable theories of the '90s and '00s -- at least to those who are willing to do real science.