In the early to mid-20th century, anglophone philosophy famously took a "linguistic turn". Debates about the mind were recast as debates about the terms or concepts we use to describe the mind; debates about free will were recast as debates about the term or concept "freedom"; debates about knowledge were recast as debates about how we do or should use the word "knowledge"; etc. Wittgenstein famously suggested that most philosophical debates were not substantive but rather confusions that arise "when language goes on holiday" (Philosophical Investigations 38); diagnose the linguistic confusion, dissipate the feeling of a problem. J.L. Austin endorsed a vision of philosophy as often largely a matter of extracting the wisdom inherent in the subtleties of ordinary language usage. Bertrand Russell and Rudolf Carnap set philosophers the task of providing logical analyses of ordinary and scientific language and concepts. By no means, of course, did all philosophers go over to an entirely linguistic view of philosophy (not even Austin, Russell, or Carnap), but the shift was substantial and profound. In broad-sweeping histories, the linguistic turn is often characterized as the single most important characteristic or contribution of 20th century philosophy.
I view the linguistic turn partly sociologically -- as driven, in part, by philosophy's need to distinguish itself from rising empirical disciplines, especially psychology, as departmental and disciplinary boundaries grew sharper in anglophone universities. The linguistic turn worked to insulate philosophical discussion from empirical science: Psychologists study the mind, we philosophers study in contrast the concept of the mind. Physicists study matter, we study in contrast the concept of the material. "Analytic" philosophers could thus justify their ignorance of and disconnection from empirical work.
This move, however, could only work when psychology was in its youth and dominated by a behaviorist focus on simple behaviors and reinforcement mechanisms (and, even earlier, introspective psychophysics). As psychology has matured, it has become quite evident -- as indeed it should have been evident all along -- that questions about our words and concepts are themselves also empirical questions and so subject to psychological (and linguistic) study. This has become especially clear recently, I think, with the rise of experimental philosophers who empirically test, using the methods of psychology, philosophers' and ordinary folks' intuitions about the application of philosophical terms. (In fact, Austin himself was fairly empirical in his study of ordinary language, reading through dictionaries, informally surveying students and colleagues.)
A priori, armchair philosophy is thus in an awkward position. The 20th-century justification for analytic philosophy as a distinct a priori discipline appears to be collapsing. I don't think a prioristic philosophers will want to hold on much longer to the view that philosophy is really about linguistic and conceptual analysis. It's clear that psychology and linguistics will soon be able to (perhaps already can) analyze our philosophical concepts and language better than armchair philosophers. Philosophers who want to reserve a space for substantive a priori knowledge through "philosophical intuition", then, have a tough metaphilosophical task cut out for them. George Bealer, Laurence BonJour, and others have been working at it, but I can't say that I myself find their results so far very satisfying.
[Note: Cross-posted at The Splintered Mind.]