In addition to the reasons already cited--politics, trying to make one's arguments impervious to refutation, etc.--I would cite at least two more reasons why academics indulge in obscurantism. First, some of them genuinely enjoy it. I can think of a few academics who get a real thrill out of weaving together impenetrable abstraction after impenetrable abstraction. The writing has a poetic aspect that becomes at least as important as, and in some cases more important than, the content. I once attended a paper about how dance has been a form of protest at certain historical moments (think
Swing Kids), except that instead of saying "dance can be a form of protest" the writer kept saying "our very bodies can become sites of performative subalternity." You could tell by the way he read the paper that the main point was simply to say that phrase and experience a brief intellectual orgasm.
Second, it's easy. Concrete facts in plain English can't be strung together at whim; you actually have to watch what you say and make sure what you're saying matches reality. But abstract concepts in large quantities offer up basically no resistance. Pontificating about post-Foucauldian disintermediation and the ideological inscriptions of sexual liminality in the discursive performance of Otherness is the easiest thing in the world. The terms are so vapid that they can be strung together without any thought at all, as evidenced by the
Postmodern Generator, a now-infamous website that randomly generates entire essays using a database of stock phrases from critical theory.