…it is difficult to express oneself. The act of writing something (which one expects or hopes will be published) is a social act; it becomes—even at its best—all but a lie. To communicate socially (as opposed to communicating personally or humanly) means that one must accept the sluggish fictions of society for at least nine-tenths of one's expression in order to present deceptively the remaining tenth which may be new. Social communication is the doom of every truly felt thought. (Norman Mailer, Advertisements for Myself, p. 244)
When I first learned about "discourse" I approached it sort of like ideology. It consisted, I thought, of all the things people said not because they're true but because they are somehow convenient to the powers that be. I thought of discourse as a theory of what Mailer above calls "social communication". But I've gotten more sophisticated (and more accurate) in my reading of Foucault since then.
It is much more constructive to think of discourse as that which makes it possible to say things that would otherwise be impossible to say, not because they would be suppressed, but because we lacked the epistemic resources to say them. The meaning of words is not defined merely by the system of language, after all. Most sentences are comprehensible only on the background of a great deal of shared knowledge, and the the more specialized your utterances get, the more specialized becomes the relevant body of knowledge. In communicating what we know, we depend crucially on the knowledge that our peers already possess. But in order to leverage that opportunity we have to grant them also a great many things that, we think, they merely believe. Things we know in our hearts are false.
This is not "the doom of every truly felt thought", but it is a real constraint. In the case of scholarship the trick is to begin and end our thinking inside the limits of discourse, i.e., without entertaining for too long thoughts that will not fly in discourse. Many "truly felt thoughts" after all are finally simply wrong. They arise in the privacy of our own minds and, when we speak of them to our close friends, we realize we are talking nonsense. They feel true, but they don't survive scrutiny. While friendships are personal, of course, they are also in another sense "social". So already here we a get a sense of what social communication implies. Scholars, researchers, scientists devote a great deal of time to thinking about things on a socially shared basis. They do not, like novelists, nurture their own private fantasy or nightmare of the society in which they live. Rather, they use their minds to solve problems that have already been acknowledged by others, and they undertake to solve those problems in terms that will be convenient to the intellectual projects of those others.
Many scholars today do, in fact, think of their community as an intellectually oppressive arrangement. But what sort of arrangement would they prefer? If scholars had the luxury of expressing themselves before an audience that held no prior beliefs about the subject and would be happy to believe whatever you tell them, then you would have explain everything from the ground up every time. Hopefully, most of the lies of discourse are lies of polite omission. We talk about the things it is possible to say within the space of a journal article and, occasionally, a book. We don't expect to "rock the century on its heels" (as the back cover of my copy of Mailer's Advertisements brags). We try to make a useful contribution of what we know to what is known.