I got my doctorate in the psychology of conversation (or psycholinguistics if you want to get fancy). In general, I studied the amazing process by which people mutually assess each other's level of understanding, and constantly and subtly adjust on a second-by-second basis to make sure they are understanding each other and being understood. It's a phenomenally complex, sophisticated, and fascinating collaborative activity.

In my first year at Stanford, I studied how people with different levels of expertise assess each other's level of knowledge and adjust to make sure they are being understood. We found that people were extremely quick and adept at picking up discrepancies between their own and their partner's knowledge and adjusted immediately. We also analysed the ways they adjusted and how effective their strategies were. This resulted in the paper References in Conversation Between Experts and Novices.

My deepest interest in conversation was in the even more subtle ways that we say things without saying them. I called it "off record" conversation, defining off record as "meaning one could reasonably have intended but which they could plausibly deny." You know, those "Gee, I'm hungry," comments that are meant to hurry you up and get ready to go to dinner, or "There's a gas station up ahead" to suggest that maybe we should stop and ask for directions. Or even, "Wanna play Quake?," to mean you're one of the gang. The claim is that there is an intention for the other person to understand the meaning, but the fact that the meaning is stated off record has significance and "softens the blow" of boldly saying it "on record." Along these lines, I wrote a paper with my advisor, Herb Clark, about ostensible invitations, the invitations people issue without wanting the other person to accept them and wanting them to know they shouldn't accept. It's what I call the "Let's Do Lunch" paper.