I study the grammar of conversational speech, using databases of spoken English made available through BYU corpora and the Linguistic Data Consortium (LDC). I explore the use conditions associated with grammatical constructions and intonation patterns—a research area often referred to as discourse pragmatics. My research in discourse pragmatics focuses on the prosodic and grammatical expression of the pragmatic roles topic and focus, and on what grammatical mechanisms people use to do conversational work like expressing an emotional reaction to something, introducing a new topic into the conversation, announcing forthcoming propositional content and assessing what has been said before. This work is a collaborative effort with several current and former CU students, including Jason Brenier (on the nonstandard 'copula doubling' pattern in English) and Hartwell Francis (on new topics that are also sentence subjects). For findings of an NSF-sponsored project on conversational reference on which Hartwell and I worked, see the Lexical Subjects site. In recent work with CU post-doctoral student Jill Duffield, we are exploring explanations for the prevalence of subject relative-clauses in English conversation. We are investigating the claim that the subject relative-clause (as represented by the bracketed portion of cars [that are designed with human beings in mind]) is as common as it is not because of general-purpose interpretive or production constaints but because it is part of an entrenched discourse routine, the presentational relative-clause construction (e.g., There are a lot of people who babysit in their homes). Recently, Jill and I have begun a series of experiments that we hope will validate a processing-based explanation (ambiguity avoidance) for one of the most enduring 'purely syntactic' constraints--the so-called that-trace effect (which we refer to, following Postal 2004, as the anti-complementizer effect, or ACE). ACE names the negative judgment that language users make about a certain kind of long-distance dependency—those in which a subject is instantiatiated in the pre-clausal filler position and the complementizer that appears before the 'gapped' argument, as in, e.g., the wh-question *Who did she believe that left? or the relative clause *the person we believed that was hired. A fact that we think will help us to explain the nature of ACE is this: the sentence Who did she believe that left? has an unwelcome reading, in which it asks 'Who among the people that left did she believe?'. We also think it is significant that 'violations' like the following are fine: What did she say that was a lie? Why? We believe that our psycholinguistic research program will provide some intuitively satsifying answers.