As conversation analysts have shown, participants can achieve a variety of social actions via talk alone (as with telephone conversations), or alternately via embodied behavior in the absence of talk. In some instances, however, people use talk in conjunction with embodied behavior, articulating precisely what it is they are doing (or alternately, what they would like recipients to believe they are doing). Drawing upon digital videos from the University of California, Santa Barbara's vast collection of naturally occurring interactions, cases in which the aforementioned practice occurs were identified, collected, and analyzed using the established methods of conversation analysis. Self-formulations of embodied actions were found in one of three sequential positions relative to the actions they explicate: before the commission of an action, after its completion or abandonment, or while an action was still underway. Generally, self-formulations either serve as a prophylactic against potential trouble or in response to trouble that has actually transpired. Self-formulations provide recipients with an "official account" of what a person is about to do, has done (or tried to do), or is involved with doing. In formulating their actions, participants show themselves to be oriented not only to possible interactional troubles, but also to the fact that recipients routinely monitor others' in-progress actions, drawing inferences about their import.