My research investigates how we manage to use language to get things (e.g. communication, science, coordinated action) done. I view this as part of the general project of cognitive science. It cannot be accomplished without taking seriously how language, the mind and the world interact. Furthermore, I see it as one of the most exciting ways of understanding how the mind works. According to the predominant approach inspired by Frege and Tarski, sentences are (complex) names for pieces of information. As competent speakers, we are all assumed to know what information is named by each sentence. We can therefore communicate information by uttering sentences that are mutually known to pick out that information. Some sentences don't name a particular piece of information, e.g. Did Bob dance? or Dance Bob!, but are held to name different kinds of contents, e.g. issues or properties. My research recognizes the deep insights of this approach but advocates for a more general and nuanced approach. I think sentences are recipes for getting at information, issues and rankings of pieces of information. On this approach, communication does not work by mutually appreciating the piece of information named by a sentence, but rather by the speaker using sentences that can be used to reconstruct the speaker's state of mind. Since that state of mind has a content, this ensures that content is communicated. This approach is inspired by heretically viewing the ideas of anti-formalist 'ordinary language' philosophers (Wittgenstein, Austin, Strawson, Grice, Searle) through the lens of work in linguistics, logic and computer science on dynamic semantics and planning. My papers and dissertation argue that this approach explains phenomena involving conditionals, imperatives and questions better than the standard picture. This is done by constructing formal models of how certain sentences work, interpreting those models in accordance with the perspective articulated above and arguing that they explain more than the models offered within the standard framework. Yet, this is more of a reconciliation than a replacement, since I too believe that formal models are necessary and that the core features of the standard approach (e.g. compositionality) are to be preserved.