Strategy and framing in text
Abstract: Research in natural language processing is increasingly taking into account the context -- physical and social -- in which linguistic communication occurs. In this talk, I will present models of text data that exploit assumptions about the social world from which those data emerged. Vector representations of political actors' policy preferences, known as "ideal points", have been an important representational tool in political science since the 1980s. I will describe how textual evidence can help to infer these embeddings in models that capture relationships between language and strategy. The first project considers high-stakes texts generated for the U.S. Supreme Court and explores the assumption that authors of amicus curiae ("friends of the court") are rational agents seeking to maximize expected utility. The second project considers a phenomenon known as "framing", where some aspects of an issue or experience are emphasized at the expense of other aspects. I will discuss our computational linguistic efforts to grapple with framing in American news discourse about immigration.