This article represents a first turn of talk in an ongoing dialogue. Our focus here is on the value of sustained dialogue across the boundaries of research discourses. We argue that such dialogue can illuminate the categories of thought and action (Bourdieu & Wacquant, 1992) that "we" in a given discourse take for granted, situate them as choices arising in particular sociohistorical circumstances, and let us imagine how things might be otherwise. In particular, we confront conventional theories and practices in psychometrics, as reflected in school-based testing, with alternatives from sociocultural and situative discourses. We illustrate our argument with contrasting examples of evidence used to address questions of learning, reform, and fairness. We close with some general suggestions for promoting dialogue across research discourses. (Contains 7 footnotes.)