They are underspecified (that is, schematic); they are generally learned by induction through recurrent exposure, rather than propositionally; and they carry with them expectations about the object or situation not available to direct perception. Thus, we use “schematic structure” as an umbrella term 3 Given the data and research questions that social demographers confront, the move from a model of outcomes to a model of processes is particularly difficult here, although perhaps also particularly important