Starting from Dennett's distinction between personal and sub-personal levels of description, I consider the relationships amongst three levels: the personal level, the level of information-processing mechanisms, and the level of neurobiology. I defend a conception of the relationship between the personal level and the sub-personal level of information-processing mechanisms as “interaction without reduction”. Even given a nonreductionist conception of persons, philosophical theorizing sometimes supports downward inferences from the personal to the sub-personal level. An example of a downward inference is provided and an objection is considered.