It is widely agreed that ontological reduction is possible, that the ontology of one theory can be shown to be nothing over and above the ontology of a distinct theory. However, it is also widely agreed that one assesses a theory’s ontology by determining what it says there is. I show that there is a tension between these orthodox positions. To resolve this tension, I propose and defend the view that the ontological commitments of a statement are sensitive to the theory in which it is embedded. / text