Abstract

It is common, following Quine, to look to what theories say to determine the ontological commitments of a scientific discipline. But methods and practices are equally telling. This paper considers early doctrines in econometrics. It argues that what is directly confirmed in tests of the theory will not support the applications to which the theory is to be put unless we can assume a kind of stability and atomism characteristic of capacities. The leap from confirmation to application will only be valid in a world where capacities are at work.