The main claim of this paper is that the ergon inference (NE I 7) does
not entail a step from the characteristic activity of man to the good of
man. The speciílcation of the ergon is required for the evaluation of
anything that has a function, but the evaluation itself is logically independent
of that specification.