Sober also points out that criticism of a competing theory, such as evolution, is not in-and-of-itself a test of ID. Proponents of ID must construct a theory that makes its own predictions in order for the theory to be testable. To contend that evolutionary processes cannot produce "irreducibly complex" adaptations merely changes the subject, Sober argues.
"When scientific theories compete with each other, the usual pattern is that independently attested auxiliary propositions allow the theories to make predictions that disagree with each other," Sober writes. "No such auxiliary propositions allow ID to do this." In developing this idea, Sober makes use of ideas that the French philosopher Pierre Duhem developed in connection with physical theories theories usually do not, all by themselves, make testable predictions. Rather, they do so only when supplemented with auxiliary information. For example, the laws of optics do not, by themselves, predict when eclipses will occur; they do so when independently justified claims about the positions of the earth, moon, and sun are taken into account.
Similarly, ID claims make predictions when they are supplemented by auxiliary claims. The problem is that these auxiliary assumptions about the putative designer's goals and abilities are not independently justified. Surprisingly, this is a point that several ID proponents concede.