When you set a ratchet to tighten a bolt, each turn in the clockwise direction tightens the bolt, but the ratchet will not loosen a bolt when turned counter-clockwise. This concept is often used to describe other processes that only go in one direction and can not be reversed.
It would seem that Behe thinks evolution is like a ratchet. You can degrade genes, but once you lose that function you can’t get it back. This is echoed in his latest essay:
So is this true? I don’t think it is.
Let’s look at tetrapod evolution as an example at the macroscopic level. Fish evolved lungs and limbs, moved onto land, and then lost the adaptations they once had for the aquatic environment their ancestors lived in. So does this mean that tetapods were forever limited to living on land and could never adapt to an aquatic environment after that? Obviously not. Multiple lineages of tetrapods have moved back into aquatic environments, such as whales, crocodiles, and turtles.
We could also look at examples at the molecular level. In one experiment, scientists knocked out the beta-galactosidase gene in E. coli. It forever lost that gene, and according to Behe it could never go back to breaking down lactose. So is that what happened? No. When scientists allowed these E. coli to compete for lactose a new beta-galactosidase gene emerged.
As Behe states, genes can be lost if they aren’t needed. What Behe ignores is the emergence of new genes. There isn’t a ratchet. The process goes in both directions.